Peripheral Neuropathy: 2-Volume Set

The Curtis Center 170 S Independence Mall W 300E Philadelphia, Pennsylvania 19106

PERIPHERAL NEUROPATHY Copyright © 2005, Elsevier Inc. All rights reserved.

ISBN 0–7216–9491–8 (2 Volume Set) (Volume 1) 9997637887 (Volume 2) 9997637895

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permissions may be sought directly from Elsevier’s Health Sciences Rights Department in Philadelphia, PA, USA: phone: (⫹1) 215 238 7869, fax: (⫹1) 215 238 2239, e-mail: [email protected]. You may also complete your request online via the Elsevier homepage (http://www.elsevier.com), by selecting ‘Customer Support’ and then ‘Obtaining Permissions.’ Exceptions to copyright are as follows: Artwork for Chapters 13, 25, 27, 28, 37, 43, 47, 99, 100, 108, 116, and 121.

NOTICE Neurology is an ever-changing field. Standard safety precautions must be followed, but as new research and clinical experience broaden our knowledge, changes in treatment and drug therapy may become necessary or appropriate. Readers are advised to check the most current product information provided by the manufacturer of each drug to be administered to verify the recommended dose, the method and duration of administration, and contraindications. It is the responsibility of the licensed prescriber, relying on experience and knowledge of the patient, to determine dosages and the best treatment for each individual patient. Neither the publisher nor the author assumes any liability for any injury and/or damage to persons or property arising from this publication. THE PUBLISHER First edition 1975. Second edition 1984. Third edition 1993.

Library of Congress Cataloging-in-Publication Data Peripheral neuropathy / [edited by] Peter James Dyck, P. K. Thomas.—4th ed. p. ; cm. Includes bibliographical references and index. ISBN 0–7216–9491–8 (set) 1. Nerves, Peripheral–Diseases. I. Dyck, Peter James, II. Thomas, P. K. (Peter Kynaston). [DNLM: 1. Peripheral Nervous System Diseases. 2. Peripheral Nerves—pathology. WL 500 P445 2005] RC409.P46 2005 616.8⬘5607–dc22 2004051387 Publishing Director, Global Medicine: Susan Pioli Developmental Editor: Jennifer Ehlers Editorial Assistant: Joan Ryan Designer: Gene Harris Marketing Manager: Michael Passanante

Working together to grow libraries in developing countries www.elsevier.com | www.bookaid.org | www.sabre.org

Printed in the United States of America Last digit is the print number:

9

8

7

6

5

4

3

2

1

Contributing Authors

AMMAR AL-CHALABI, PH.D., F.R.C.P. Senior Lecturer in Neurology, Institute of Psychiatry, London, United Kingdom; Instructor in Complex Genetics, Cold Spring Harbor Laboratory, Cold Spring Harbor, New York; Visiting Scientist in Neurology, Massachusetts General Hospital, Boston, Massachusetts; Consultant Neurologist, King’s College Hospital, London,United Kingdom Sporadic Motor Neuron Degeneration; Inherited Motor Neuron Degeneration

DORIS-EVA BAMIOU, M.D., M.SC.(DISTINCTION) Clinical Research Fellow, Neuro-otology Department, The National Hospital for Neurology and Neurosurgery, London, United Kingdom Diseases of the Eighth Cranial Nerve

ROBERT W. BANKS, PH.D., D.SC. Lecturer, School of Biological and Biomedical Sciences, University of Durham, Durham, United Kingdom The Muscle Spindle

RICHARD J. BAROHN, M.D. Professor and Chairman, Department of Neurology, and Professor of Pathology, University of Kansas Medical Center, Kansas City, Kansas Polyneuropathy Caused by Nutritional and Vitamin Deficiency

TIMOTHY J. BENSTEAD, M.D., F.R.C.P.(C) Professor, Division of Neurology, Dalhousie University; QEII Health Sciences Centre, Halifax, Nova Scotia, Canada Differential Diagnosis of Polyneuropathy

ALAN R. BERGER, M.D. Professor and Associate Chairman, Department of Neurology, University of Florida; Director, Neuroscience Institute, Shands Jacksonville, Jacksonville, Florida Human Toxic Neuropathy Caused by Industrial Agents

C.-H. BERTHOLD, PH.D., M.D. Professor Emeritus, Section of Neuroanatomy, Department of Anatomy and Cell Biology, Göteborg University, Göteborg, Sweden Microscopic Anatomy of the Peripheral Nervous System

ADIL E. BHARUCHA, M.D. Associate Professor of Medicine, Mayo Clinic College of Medicine; Consultant in Gastroenterology and Hepatology, Mayo Clinic, Rochester, Minnesota Autonomic and Somatic Systems to the Anorectum and Pelvic Floor; Management of Gut Dysmotility

ROLFE BIRCH, M.A., M.B., B.CHIR., F.R.C.S.EDIN., F.R.C.S.&P.GLAS., F.R.C.S.ENG., M.CHIR. (Personal Chair) Professor of Neurological Orthopaedic Surgery, University College, London; Visiting Professor, Imperial College, London; Orthopaedic Surgeon, and Head of Department, Peripheral Nerve Injury Unit, Royal National Orthopaedic Hospital, Stanmore, Middlesex, United Kingdom Operating on Peripheral Nerves

HERBERT L. BONKOVSKY, M.D. Professor, and Director, The Liver-Biliary-Pancreatic Center, Department of Medicine, Biochemistry & Molecular Biology, University of Massachusetts, Worcester, Massachusetts Porphyric Neuropathy

AUGUST M. BOOTH, PH.D. Research Assistant Professor Emeritus, Department of Pharmacology, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania Autonomic Systems to the Urinary Bladder and Sexual Organs

E. PETER BOSCH, M.D. Professor of Neurology, Mayo Medical School, Rochester, Minnesota; Consultant, Department of Neurology, Mayo Clinic Scottsdale, Scottsdale, Arizona Peripheral Neuropathy Associated with Lymphoma, Leukemia, and Myeloproliferative Disorders

v

vi

Contributing Authors

HUGH BOSTOCK, M.SC., PH.D. Professor of Neurophysiology, Institute of Neurology, University College London, London, United Kingdom Nerve Excitability Measures: Biophysical Basis and Use in the Investigation of Peripheral Nerve Disease

FRANK BRADKE, PH.D. Independent Principal Investigator on Associate Professor Level, Max Planck Institute of Neurobiology, Munich, Germany Guidance of Axons to Targets in Development and in Disease

ROSCOE O. BRADY, M.D. Branch Chief, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland Fabry’s Disease

STEPHEN BRIMIJOIN, PH.D. Professor of Pharmacology, Mayo Clinic College of Medicine, Rochester, Minnesota Axonal Transport: Properties, Mechanisms, and Role in Nerve Disease

DEBORAH BUCK, PH.D. Senior Research Associate, Department of Primary Care, University of Liverpool, Liverpool, United Kingdom Health Outcomes and Quality of Life

RICHARD P. BUNGE, M.D.* Professor of Neurological Surgery, Cell Biology and Anatomy, and Neurology; Scientific Director, The Miami Project to Cure Paralysis; University of Miami School of Medicine, Miami, Florida Gross Anatomy of the Peripheral Nervous System

DAVID BURKE, M.D., D.SC. Dean of Research and Development, College of Health Sciences, University of Sydney, Sydney, New South Wales, Australia Nerve Excitability Measures: Biophysical Basis and Use in the Investigation of Peripheral Nerve Disease

JAMES P. BURKE, PH.D. Assistant Professor of Epidemiology, Mayo Medical School, Mayo Clinic, Rochester, Minnesota

MICHAEL CAMILLERI, M.D., M.PHIL. (LOND.), F.R.C.P.(LOND.), F.R.C.P.(EDIN.), F.A.C.P., F.A.C.G. Atherton and Winifred W. Bean Professor, Professor of Medicine and Physiology, Mayo Medical School; Consultant in Gastroenterology, Mayo Clinic, Rochester, Minnesota Management of Gut Dysmotility

J. AIDAN CARNEY, M.D., PH.D. Professor of Pathology (Emeritus), Mayo Medical School; Consultant in Pathology (Emeritus), Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota Multiple Endocrine Neoplasia, Type 2B

COLIN CHALK, M.D., C.M., F.R.C.P.(C.) Associate Professor, Department of Neurology and Neurosurgery, McGill University; Associate Physician, Division of Neurology, Department of Medicine, McGill University Health Centre, Montréal General Hospital site, Montréal, Québec, Canada Diseases of Spinal Roots

PHILLIP F. CHANCE, M.D. Professor of Pediatrics and Neurology, Division Head, Genetics and Development, Research Affiliate, Center on Human Development and Disability, and Coordinator, Research Emphasis Area on Joubert Syndrome, University of Washington, Seattle, Washington Hereditary Motor and Sensory Neuropathies: An Overview of Clinical, Genetic, Electrophysiologic, and Pathologic Features; Hereditary Motor and Sensory Neuropathies Involving Altered Dosage or Mutation of PMP22: The CMT1A Duplication and HNPP Deletion

S. Y. CHIU, PH.D. Professor of Physiology, University of Wisconsin, Madison, Wisconsin Channel Function in Mammalian Axons and Support Cells

MICHAEL P. COLLINS, M.D. Neuromuscular Diseases Consultant, Neurosciences Department, Marshfield Clinic, Marshfield, Wisconsin Neuropathies with Systemic Vasculitis

Epidemiologic Approaches to Peripheral Neuropathy

TED M. BURNS, M.D. Assistant Professor, Department of Neurology, University of Virginia, Charlottesville, Virginia Mechanisms of Acute and Chronic Compression Neuropathy; Peripheral Neuropathies in Infants and Children: Polyneuropathies, Mononeuropathies, Plexopathies, and Radiculopathies

*Deceased.

JOHN H. COOTE, PH.D., D.SC. Professor of Physiology (Emeritus), University of Birmingham, Birmingham, United Kingdom Neural Control of Cardiac Function

JAMES J. CORBETT, M.D. McCarty Professor and Chairman of Neurology, and Professor of Ophthalmology, University of Mississippi School of Medicine, Jackson, Mississippi; Instructor Visiting Lecturer in Ophthalmology, Harvard Medical School, Boston, Massachusetts The Pupil

Contributing Authors

DAVID R. CORNBLATH, M.D. Professor of Neurology, Johns Hopkins University School of Medicine; Neurologist, and Director, Neurology EMG Laboratory, The Johns Hopkins Hospital, Baltimore, Maryland Peripheral Neuropathies in Human Immunodeficiency Virus Infection

T. COWEN, PH.D. Professor of Neurobiology of Aging, Department of Anatomy and Developmental Biology, University College London, Royal Free Campus, London, United Kingdom Aging in the Peripheral Nervous System

PAULA CUDIA, M.D. Research Fellow, Centre for Neuromuscular Disease, The National Hospital for Neurology, London, United Kingdom Peripheral Nerve Diseases Associated with Mitochondrial Respiratory Chain Dysfunction

BASIL T. DARRAS, M.D. Professor of Neurology (Pediatrics), Harvard Medical School; Senior Associate in Neurology, Director, Neuromuscular Program, and Director, Residency Training Program and Outpatient Clinics, Department of Neurology, Children’s Hospital Boston, Boston, Massachusetts Peripheral Neuropathies in Infants and Children: Polyneuropathies, Mononeuropathies, Plexopathies, and Radiculopathies

JENNY L. DAVIES, B.A. Data Analyst II, Peripheral Neuropathy Research Center, Mayo Clinic, Rochester, Minnesota Nerve Tests Expressed as Percentiles, Normal Deviates, and Composite Scores

WILLIAM C. DE GROAT, M.SC., PH.D. Department of Pharmacology, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania

vii

MICHAEL DONAGHY, D.PHIL., F.R.C.P. Reader in Clinical Neurology, University of Oxford, Oxford; Consultant Neurologist, Department of Clinical Neurology, The Radcliffe Infirmary, Oxford; Honorary Civilian Consultant in Neurology to the Army, United Kingdom Lumbosacral Plexus Lesions

PETER J. DYCK, M.D. Professor of Neurology and Roy E. and Merle Meyer Professor of Neuroscience, Mayo Clinic College of Medicine; Head of the Peripheral Neuropathy Research Center and Consultant in Neurology, Mayo Clinic, Rochester, Minnesota Pathologic Alterations of Nerves; Nerve Tests Expressed as Percentiles, Normal Deviates, and Composite Scores; Compound Action Potentials of Sural Nerve in Vitro in Peripheral Neuropathy; Quantitating Overall Neuropathic Symptoms, Impairments, and Outcomes; Quantitative Sensation Testing; Hereditary Motor and Sensory Neuropathies: An Overview of Clinical, Genetic, Electrophysiologic, and Pathologic Features; HMSN II (CMT2) and Miscellaneous Inherited System Atrophies of Nerve Axon: Clinical–Molecular Genetic Correlates; HSANs: Clinical Features, Pathologic Classification, and Molecular Genetics; Chronic Inflammatory Demyelinating Polyradiculoneuropathy; Neuropathy Associated with the Monoclonal Gammopathies; Nonmalignant Inflammatory Sensory Polyganglionopathy; Microvasculitis; Amyloidosis and Neuropathy

P. JAMES B. DYCK, M.D. Associate Professor of Neurology, Mayo Medical School; Consultant in Neurology, Mayo Clinic, Rochester, Minnesota Pathologic Alterations of Nerves; Nerve Tests Expressed as Percentiles, Normal Deviates, and Composite Scores; Quantitative Sensation Testing; Radiculoplexus Neuropathies: Diabetic and Nondiabetic Varieties; Microvasculitis

ANDREW G. ENGEL, M.D. McKnight-3M Professor of Neuroscience, Mayo Clinic College of Medicine, Rochester, Minnesota Diseases of the Neuromuscular Junction

Autonomic Systems to the Urinary Bladder and Sexual Organs

ANGELA DISPENZIERI, M.D. Assistant Professor of Medicine, Mayo Graduate School of Medicine; Consultant, Mayo Clinic, Rochester, Minnesota POEMS Syndrome (Osteosclerotic Myeloma)

MARY L. DOMBOVY, M.D., M.H.S.A. Clinical Associate Professor of Neurology and Physical Medicine and Rehabilitation, University of Rochester; Department Chair, Physical Medicine and Rehabilitation, Unity Health System, Rochester, New York Rehabilitation Management of Neuropathies

JANEAN ENGELSTAD, H.T. Histology Technician, Peripheral Neuropathy Research Center, Mayo Clinic, Rochester, Minnesota Pathologic Alterations of Nerves; Microvasculitis

MARK A. FERRANTE, M.D. Clinical Associate Professor of Neurology, Tulane University Medical Center, New Orleans, Louisiana; Director, EMG Laboratory, Bienville Orthopaedic Specialists, Biloxi, Mississippi Upper Limb Neuropathies: Long Thoracic (Nerve to the Serratus Anterior), Suprascapular, Axillary, Musculocutaneous, Radial, Ulnar, and Medial Antebrachial Cutaneous

viii

Contributing Authors

JOHN P. FRAHER, M.B., B.CH., B.A.O., F.R.C.S.ED., PH.D., D.SC., M.R.I.A. Professor of Anatomy, University College Cork, Cork, Ireland Microscopic Anatomy of the Peripheral Nervous System

MASON W. FREEMAN, M.D. Associate Professor of Medicine, Harvard Medical School; Physician, and Chief, Lipid Metabolism Unit, Massachusetts General Hospital, Boston, Massachusetts Tangier Disease and Neuropathy

ROY FREEMAN, M.B., CH.B. Associate Professor of Neurology, Harvard Medical School; Director, Autonomic and Peripheral Nerve Laboratory, Beth Israel Deaconess Medical Center, Boston, Massachusetts Treatment of Cardiovascular Autonomic Failure

THOMAS R. FRITSCHE, M.D., PH.D. Associate Director, The JONES Group/JMI Laboratories, North Liberty, Iowa

RALF GOLD, M.D. Consultant, and Head, Research Group for MS and Neuroimmunology, Julius-Maximilians-University Faculty of Medicine; Professor of Neurology, Department of Neurology, University Hospital Würzburg, Würzburg, Germany Introduction to Immune Reactions in the Peripheral Nervous System; Experimental Autoimmune Neuritis

IAN A. GRANT, M.D., F.R.C.P.(C) Assistant Professor, Division of Neurology, Dalhousie University; QEII Health Sciences Centre, Halifax, Nova Scotia, Canada Differential Diagnosis of Polyneuropathy; Cryptogenic Sensory Polyneuropathy

NORMAN A. GREGSON, PH.D. Reader in Neuroimmunology, Department of Clinical Neurosciences, Guy’s, King’s and St. Thomas’ School of Medicine, London, United Kingdom Peripheral Nerve Antigens

Parasitic Infections of the Peripheral Nervous System

ANNEKE GABREËLS-FESTEN, M.D., PH.D. Assistant Professor, University Nijmegen; Institute of Neurology, University Medical Center Nijmegen, Nijmegen, The Netherlands Autosomal Recessive Hereditary Motor and Sensory Neuropathies

ERNEST D. GARDNER, M.D.* Professor of Neurology, Orthopaedic Surgery, and Anatomy, University of California, Davis, School of Medicine, Davis, California Gross Anatomy of the Peripheral Nervous System

JOHN W. GRIFFIN, M.D. Professor and Director, Department of Neurology, Johns Hopkins University School of Medicine, Baltimore, Maryland The Control of Axonal Caliber; The Guillain-Barré Syndromes

MICHAEL J. GROVES, B.SC.(HONS.), PH.D. Non-Clinical Lecturer in Peripheral Nerve Pathology, Division of Neuropathology, Institute of Neurology, University College London, London, United Kingdom Pathology of Peripheral Neuron Cell Bodies

CATERINA GIANNINI, M.D. Assistant Professor, Department of Pathology, and Consultant, Mayo Clinic, Rochester, Minnesota Tumors and Tumor-like Conditions of Peripheral Nerve

DONALD H. GILDEN, M.D. Louise Baum Professor and Chairman, Department of Neurology, and Professor of Microbiology, University of Colorado Health Sciences Center; Professor and Chairman, Department of Neurology, University of Colorado Hospital, Denver, Colorado Herpesvirus Infection and Peripheral Neuropathy

HANS H. GOEBEL, M.D. Professor of Neuropathology, University of Mainz, Medical Center; Professor of Neuropathology, Johannes Gutenberg University, Mainz, Germany Lysosomal and Peroxisomal Disorders

*Deceased.

THOMAS M. HABERMANN, M.D. Professor of Medicine, Mayo Medical School; Consultant, Division of Hematology and Internal Medicine, Mayo Clinic Rochester, Rochester, Minnesota Peripheral Neuropathy Associated with Lymphoma, Leukemia, and Myeloproliferative Disorders

ANGELIKA F. HAHN, M.D., F.R.C.P.(C) Professor of Neurology, University of Western Ontario, and London Health Sciences Centre, London, Ontario, Canada Chronic Inflammatory Demyelinating Polyradiculoneuropathy

SUSAN HALL, B.SC., PH.D., D.SC. Head, Division of Anatomy, Cell and Human Biology, Guy’s, King’s and St. Thomas’ School of Biomedical Sciences, London, United Kingdom Mechanisms of Repair after Traumatic Injury

Contributing Authors

JOHN R. HALLIWILL, PH.D. Assistant Professor, Exercise and Movement Science, University of Oregon, Eugene, Oregon Sympathetic Nerves and Control of Blood Vessels to Human Limbs

MICHAEL G. HANNA, M.B.CH.B.(HONS.), M.D., F.R.C.P.(UK) Reader in Clinical Neurology, Department of Molecular Neuroscience, Institute of Neurology, University College London; Consultant Neurologist, Centre for Neuromuscular Disease, The National Hospital for Neurology and Neurosurgery, London, United Kingdom Peripheral Nerve Diseases Associated with Mitochondrial Respiratory Chain Dysfunction

A. E. HARDING, M.D., F.R.C.P.* Professor of Clinical Neurology, Institute of Neurology, London; Consultant Neurologist, The National Hospital for Neurology and Neurosurgery, London, United Kingdom Inherited Neuronal Atrophy and Degeneration Predominantly of Lower Motor Neurons

HANS-PETER HARTUNG, M.D. Professor and Chairman, Department of Neurology, Heinrich-Heine Universität Düsseldorf, Düsseldorf, Germany Introduction to Immune Reactions in the Peripheral Nervous System; Experimental Autoimmune Neuritis; Principles of Immunotherapy; Chronic Inflammatory Demyelinating Polyradiculoneuropathy

STEVEN HERSKOVITZ, M.D. Associate Professor of Clinical Neurology, Albert Einstein College of Medicine; Montefiore Medical Center, Bronx, New York Neuropathy Caused by Drugs

AHMET HÖKE, M.D., PH.D. Assistant Professor, Departments of Neurology and Neuroscience, Johns Hopkins University School of Medicine; Staff Neurologist and Pathologist, and Assistant Director of Neuromuscular Histopathology Laboratory, The Johns Hopkins Hospital, Baltimore, Maryland The Control of Axonal Caliber; Peripheral Neuropathies in Human Immunodeficiency Virus Infection

RICHARD A. C. HUGHES, M.D., F.R.C.P., F.MED.SCI. Professor and Head, Department of Clinical Neurosciences, Guy’s King’s and St. Thomas’ School of Medicine, King’s College London, University of London, London, United Kingdom Peripheral Nerve Antigens; Principles of Immunotherapy; Quantitating Overall Neuropathic Symptoms, Impairments, and Outcomes; Diseases of the Fifth Cranial Nerve

*Deceased.

ix

CLARE HUXLEY, PH.D. Reader, Imperial College London, London, United Kingdom Transgenic Models of Inherited Neuropathy

ROBERT R. JACOBSON, M.D., PH.D. Former Director, Leprosy Consultant, and Lecturer, National Hansen’s Disease Center, Baton Rouge, Louisiana Leprosy

ANN JACOBY, PH.D. Professor of Medical Sociology, Department of Primary Care, University of Liverpool, Liverpool, United Kingdom Health Outcomes and Quality of Life

KRISTJÁN R. JESSEN, M.SC., PH.D. Professor of Developmental Neurobiology, Department of Anatomy and Developmental Biology, University College London, London, United Kingdom Molecular Signaling in Schwann Cell Development

DAVID M. JOHNSON, B.S. Engineering Technical Specialist, Division of Engineering and Technology Services, Mayo Clinic, Rochester, Minnesota Quantitative Sensation Testing

H. ROYDEN JONES, JR., M.D. Jamie Ortiz-Patino Chair in Neurology, and Chairman Emeritus of the Division of Medical Specialties and Department of Neurology, Lahey Clinic, Burlington; Clinical Professor of Neurology, Harvard Medical School, Boston; Director of the EMG Laboratory, Children’s Hospital Boston, Boston, Massachusetts Peripheral Neuropathies in Infants and Children: Polyneuropathies, Mononeuropathies, Plexopathies, and Radiculopathies

MICHAEL J. JOYNER, M.D. Professor of Anesthesiology, Mayo Medical School; Professor of Anesthesiology, and Vice-Chair, Department of Physiology, Mayo Clinic, Rochester, Minnesota Sympathetic Nerves and Control of Blood Vessels to Human Limbs

BASHAR KATIRJI, M.D. Professor of Neurology, Case Western Reserve University; Chief, Neuromuscular Division, and Director, EMG Laboratory, University Hospitals of Cleveland, Cleveland, Ohio Mononeuropathies of the Lower Limb

KENTON R. KAUFMAN, PH.D. Associate Professor of Bioengineering, Mayo Medical School; Director, Motion Analysis Laboratory, and Consultant, Department of Orthopedic Surgery, Mayo Clinic, Rochester, Minnesota Quantitative Muscle Strength Assessment

x

Contributing Authors

JOHN J. KELLY, M.D. Professor and Chair, Department of Neurology, George Washington University, Washington, D.C. Amyloidosis and Neuropathy

WILLIAM R. KENNEDY, M.S., M.D. Professor of Neurology, Department of Neurology, University of Minnesota, Minneapolis, Minnesota Pathology and Quantitation of Cutaneous Innervation

MATTHEW C. KIERNAN, M.R.G.S.(HONS.), PH.D., F.R.A.C.P. Senior Lecturer in Neurology, University of New South Wales and Prince of Wales Medical Research Institute; Consultant Neurologist, Princes of Wales Hospital, Randwick, Sydney, New South Wales, Australia Nerve Excitability Measures: Biophysical Basis and Use in the Investigation of Peripheral Nerve Disease

BERND C. KIESEIER, M.D. Assistant Professor of Neurology, Heinrich-HeineUniversity, Düsseldorf, Germany Introduction to Immune Reactions in the Peripheral Nervous System; Experimental Autoimmune Neuritis

JUN KIMURA, M.D. Professor, Department of Neurology, University of Iowa Health Care, and University of Iowa, Iowa City, Iowa Nerve Conduction and Needle Electromyography

R. H. M. KING, PH.D., M.SC., F.R.C.PATH. Principal Research Fellow, Department of Clinical Neurosciences, Royal Free and University College Medical School, London, United Kingdom Microscopic Anatomy of the Peripheral Nervous System; Aging in the Peripheral Nervous System

CHRISTOPHER J. KLEIN, M.D. Assistant Professor of Neurology, Mayo Medical School; Consultant in Neurology, Mayo Clinic and Mayo Foundation, Rochester, Minnesota Nerve Tests Expressed as Percentiles, Normal Deviates, and Composite Scores; Quantitative Sensation Testing; Hereditary Motor and Sensory Neuropathies: An Overview of Clinical, Genetic, Electrophysiologic, and Pathologic Features; HMSN II (CMT2) and Miscellaneous Inherited System Atrophies of Nerve Axon: Clinical–Molecular Genetic Correlates; Hereditary Brachial Plexus Neuropathy; HSANs: Clinical Features, Pathologic Classification, and Molecular Genetics

KLEOPAS A. KLEOPA, M.D. Senior Consultant Neurologist, The Cyprus Institute of Neurology and Genetics, Nicosia, Cyprus X-linked Charcot-Marie-Tooth Disease

CHRISTOPHER J. KLINGELE, M.D. Assistant Professor of Obstetrics and Gynecology, Mayo Clinic College of Medicine; Consultant in Obstetrics and Gynecology, Mayo Clinic, Rochester, Minnesota Autonomic and Somatic Systems to the Anorectum and Pelvic Floor

DAVID L. KREULEN, PH.D. Professor, Departments of Physiology and Neurology, Michigan State University, East Lansing, Michigan Neurobiology of Autonomic Ganglia

ROBERT A. KYLE, M.D. Consultant, Division of Hematology and Internal Medicine, Mayo Clinic; Professor of Medicine and Laboratory Medicine, Mayo Medical School, Rochester, Minnesota Neuropathy Associated with the Monoclonal Gammopathies; Amyloidosis and Neuropathy; POEMS Syndrome (Osteosclerotic Myeloma)

CATHERINE LACROIX, M.D. Hospital Practitioner – Neuropathy, Centre Hospitalier de Bicetre, Le Kremlin, Bicetre, France Sarcoid Neuropathy

JOHN T. KISSEL, M.D. Professor and Vice-Chair, Department of Neurology, and Director, Division of Neuromuscular Disease, The Ohio State University, Columbus, Ohio Neuropathies with Systemic Vasculitis

TERRENCE D. LAGERLUND, M.D., PH.D. Associate Professor of Neurology, Mayo Clinic College of Medicine; Consultant in Neurology, Mayo Clinic, Rochester, Minnesota Nerve Blood Flow and Microenvironment

CAROLINE M. KLEIN, M.D., PH.D. Assistant Professor of Neurology, Department of Neurology, University of North Carolina School of Medicine, Chapel Hill, North Carolina Diseases of the Seventh Cranial Nerve; The Peripheral Nerve Involvement of Spinal Cord, Spinal Roots, and Meningeal Disease

EDWARD H. LAMBERT, M.D., PH.D.* Professor Emeritus, Mayo Clinic, Rochester, Minnesota Compound Action Potentials of Sural Nerve in Vitro in Peripheral Neuropathy

*Deceased.

Contributing Authors

SALLY N. LAWSON, PH.D. Professor, Department of Physiology, School of Medical Sciences, University of Bristol, University Walk, Bristol, Avon, United Kingdom The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons

xi

LINDA M. LUXON, B.SC.(HONS.), M.B., F.R.C.P. Professor of Audiological Medicine, University College London (Institute of Child Health), University of London; Consultant Physician in Neuro-otology, The National Hospital for Neurology and Neurosurgery, London, United Kingdom Diseases of the Eighth Cranial Nerve

JACQUELINE A. LEAVITT, M.D. Associate Professor of Ophthalmology, Mayo Medical School; Consultant, Department of Ophthalmology, Rochester Methodist Hospital, St. Mary’s Hospital, Rochester, Minnesota Diseases of the Third, Fourth, and Sixth Cranial Nerves

P. NIGEL LEIGH, PH.D., M.B.B.S., F.R.C.P.(UK) Professor of Clinical Neurology, and Head, Department of Neurology, Institute of Psychiatry; Director, King’s MND Care & Research Centre, King’s College Hospital, London, United Kingdom Sporadic Motor Neuron Degeneration; Inherited Motor Neuron Degeneration

J. G. LLEWELYN, B.SC.(HONS), M.D., F.R.C.P. Consultant Neurologist, Royal Gwent Hospital, Newport and Cardiff University School of Medicine, Cardiff, Wales; Honorary Consultant Neurologist, The National Hospital for Neurology and Neurosurgery, London, United Kingdom Diabetic Neuropathies

GLENN LOPATE, M.D. Assistant Professor of Neurology, Department of Neurology, Washington University; Assistant Professor of Neurology, Barnes-Jewish Hospital, St. Louis, Missouri Polyneuropathies and Antibodies to Nerve Components

PHILLIP A. LOW, M.D., F.R.A.C.P. Professor of Neurology, and Chairman, Division of Clinical Neurophysiology, Mayo Medical School; Consultant in Neurology, Mayo Clinic and Mayo Foundation, Rochester, Minnesota Oxidative Stress and Excitatory Neurotoxins in Neuropathy; Nerve Blood Flow and Microenvironment; Quantitation of Autonomic Impairment; Management of Autonomic Failure

JAMES R. LUPSKI, M.D., PH.D. Cullen Professor of Molecular and Human Genetics, and Professor of Pediatrics, Baylor College of Medicine; Consulting Medical Geneticist and Pediatrician, Texas Children’s Hospital and Ben Taub General Hospital; Consulting Geneticist, The Methodist Hospital, Texas Women’s Hospital, and St. Joseph’s Hospital, Baylor College of Medicine, Houston, Texas Hereditary Motor and Sensory Neuropathies: An Overview of Clinical, Genetic, Electrophysiologic, and Pathologic Features; Hereditary Motor and Sensory Neuropathies Involving Altered Dosage or Mutation of PMP22: The CMT1A Duplication and HNPP Deletion; Hereditary Motor and Sensory Neuropathy Related to Early Growth Response 2 (EGR2) Gene

RUDOLF MARTINI, PH.D. Professor and Head, Developmental Neurobiology, Department of Neurology, School of Medicine, University of Wuerzburg, Wuerzburg, Germany Myelination; Transgenic Models of Nerve Degeneration

CHRISTOPHER J. MATHIAS, D.PHIL., D.SC., F.R.C.P., F.MED.SCI. Professor of Neurovascular Medicine, Division of Neuroscience and Psychological Medicine, and Faculty of Medicine, Imperial College, London; Division of Clinical Neurology, Institute of Neurology, University College London; Consultant Physician and Director, Neurovascular Medicine Unit, St. Mary’s Hospital, London; Autonomic Unit, The National Hospital for Neurology and Neurosurgery, London, United Kingdom Quantitation of Autonomic Impairment; Diseases of the Ninth, Tenth, Eleventh, and Twelfth Cranial Nerves

JUSTIN C. MCARTHUR, M.B.B.S., M.P.H. Professor of Neurology and Epidemiology, Johns Hopkins University; Deputy Director, Department of Neurology, Johns Hopkins Hospital, Baltimore, Maryland Pathology and Quantitation of Cutaneous Innervation

ELIZABETH S. MCDONALD, M.D., PH.D. Radiology Resident, Mayo Clinic College of Medicine, Rochester, Minnesota Neurotrophic Factors in the Peripheral Nervous System

JAMES G. MCLEOD, M.B.B.S., D.PHIL., D.SC., F.R.A.C.P., F.R.C.P. Emeritus Professor, University of Sydney; Honorary Consulting Neurologist, Royal Prince Alfred Hospital, Camperdown, New South Wales, Australia Paraneoplastic Neuropathy

PHILIP G. MCMANIS, M.B.B.S., M.D., F.R.A.C.P.* Associate Professor, University of Sydney; Senior Staff Neurologist and Director of Clinical Neurophysiology, Royal North Shore Hospital, Sydney, New South Wales, Australia Nerve Blood Flow and Microenvironment

L. JOSEPH MELTON III, M.D. Professor of Epidemiology, Mayo Medical School, Mayo Clinic; Mayo Clinic, Rochester, Minnesota Epidemiologic Approaches to Peripheral Neuropathy

*Deceased.

xii

Contributing Authors

ALBEE MESSING, V.M.D., PH.D. Professor of Pathology and Neuroscience, School of Veterinary Medicine and Waisman Center, University of Wisconsin-Madison, Madison, Wisconsin Diphtheritic Polyneuropathy

VIRGINIA V. MICHELS, M.D. Professor of Medical Genetics, Mayo Clinic/Foundation, Rochester, Minnesota Mendelian and Mitochondrial Inheritance, Gene Identification, and Clinical Testing

RHONA MIRSKY, PH.D. Professor of Developmental Neurobiology, Department of Anatomy and Developmental Biology, University College London, London, United Kingdom Molecular Signaling in Schwann Cell Development

JOHN D. POLLARD, M.B.B.S., B.SC.(MED), PH.D., F.R.A.C.P. Bushell Professor of Neurology, University of Sydney; Head of Neurology, Royal Prince Alfred Hospital, Camperdown, New South Wales, Australia Principles of Immunotherapy; Neuropathy in Diseases of the Thyroid and Pituitary Glands

MICHAEL POLYDEFKIS, M.D. Assistant Professor of Neurology, Johns Hopkins University, Baltimore, Maryland Pathology and Quantitation of Cutaneous Innervation

SUDHA POTTUMARTHY, M.B.B.S., F.R.C.P.A. Senior Fellow, Department of Laboratory Medicine, and Acting Director, Antimicrobics Testing Laboratory, University of Washington School of Medicine, Seattle, Washington Parasitic Infections of the Peripheral Nervous System

PETER C. O’BRIEN, PH.D. Professor of Biostatistics, Division of Biostatistics, Mayo Clinic, Rochester, Minnesota Nerve Tests Expressed as Percentiles, Normal Deviates, and Composite Scores; Quantitating Overall Neuropathic Symptoms, Impairments, and Outcomes; Quantitative Sensation Testing

GRAHAM M. O’HANLON, PH.D. Research Assistant, Division of Clinical Neurosciences, University of Glasgow, Glasgow, Scotland, United Kingdom

MARY M. REILLY, M.D., F.R.C.P.I., F.R.C.P. Honorary Senior Lecturer, Department of Molecular Neurosciences, Institute of Neurology; Consultant Neurologist, The National Hospital for Neurology and Neurosurgery, London, United Kingdom Hereditary Amyloid Neuropathy

ANDREA ROBERTSON, PH.D. Career Development Fellow, Medical Research Council, London, United Kingdom Transgenic Models of Inherited Neuropathy

Peripheral Nerve Antigens

GILMORE N. O’NEILL, M.B., M.R.C.P.I. Instructor in Neurology and Assistant in Neurology, Massachusetts General Hospital, Boston, Massachusetts Tangier Disease and Neuropathy

GUSTAVO C. ROMAN, M.D., F.A.C.P., F.A.A.N., F.R.S.M.(LOND.) Professor of Medicine/Neurology, Division of Neurology, Department of Medicine, University of Texas Health Science Center at San Antonio; Neurologist, University Hospital and VA Hospital, San Antonio, Texas Tropical Myeloneuropathies

DAVID J. PATERSON, M.SC., M.A., D.PHIL. Professor of Physiology, University of Oxford, Director of Pre-clinical Studies, Fellow of Merton College, Oxford, University of Oxford, Oxford, United Kingdom Neural Control of Cardiac Function

ALAN PESTRONK, M.D. Professor of Neurology and Pathology, Washington University Medical School; Professor, Barnes-Jewish Hospitals, St. Louis, Missouri Polyneuropathies and Antibodies to Nerve Components

DAVID PLEASURE, M.D. Professor, Neurology and Pediatrics, University of Pennsylvania School of Medicine; Director, Joseph Stokes Jr. Research Institute, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania Diphtheritic Polyneuropathy

MICHAEL C. ROWBOTHAM, M.D. Professor of Clinical Neurology and Anesthesia, University of California, San Francisco; Director, UCSF Pain Clinical Research Center; Senior Attending Neurologist, UCSF-Mount Zion Pain Management Center, San Francisco, California Mechanisms and Pharmacologic Management of Neuropathic Pain

MONIQUE M. RYAN, M.B.B.S., M. MED., F.R.A.C.P. Paediatric Neurologist, Institute for Neuromuscular Research, and Senior Lecturer, Discipline of Paediatrics and Child Health, University of Sydney, The Children’s Hospital at Westmead, Westmead, New South Wales, Australia Peripheral Neuropathies in Infants and Children: Polyneuropathies, Mononeuropathies, Plexopathies, and Radiculopathies

Contributing Authors

MARTIN RYDMARK, M.D., PH.D. Associate Professor, Senior Lecturer, Mednet – Medical Informatics & Computer Assisted Education, Institute of Anatomy and Cell Biology, The Sahlgrenska Academy at Göteborg University, Göteborg, Sweden

xiii

MARTIN SCHMELZ, M.D., PH.D. Professor, Department of Anesthesiology and Intensive Care Medicine, Faculty of Clinical Medicine Mannheim, University of Heidelberg, Mannheim, Germany Single-Unit Recordings of Afferent Human Peripheral Nerves by Microneurography

Microscopic Anatomy of the Peripheral Nervous System

THOMAS D. SABIN, M.D. Professor and Vice Chair of Neurology, Tufts University School of Medicine; Acting Chief, Department of Neurology, Tufts-New England Medical Center, Boston, Massachusetts Leprosy

GÉRARD SAID, M.D. Professor of Neurology, University Paris-Sud, Paris; Chairman, Service of Neurology, Hopital de Bicetre, Le Kremlin Bicetre, Val De Marne, France Lyme Disease; Sarcoid Neuropathy

DAVID S. SAPERSTEIN, M.D. Assistant Professor of Neurology and Pathology, and Chief, Neuromuscular Disease Service, University of Kansas Medical Center, Kansas City, Kansas Polyneuropathy Caused by Nutritional and Vitamin Deficiency

FRANCESCO SCARAVILLI, M.D., PH.D., D.SC., F.R.C.PATH. Professor of Neuropathology, Honorary Consultant Neuropathologist, and Head of Division of Neuropathology, Institute of Neurology, The National Hospital for Neurology and Neurosurgery, University College London, London,United Kingdom Pathology of Peripheral Neuron Cell Bodies

JON J. A. SCOTT, B.SC., PH.D. Senior Lecturer in Physiology, Department of Pre-Clinical Sciences, University of Leicester, Leicester, United Kingdom The Golgi Tendon Organ

KAZIM SHEIKH, M.D. Assistant Professor, Department of Neurology, Johns Hopkins University School of Medicine, Baltimore, Maryland The Guillain-Barré Syndromes

JOHN T. SHEPHERD, M.D. Professor Emeritus, Mayo Clinic, Rochester, Minnesota Sympathetic Nerves and Control of Blood Vessels to Human Limbs

MICHAEL E. SHY, M.D. Professor of Neurology, Professor of Molecular Medicine and Genetics, Director of the Inherited Neuropathy Clinic, and Co-Director of the Neuromuscular Program, Wayne State University School of Medicine, Detroit, Michigan Hereditary Motor and Sensory Neuropathies: An Overview of Clinical, Genetic, Electrophysiologic, and Pathologic Features; Hereditary Motor and Sensory Neuropathies Related to MPZ (P0) Mutations

WOLFGANG SINGER, M.D. Resident, Neurology, Mayo Clinic, Rochester, Minnesota Management of Autonomic Failure

HERBERT H. SCHAUMBURG, M.D. Edwin S. Lowe Professor and Chairman of Neurology, Albert Einstein College of Medicine; Chairman of Neurology, Montefiore Medical Center, Bronx, New York Human Toxic Neuropathy Caused by Industrial Agents; Neuropathy Caused by Drugs

STEVEN S. SCHERER, M.D., PH.D. William N. Kelley Professor of Neurology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania X-linked Charcot-Marie-Tooth Disease

RAPHAEL SCHIFFMANN, M.D. Section Chief, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland Fabry’s Disease

BENN E. SMITH, M.D. Assistant Professor of Neurology, Mayo Clinic College of Medicine, Rochester, Minnesota Nonmalignant Inflammatory Sensory Polyganglionopathy

ERIC J. SORENSON, M.D. Assistant Professor, Mayo Foundation Hospitals, Mayo Medical School; Head of Section, Neuromuscular Diseases, and Director of Motor Neuron Program, Mayo Clinic, Rochester, Minnesota Transgenic Animal Models of Amyotrophic Lateral Sclerosis

JUDITH M. SPIES, M.B.B.S., PH.D., F.R.A.C.P. Senior Lecturer, University of Sydney; Staff Specialist Neurologist, Royal Prince Alfred Hospital, Camperdown, New South Wales, Australia Paraneoplastic Neuropathy

xiv

Contributing Authors

ERIK V. STÅLBERG, M.D., PH.D., F.R.C.P. Professor Emeritus, Department of Neurosciences, Clinical Neurophysiology, Uppsala University; Professor Emeritus, Department of Clinical Neurophysiology, Uppsala University Hospital, Uppsala, Sweden Single-Fiber Electromyography and Other Electrophysiologic Techniques for the Study of the Motor Unit

J. CLARKE STEVENS, M.D. Professor of Neurology, Mayo Medical School, Rochester, Minnesota Median Neuropathy

GUIDO STOLL, M.D. Consultant, Julius-Maximilians-University Faculty of Medicine; Professor of Neurology, Department of Neurology, University Hospital Würzburg, Würzburg, Germany Introduction to Immune Reactions in the Peripheral Nervous System; Experimental Autoimmune Neuritis

GUILLERMO A. SUAREZ, M.D. Associate Professor of Neurology, Mayo Clinic College of Medicine; Consultant in Neurology, Mayo Clinic, Rochester, Minnesota Immune Brachial Plexus Neuropathy; POEMS Syndrome (Osteosclerotic Myeloma)

UELI SUTER, PH.D. Professor, Institute of Cell Biology, Swiss Federal Institute of Technology, ETH Zu¯rich, ETH Ho¯nggerberg, Zu¯rich, Switzerland Myelination

THOMAS R. SWIFT, M.D. Professor Emeritus of Neurology, Medical College of Georgia, Augusta, Georgia Leprosy

BRUCE V. TAYLOR, M.B., M.D., F.R.A.C.P. Clinical Senior Lecturer, University of Tasmania; Head, Department of Neurology, Royal Hobart Hospital, Hobart, Tasmania, Australia Multifocal Motor Neuropathy and Conduction Block

AYALEW TEFFERI, M.D. Professor of Medicine and Hematology, Mayo Medical School and Mayo Clinic, Rochester, Minnesota Peripheral Neuropathy Associated with Lymphoma, Leukemia, and Myeloproliferative Disorders

STEPHEN N. THIBODEAU, PH.D. Professor of Laboratory Medicine, Mayo Clinic/Foundation, Rochester, Minnesota Mendelian and Mitochondrial Inheritance, Gene Identification, and Clinical Testing

P. K. THOMAS, C.B.E., M.D., D.SC., F.R.C.P., F.R.C.(PATH.) Emeritus Professor of Neurology, University College London School of Medicine and The National Hospital for Neurology and Neurosurgery, London, United Kingdom Clinical Patterns of Peripheral Neuropathy; Diseases of the Ninth, Tenth, Eleventh, and Twelfth Cranial Nerves; Autosomal Recessive Hereditary Motor and Sensory Neuropathies; Lysosomal and Peroxisomal Disorders; Diabetic Neuropathies

PHILIP D. THOMPSON, M.B.B.S., PH.D., F.R.A.C.P. Professor of Neurology, University of Adelaide; Head, Department of Neurology, Royal Adelaide Hospital, Adelaide, South Australia Clinical Patterns of Peripheral Neuropathy

ERIK C. THORLAND, PH.D. Fellow, Clinical Molecular Genetics, Mayo Clinic/Foundation, Rochester, Minnesota Mendelian and Mitochondrial Inheritance, Gene Identification, and Clinical Testing

D. R. TOMLINSON, PH.D., D.SC. Professor of Neuropharmacology, Family of Life Sciences, The University of Manchester, Manchester, England Diabetic Neuropathies

ERIK TOREBJÖRK, M.D., PH.D. Professor, Department of Clinical Neurophysiology, University of Uppsala, Uppsala, Sweden Single-Unit Recordings of Afferent Human Peripheral Nerves by Microneurography

KLAUS V. TOYKA, M.D., F.R.C.P. Professor of Neurology, Julius-Maximilians-University Faculty of Medicine; Neurologist-In-Chief, Department of Neurology, University Hospital Würzburg, Würzburg, Germany Introduction to Immune Reactions in the Peripheral Nervous System; Experimental Autoimmune Neuritis

JO Zˇ E V. TRONTELJ, M.D., PH.D. Professor, Institute of Clinical Neurophysiology, University Medical Center, Ljubljana, Slovenia Single Fiber Electromyography and Other Electrophysiologic Techniques for the Study of the Motor Unit

KENNETH L. TYLER, M.D. Reuler-Lewin Family Professor of Neurology and Professor of Medicine, Microbiology and Immunology, University of Colorado Health Sciences Center; Chief, Neurology Service, Denver Veterans Affairs Medical Center, Denver, Colorado Herpesvirus Infection and Peripheral Neuropathy

Contributing Authors

B. ULFHAKE, PH.D. Professor, Department of Neuroscience, Karolinska Institutet, Stockholm, Sweden Aging in the Peripheral Nervous System

PAUL M. VANHOUTTE, M.D., PH.D. Distinguished Visiting Professor, Department of Pharmacology, University of Hong Kong, Hong Kong Sympathetic Nerves and Control of Blood Vessels to Human Limbs

ANNABEL K. WANG, M.D. Assistant Professor, Department of Neurology, Mount Sinai School of Medicine; Assistant Attending Physician, Mount Sinai Hospital, New York, New York The Peripheral Nerve Involvement of Spinal Cord, Spinal Roots, and Meningeal Disease

LAURA E. WARNER, PH.D. Research Scientist, University of Washington, Seattle, Washington Hereditary Motor and Sensory Neuropathy Related to Early Growth Response 2 (EGR2) Gene

HENRY DEF. WEBSTER, M.D. Emeritus Scientist, NINDS, National Institutes of Health, Bethesda, Maryland Introduction

ANANDA WEERASURIYA, M.PHIL., PH.D. Professor of Neuroscience and Physiology, Mercer University School of Medicine, Macon, Georgia Blood-Nerve Interface and Endoneurial Homeostasis

GWEN WENDELSCHAFER-CRABB, M.S. Senior Scientist, Department of Neurology, University of Minnesota, Minneapolis, Minnesota Pathology and Quantitation of Cutaneous Innervation

EELCO F. M. WIJDICKS, M.D. Professor of Neurology, and Chair, Division of Critical Care Neurology, Department of Neurology, Mayo Clinic, Rochester, Minnesota Management of Patients with Acute Neuromuscular Disease in the Intensive Care Unit

ASA J. WILBOURN, M.D. Clinical Professor of Neurology, Case Western Reserve University; Director, EMG Laboratory, Department of Neurology, Cleveland Clinic, Cleveland, Ohio Brachial Plexus Lesions; Upper Limb Neuropathies: Long Thoracic (Nerve to the Serratus Anterior), Suprascapular, Axillary, Musculocutaneous, Radial, Ulnar, and Medial Antebrachial Cutaneous; Mononeuropathies of the Lower Limb

xv

HUGH J. WILLISON, M.B.B.S., PH.D., F.R.C.P. Professor of Neurology, Division of Clinical Neurosciences, University of Glasgow; Honorary Consultant Neurologist, Institute of Neurological Sciences, Southern General Hospital, Glasgow, Scotland, United Kingdom Peripheral Nerve Antigens; Multifocal Motor Neuropathy and Conduction Block

ANTHONY J. WINDEBANK, M.D. Professor of Neurology, Mayo Graduate School and Mayo Medical School; Consultant in Neurology, Mayo Clinic and Mayo Foundation; Dean, Mayo Medical School, Mayo Clinic College of Medicine, Rochester, Minnesota Neurotrophic Factors in the Peripheral Nervous System; Hereditary Brachial Plexus Neuropathy; Porphyric Neuropathy; Nonmalignant Inflammatory Sensory Polyganglionopathy; Metal Neuropathy

HARALD WITTE Graduate Student, Max-Planck-Institute of Neurobiology, Munich, Germany Guidance of Axons to Targets in Development and in Disease

JACKIE D. WOOD, M.S., PH.D. Professor of Physiology and Cell Biology and Internal Medicine, Ohio State University College of Medicine and Public Health, Columbus, Ohio Neurobiology of the Enteric Nervous System

BRIAN R. YOUNGE, M.D. Associate Professor of Ophthalmology, Mayo Medical School, Mayo Graduate School; Consultant in Ophthalmology, Mayo Clinic, Rochester, Minnesota Diseases of the Third, Fourth, and Sixth Cranial Nerves

DOUGLAS W. ZOCHODNE, M.D., F.R.C.P.(C.) Professor, University of Calgary; Consultant Neurologist, Calgary Health Region, Foothills Hospital, Calgary, Alberta, Canada Neuropathies Associated with Renal Failure, Hepatic Disorders, Chronic Respiratory Disease, and Critical Illness

Preface

More than a decade has passed since the third edition of Peripheral Neuropathy. The editors have found it necessary to create an essentially new textbook in order to encompass the rapid, exciting advances in neurobiology, molecular genetics, chemical and cellular pathology, and treatment. With the exception of chapters on gross anatomy, the compound action potential of the sural nerve in vitro, and the chapter on progressive muscular atrophy, most chapters are completely rewritten to produce the broadest and most up-to-date reviews of the neurobiology of the nerve and its diseases. The editors hope that the readers will find the new edition comprehensive and informative. The only credit the editors take is for their choice of authors, who are authorities in their fields. We are deeply grateful for their hard work and new insights. As in previous editions, we honor a pioneer and friend who has made major contributions to our understanding of nerve biology and disease. In the first edition, Wilhelm Krücke, of Frankfurt am Main, was chosen as he was considered to be the Dean of Neuropathology of peripheral nerve. It was not surprising that in his Introduction he mentioned the contributions of T. Schwann (description of the myelin cell named after him), A. Waller and R. Cajal (key contributors to cellular events of nerve fiber degeneration), L. Ranvier (the node that bears his name), A. Gombault (segmental demyelination), R. Virchow (inflammation), and P. Weis (axonal flow). For the second edition, we honored Fritz Buchthal, who along with E. H. Lambert, pioneered clinical nerve conduction and clinical electromyography. It is of some interest that these new techniques have become so informative that the earlier use of direct nerve and muscle stimulation with galvanic and faradic currents (Duchenne, and later Erb) are hardly mentioned in modern textbooks. In this fourth edition, Kiernan and colleagues (Chapter 5) argue that these earlier approaches (or modifications) might still be used in specific indications. J. Z. Young, the special mentor of one of us (PKT), wrote the Introduction to the third edition. Young

(with P. K. Thomas) had returned to the old technique of teasing peripheral nerve fibers—a methodology that remains useful and is extensively described in Chapter 32. In the present edition, we honor Henry deF. Webster. Harry characterized the ultrastructural features of mammalian nerve both in development and after maturation, and provided early studies of pathologic alteration in disease. His electron micrographs in Peters, Palay, and deF. Webster, “The Fine Structure of the Nervous System: The Neurons and Supporting Cells” (W. B. Saunders, 1976), remains as an example of how it should be done. Quite unfairly to authors whose chapters are not mentioned here, we list some of the special coverage. Sally Lawson in her chapter on dorsal root ganglion (spinal ganglion) neurons, reviews sensory properties, electrophysiology, differences between non-nociceptor and nociceptor somas, immunocytochemical properties, trophic factors, cytokines, receptors, ion channels, membrane properties, genes, and injury reactions—a real tour de force! Jackie Wood makes the case for a third nervous system (in addition to the central and peripheral nervous systems), the enteric system. An insight into the system is crucial for the understanding of diseases affecting the gut. The chapter by A. G. Engel provides a concise review of the diseases of the neuromuscular junction. His electron micrographs of pathologic alterations are unmatched. The chapters on quantitating neuropathic impairment, disability, symptoms, outcomes, and quality of life as it relates to neuropathy are more focused on neuropathy and more comprehensive than discussed elsewhere. Ian Grant’s chapter on differential diagnosis of neuropathy is an important read for physicians who want to improve their diagnostic ability. The sections on inherited motor, motor and sensory, and sensory and autonomic neuropathies have undergone major changes. Comprehensive and detailed reviews are focused first on clinical features (inheritance and natural history pattern, population, level and pathologic type of neuron involvement, electrophysiologic and pathologic characteristics) followed by molecular xvii

xviii

Preface

genetic causes. P. James B. Dyck emphasizes the need to recognize that diabetic neuropathy is not a single disorder. Discussing diabetic radiculoplexus neuropathies, he stresses underlying immune mechanisms and the possible benefit of immune-modulating therapy. The chapter on necrotizing vasculitis of the peripheral nervous system by Michael Collins and John Kissel is a definitive and beautifully illustrated contribution. In the pathology chapter, we include the new approaches used in identifying not only the focal or multifocal abnormality of proximal nerves, but also the underlying cause. Many chapters deal with an increasing emphasis on diagnosis and management of diseases of the autonomic nervous system. The strong support of P. A. Low, MD is acknowledged.

The editors are grateful to the editors at Elsevier (Jennifer Ehlers) and at Bermedica Production, Ltd. (Berta Steiner). PJD is grateful to Mary Lou Hunziker for her enormous help. I (PJD) hereby also acknowledges the heroic help of Nok (Sam) Ponsford, M.D., who stepped in to help when one of us, her husband (PKT), developed a stroke, and other health complications, during the editing of this book. Colleagues from The Royal Free Hospital also helped. NOK and PK—thank you. One of us (PJD) is grateful for the forbearance of Isabelle, Ernest Carl, Fred Howard, P. James B., and M. Katharine E. and to grandchildren (Sophie, Jacob, Chloe, and Abbie), and to Marian and Scott. PETER J. DYCK P. K. THOMAS

1 Introduction HENRY DEF. WEBSTER

Early Electron Microscopic Studies of Peripheral Nerve Segmental Demyelination and Remyelination Wallerian Degeneration and Regeneration

Development of Peripheral Nerve Fibers Early Axon–Schwann Cell Interactions Early Schwann Cell–Axon Relationships in Nerves

EARLY ELECTRON MICROSCOPIC STUDIES OF PERIPHERAL NERVE About 50 years ago, the first electron microscopic observations of peripheral nerve were reported. The limited resolution of the light microscope had left some important questions unanswered and had only partially answered others. What were the relationships of axons, their myelin sheaths and Schwann cells? Did myelin have a layered structure and were there discontinuities in the myelin sheath at Schmidt-Lanterman clefts? How was myelin formed and how did myelinated and unmyelinated fibers develop? What structural changes did trauma or disease produce in peripheral nerves? Finally, what cellular changes occurred during nerve regeneration and how did they differ from those seen during development? The electron microscope permitted investigators to visualize relationships of myelin, cell membranes, organelles, fibrils and other cellular inclusions at substantially higher resolution. This was an important advantage and led to a substantial increase in the use of this instrument. Increased use of the electron microscope led investigators to describe the fine structure of the nervous system,27 to quantitate nerve biopsy findings,14,24,36 and to create better methods for preserving peripheral nervous tissue.9,45,47,50,52 Subsequent advances in methodology included immunocytochemical techniques for localizing myelin constituents16,25,38 and an in situ hybridization method to study the distribution of myelin protein messenger RNAs (mRNAs) in sections prepared for electron microscopy.48

Formation and Growth of Myelin Sheaths Distribution of the Protein Constituents of Myelin Concluding Comment

Early studies using the electron microscope showed that axons in the peripheral nervous system were ensheathed by Schwann cells and that myelin was located in the cytoplasm of Schwann cells.13,26,30 Myelin segments terminated at nodes of Ranvier in a series of loops32,41 and had a compact layered structure except in Schmidt-Lanterman clefts.31 There, myelin lamellae were separated by helically arranged strips of cytoplasm. The layered pattern of compact myelin was evident after fixation and was thought to be consistent with molecular models of myelin derived from observations utilizing polarized light and x-ray diffraction.11,23,28,33 Early in development, Schwann cells surrounded bundles of axons and segregated larger ones for subsequent myelination.26 Then myelin formation occurred by spiral growth of the mesaxon, an extension of the Schwann cell surface membrane.13,30 Further myelin membrane growth produced a compact spiral of myelin membrane with inner and outer mesaxons connecting it to the Schwann cell’s periaxonal and external surface membranes. The above findings were important and proved to be useful for investigators who began studying peripheral nervous system pathology with the electron microscope.

SEGMENTAL DEMYELINATION AND REMYELINATION Segmental demyelination, a sequence of changes characterized by myelin damage and relative sparing of axons, is a cardinal feature of human diphtheritic neuritis12 and its 3

4

Introduction

model, experimental diphtheritic neuritis in guinea pigs.43 Because of the importance of segmental demyelination in human demyelinating diseases, we selected this pathologic process for our first electron microscopic study.52 Early observations showed that the preservation of large myelin sheaths had to be improved so that lesions could be distinguished from preparative artifacts. Then early focal lesions were identified paranodally and elsewhere along the length of myelin segments without evidence of inflammation.52 Later breakdown of myelin into ovoids occurred within Schwann cells; nearby macrophages also contained myelin remnants. Severely demyelinated nerves also contained some Schwann cells that were actively regenerating myelin sheaths. Interestingly, some of these sheaths had major variations in contour along their lengths52 that resembled those seen paranodally in normal adult fibers.51 Since published models of myelination did not explain how these myelin irregularities were formed, their development and distribution were investigated in a later study of myelin formation.45 In acute idiopathic neuritis (Guillain-Barré syndrome) and its experimental model, experimental allergic neuritis (EAN42), segmental demyelination is preceded by a perivascular invasion of lymphocytes and mononuclear cells. In a serial section study of early EAN lesions, Åström found that lymphocytes migrated through endothelial cells of rat sciatic nerve venules, a process he considered to be a form of emperipolesis.2,3 Demyelination then began within the perivenular inflammatory infiltrates and involved contact of mononuclear cells with myelin and subsequent stripping of myelin remnants from axon surfaces by mononuclear cells.18

WALLERIAN DEGENERATION AND REGENERATION Nerve transection or crush produces distal degeneration of axons and their myelin sheaths, a process called wallerian degeneration. When the early stages of this process were examined, there were transient accumulations of mitochondria in paranodal regions of large myelinated fibers.44 They occurred also in some unmyelinated axons. Comparisons of proximal and distal paranodal regions showed that most accumulations were in distal paranodal axoplasm. Migration of mitochondria to paranodal regions and their proliferation were considered as possible explanations of these transient accumulations. A later study by others demonstrated that both anterograde and retrograde transport of axonal mitochondria and other organelles occurred during wallerian degeneration.40 After nerve transection, axons and their myelin sheaths regenerate. This process begins in the distal end of the proximal stump, the site we selected for our studies of nerve regeneration. When axons form growth cones and begin regenerating, Schwann cells divide rapidly. They

also synthesize laminin, a major constituent of their basement membranes. Laminin stimulates cell division17 and has an important role in the extension of neurites during nerve development and regeneration.4 After transecting sciatic nerves, we prepared supernatants of proximal and distal stumps and compared their effects on Schwann cell proliferation and laminin production in vitro.53 Lysates of cultures treated with 24-hour proximal supernatants contained significantly higher levels of laminin than those prepared from other supernatant intervals, from distal segments, or from control nerves. Twenty-four–hour and 2-day proximal supernatants also increased proliferation of cultured Schwann cells significantly. These findings suggested that proximal stump axons were possible sources of substances that were responsible for these effects.53 Since the addition of vasoactive intestinal peptide (VIP) to cultured Schwann cells also increased laminin synthesis,54 Zhang and collaborators then tested the effects of locally administered VIP on regenerative responses of transected nerve fibers. They found that injections of VIP into transection sites twice daily for 2 weeks accelerated ensheathment and myelination of regenerating axons.55

DEVELOPMENT OF PERIPHERAL NERVE FIBERS As noted above, the contour variations we described in normal adult51 and regenerating52 myelin sheaths were not explained by contemporary models of myelination. Our interest in their formation and distribution led to a series of later studies of Schwann cell and nerve fiber development that are described and illustrated in an earlier edition of this text.46

Early Axon–Schwann Cell Interactions In the developing peripheral nervous system, outgrowth of neurites is soon followed by the appearance of Schwann cells that originate from the neural crest.21 Speidel described their early association with axons in the transparent tail fins of tadpoles, along with important parameters of myelin formation.34 Billings-Gagliardi extended these findings by correlating in vivo observations of early interactions of Schwann cells and axons in Xenopus tadpoles with their appearance in electron micrographs.5,6 Schwann cells, which were seen moving freely between axons, did not have a basal lamina and did not resemble either fibroblasts or macrophages. These Schwann cells were ovoid in shape and had several long processes that ended in blunt expansions.6 When first interacting with axons, Schwann cells moved “inchworm” style along axonal surfaces. Movement was sporadic with alternating periods of rapid (5 ␮m/min) and no movement.5 Later, when Schwann cells spread along axons and ensheathed them,

5

Introduction

they became more spindle shaped and acquired a basal lamina.6 Subsequent tissue culture studies clearly demonstrated that axon–Schwann cell contact was required for basal lamina formation.7,8 The in vivo and electron microscopic observations of tadpole nerve fibers by BillingsGagliardi5,6 also showed that relationships between ensheathing Schwann cells and axons were more complex than those described by Speidel.34 Instead, they resembled those found in developing mouse and rat nerves.

By the end of prophase, the Schwann cell was spindle shaped and remained so through metaphase and much of anaphase. The axis of mitosis was parallel to the long axis of the cell. Radial processes that surrounded axons reappeared in telophase and extended along processes of neighboring interphase Schwann cells or the basal lamina that enclosed the family. These shape changes that occur during mitosis were thought to have a role in increasing the number of axons contacted by Schwann cells early in nerve development.22

Early Schwann Cell–Axon Relationships in Nerves

Formation and Growth of Myelin Sheaths

In developing mammalian nerves, the relationships between Schwann cells and axons change rapidly. During this complex sequence of events, axons are sorted into a population of fibers that becomes myelinated and another population that remains unmyelinated. Some geometric and quantitative aspects of these changing relationships were examined in skip serial sections of a fiber population found at the margin of the rat sciatic nerve’s posterior tibial fascicle.22,49 At birth, none of this marginal bundle’s axons was myelinated and almost all of the bundle’s transverse area was occupied by “Schwann cell families,” a term used to describe all of the axons and Schwann cell processes found within a common basal lamina. Big axon bundles were located in the center of each family, and larger axons were found more commonly at the edge of a bundle, segregated in a separate furrow, or in a 1⬊1 relationship with a Schwann cell located on the family’s outer surface. This concentric arrangement, which persisted during axon bundle subdivision and the onset of myelination, suggested that radial sorting of axons destined to be myelinated occurred in sheaths formed by longitudinal columns of Schwann cell families. The sorting sequence included initial contact with a Schwann cell process, segregation in a separate furrow of a family sheath, Schwann cell division, and establishment of a 1⬊1 relationship with one of the daughter Schwann cells, which then became isolated from the family sheath before myelination began.49 Counts during the week after birth also showed that Schwann cell families did the sorting in this population of fibers and permitted us to estimate rates for axon bundle subdivision, segregation of larger axons in separate furrows, and establishment of 1⬊1 relationships. Because approximately half of the Schwann cells that surrounded axon bundles also enveloped larger axons in a separate furrow, this process of segregation was thought to be an essential intermediate step in the establishment of the 1:1 relationship that preceded myelination.49 In order to understand the geometry and dynamics of this sorting process, we also examined the relationships of dividing Schwann cells.22 In newborn rat sciatic nerves, virtually all of the mitotic Schwann cells were located in the family sheaths described above. As mitosis began, the radial extent of the processes that surrounded axons decreased.

The electron microscopic observations and hypothesis of Geren clearly established the basic morphologic parameters of peripheral myelination.13 The mesaxon, an extension of the Schwann cell surface membrane, grows and forms a spiral sheet around the axon. Further growth and apposition of the spiral layers occur as the myelin sheath matures. Her findings, which were confirmed by many others, did not explain two important features of myelin sheath development. First, the contour of the myelin spiral is not uniform along the internode; complex variations occur.51 Second, after the compact sheath is formed, its internal circumference increases to accommodate the growing axon. Because little was known about the geometry and dimensions of the myelin spiral as it formed and grew in a Schwann cell, these parameters were studied in skip serial sections of the marginal bundle of the rat sciatic nerve’s posterior tibial fascicle.45 All of this bundle’s fibers were unmyelinated at birth, so the onset and duration of myelination were easily established. At appropriate intervals, approximate dimensions of the bundle and its largest fibers were measured and calculated at the same relative level in littermates’ nerves. These data showed that the myelin membrane’s area and transverse length increased exponentially with time; the growth rate increased rapidly during the formation of the first four to six spiral layers and remained relatively constant during the subsequent rapid enlargement of the compact sheath.45 Later more precise measurements in the developing sixth cranial nerve also showed that the myelin membrane grows exponentially during myelin formation.15 Along myelin internodes composed of two to six spiral turns, there were many variations in the number of lamellae and their contour. Near the mesaxon’s origin, longitudinal strips of cytoplasm separated the myelin layers. Thicker sheaths were larger in circumference, more circular in transverse sections and more uniform at different levels. Separation of lamellae by cytoplasm became discontinuous and generally occurred at Schmidt-Lanterman clefts. Variations in sheath contour similar to those described earlier51 became less frequent and usually were located in paranodal regions.45 These findings showed that myelin

6

Introduction

remodeling and redistribution of Schwann cell cytoplasm both occur during peripheral nerve myelination and that they are important features of this dynamic process.45 How does a Schwann cell in developing nerve rapidly increase the internal circumference, the number of compact layers, and the length of its myelin segment? There is general agreement that this is a difficult unsolved problem. Several mechanisms have been suggested and are being investigated.

and diseased peripheral nervous system has expanded rapidly. The electron microscope is still widely used, especially in correlative studies. Three recent examples concern the myelinated fiber’s molecular architecture1 and the role of an integrin receptor in myelination.10,29 These reports and the contents of this edition suggest a bright future for research that will continue to increase our understanding of the etiology, pathogenesis and treatment of peripheral nerve diseases.

Distribution of the Protein Constituents of Myelin

REFERENCES

In peripheral myelin, the major protein constituent is a glycoprotein called P0. To study its localization in Schwann cells, Trapp and his collaborators immunostained semithin Epon sections with P0 antiserum and traced the distribution of cytoplasmic staining on electron micrographs of the same Schwann cells in adjacent thin sections.38 They showed that very thin myelin sheaths in newborn trigeminal nerve were detected more easily in semithin sections stained with P0 antiserum than with other cellular stains. In developing and adult myelin-forming Schwann cells, P0 staining also was present in cytoplasmic regions occupied by Golgi profiles.38 Later P0 gene expression was studied by using light and electron microscopic in situ hybridization techniques to explore the distribution of P0 mRNA during the postnatal development of the trigeminal ganglion.19,20,48 Of particular interest was the transient detection of P0 mRNA in developing Schwann cells that surround sensory neurons, even though these cells never form myelin.19 There are two basic proteins in nerve myelin called P1 and P2 with molecular weights of 18,500 and 13,500 respectively. P1 was localized in dense line regions of compact myelin25 and, like P0, was readily detected in myelin sheaths of newborn rats.16,39 P2 is present in smaller amounts and was not detected until day 4 in a quantitative densitometric study of developing rat sixth nerves.16 However, since it was detected in 85% of myelin sheaths at day 20, the authors concluded that the sensitivity of the immunostaining method was limited and that all myelinforming Schwann cells probably expressed P2.16 Myelin-associated glycoprotein, a member of the immunoglobulin super family, is present in developing peripheral myelin.35,46 In mature sheaths, it is found in Schmidt-Lanterman clefts, paranodal areas, and periaxonal regions of myelin and Schwann cells.37

CONCLUDING COMMENT During the 30 odd years since the first edition of this reference was planned, the literature devoted to the structure, function, biochemistry and genetics of the normal

1. Arroyo, E. J., and Scherer, S. S.: On the molecular architecture of myelinated fibers. Histochem. Cell Biol. 113:1, 2000. 2. Åström, K. E.: Migration of lymphocytes through the endothelium of venules in experimental allergic neuritis. Experientia 24:589, 1968. 3. Åström, K. E., Webster, H. deF., and Arnason, B. E.: The initial lesion in experimental allergic neuritis: a phase and electron microscopic study. J. Exp. Med. 128:469, 2002. 4. Baron-Van Evercooren, A., Kleinman, H. K., Ohno, S., et al.: Nerve growth factor, laminin, and fibronectin promote neurite growth in human fetal sensory ganglia cultures. J. Neurosci. Res. 8:179, 1982. 5. Billings-Gagliardi, S.: Mode of locomotion of Schwann cells migrating in vivo. Am. J. Anat. 150:73, 1977. 6. Billings-Gagliardi, S., Webster, H. deF., and O’Connell, M. F.: In vivo and electron microscopic observations on Schwann cells in developing tadpole nerve fibers. Am. J. Anat. 141:375, 1974. 7. Bunge, M. B., Williams, A. K., and Wood, P. M.: NeuronSchwann cell interaction in basal lamina formation. Dev. Biol. 92:449, 1982. 8. Bunge, M. B., Williams, A. K., and Wood, P. M., et al.: Comparison of nerve cell and nerve cell plus Schwann cell cultures, with particular emphasis on basal lamina and collagen formation. J. Cell Biol. 84:184, 1980. 9. Descarries, L., and Schröder, J. M.: Fixation du tissu nerveux par perfusion a grand débit. J. Microscopie 7:281, 1968. 10. Feltri, M. L., Porta, D. G., Previtali, S. G., et al.: Conditional disruption of beta 1 integrin in Schwann cells impedes interactions with Schwann cells. J. Cell Biol. 156:199, 2002. 11. Fernandez-Moran, H., and Finean, J. B.: Electron microscope and low-angle x-ray diffraction studies of the nerve myelin sheath. J. Biophys. Biochem. Cytol. 3:725, 1957. 12. Fisher, C. M., and Adams, R. D.: Diphtheritic polyneuritis—a pathological study. J. Neuropathol. Exp. Neurol. 15:243, 1956. 13. Geren, B. B.: The formation from the Schwann cell surface of myelin in the peripheral nerves of chick embryos. Exp. Cell Res. 7:558, 1954. 14. Gutrecht, J. A., and Dyck, P. J.: Quantitative teased-fiber and histologic studies of human sural nerve during postnatal development. J. Comp. Neurol. 138:117, 1970. 15. Hahn, A. F., Chang, Y., and Webster, H. deF.: Development of myelinated nerve fibers in the sixth cranial nerve of the rat: a quantitative electron microscope study. J. Comp. Neurol. 260:491, 1987.

Introduction 16. Hahn, A. F., Whitaker, J. N., Kachar, B., and Webster, H. deF.: P2, P1, and P0 myelin protein expression in developing rat sixth nerve: a quantitative immunocytochemical study. J. Comp. Neurol. 260:501, 1987. 17. Kleinman, H. K., Cannon, F. B., Laurie, G. W., et al.: Biological activities of laminin. J. Cell. Biochem. 27:317, 1985. 18. Lampert, P. W.: Mechanism of demyelination in experimental allergic neuritis. Lab. Invest. 20:127, 1969. 19. Lamperth, L., Manuelidis, L., and Webster, H. deF.: Non myelin-forming perineuronal Schwann cells in rat trigeminal ganglia express P0 myelin glycoprotein mRNA during postnatal development. Mol. Brain Res. 5:177, 1989. 20. Lamperth, L., Manuelidis, L., and Webster, H. deF.: P0 glycoprotein mRNA distribution in myelin-forming Schwann cells of the developing rat trigeminal ganglion. J. Neurocytol. 19:756, 1990. 21. Le Douarin, N.: The Neural Crest, 2nd ed. Cambridge, UK, Cambridge University Press, 1999. 22. Martin, J. R., and Webster, H. deF.: Mitotic Schwann cells in developing nerve: their changes in shape, fine structure and axon relationships. Dev. Biol. 32:417, 1973. 23. Napolitano, L. M., and Scallen, T. J.: Observations on the fine structure of peripheral nerve myelin. Anat. Rec. 163:1, 1969. 24. Ochoa, J., and Mair, W. G. P.: The normal sural nerve in man. I. Ultrastructure and numbers of fibres and cells. Acta Neuropathol. (Berl.) 13:197, 1969. 25. Omlin, F. X., Webster, H. deF., Palkovits, C. G., and Cohen, S. R.: Immunocytochemical localization of basic protein in major dense line regions of central and peripheral myelin. J. Cell Biol. 95:242, 1982. 26. Peters, A., and Muir, A. R.: The relationship between axons and Schwann cells during development of peripheral nerves in the rat. Q. J. Exp. Physiol. 44:117, 1959. 27. Peters, A., Palay, S. L., and Webster, H. deF.: The Fine Structure of the Nervous System: The Cells and Their Processes, New York, Harper & Row, 1970. 28. Peterson, R. G., and Pease, D. C.: Myelin embedded in polymerized gluteraldehyde-urea. J. Ultrastruct. Res. 41:115, 1972. 29. Podratz, J. L., Rodriguez, E., and Windebank, A. J.: Role of the extracellular matrix in myelination of peripheral nerve. Glia 35:35, 2001. 30. Robertson, J. D.: The ultrastructure of adult vertebrate peripheral myelinated fibers in relation to myelinogenesis. J. Biophys. Biochem. Cytol. 1:271, 1955. 31. Robertson, J. D.: The ultrastructure of Schmidt-Lanterman clefts and related shearing defects of the myelin sheath. J. Biophys. Biochem. Cytol. 4:39, 1958. 32. Robertson, J. D.: Preliminary observations on the ultrastructure of nodes of Ranvier. Z. Zellforsch. 50:553, 1959. 33. Robertson, J. D.: Origin of the unit membrane concept. Protoplasma 63:218, 1967. 34. Speidel, C. C.: In vivo studies of myelinated nerve fibers. Int. Rev. Cytol. 16:173, 1964. 35. Sternberger, N. H., Quarles, R. H., Itoyama, Y., and Webster, H. deF.: Myelin-associated glycoprotein demonstrated immunocytochemically in myelin and myelinforming cells of developing rats. Proc. Natl. Acad. Sci. U. S. A. 76:1510, 1979.

7

36. Thomas, P. K.: The quantitation of nerve biopsy findings. J. Neurol. Sci. 11:285, 1970. 37. Trapp, B. D., Andrews, S. B., Wong, A., et al.: Colocalization of the myelin-associated glycoprotein and the microfilament components, F-actin and spectrin, in Schwann cells of myelinated nerve fibres. J. Neurocytol. 18:47, 1989. 38. Trapp, B. D., Itoyama, Y., Sternberger, N. H., et al.: Immunocytochemical localization of P0 protein in Golgi complex membranes and myelin of developing rat Schwann cells. J. Cell Biol. 90:1, 1981. 39. Trapp, B. D., Mclntyre, L. J., Quarles, R. H., et al.: Immunocytochemical localization of rat peripheral nervous system myelin proteins: P2 protein is not a component of all peripheral nervous system myelin sheaths. Proc. Natl. Acad. Sci. U. S. A. 76:3552, 1979. 40. Tsukita, S., and Ishikawa, H.: The movement of membranous organelles in axons. J. Cell Biol. 84:513, 1980. 41. Uzman, B. G., and Nogueira-Graf, G.: Electron microscope studies of the formation of nodes of Ranvier in mouse sciatic nerves. J. Biophys. Biochem. Cytol. 3:589, 1957. 42. Waksman, B. H., and Adams, R. D.: A comparative study of experimental allergic neuritis in the rabbit, guinea pig and mouse. J. Neuropathol. Exp. Neurol. 15:293, 1956. 43. Waksman, B. H., Adams, R. D., and Mansmann, H. C.: Experimental study of diphtheritic polyneuritis in the rabbit and guinea pig. J. Exp. Med. 105:591, 1957. 44. Webster, H. deF.: Transient, focal accumulation of axonal mitochondria during the early stages of wallerian degeneration. J. Cell Biol. 12:361, 1962. 45. Webster, H. deF.: The geometry of peripheral myelin sheaths during their formation and growth in rat sciatic nerves. J. Cell Biol. 48:348, 1971. 46. Webster, H. deF.: Development of peripheral nerve fibers. In Dyck, P. J., Thomas, P. K., Griffen, J. W., et al.: (eds.): Peripheral Neuropathy, 3rd ed, Philadelphia, W. B. Saunders, p. 243–266, 1993. 47. Webster, H. deF., and Collins, G. H.: Comparison of osmium tetroxide and glutaraldehyde perfusion fixation for the electron microscopic study of the normal rat peripheral nervous system. J. Neuropathol. Exp. Neurol. 23:109, 1964. 48. Webster, H. deF., Lamperth, L., Favilla, J. T., et al.: Use of a biotinylated probe and in situ hybridization for light and electron microscopic localization of P0 mRNA in myelin-forming Schwann cells. Histochemistry. 86:441, 1987. 49. Webster, H. deF., Martin, J. R., and O’Connell, M. F.: The relationships between interphase Schwann cells and axons before myelination: a quantitative electron microscopic study. Dev. Biol. 32:401, 1973. 50. Webster, H. deF., Schröder, J. M., Asbury, A. K., and Adams, R. D.: The role of Schwann cells in the formation of “onion bulbs” found in chronic neuropathies. J. Neuropathol. Exp. Neurol. 26:276, 1967. 51. Webster, H. deF., and Spiro, D.: Phase and electron microscopic studies of experimental demyelination. I. Variations in myelin sheath contour in normal guinea pig sciatic nerve. J. Neuropathol. Exp. Neurol. 19:42, 1960. 52. Webster, H. deF., Spiro, D., Waksman, B. H., and Adams, R. D.: Phase and electron microscopic studies of experimental

8

Introduction

demyelination. II. Scwann cell changes in guinea pig sciatic nerves during experimental diphtheritic neuritis. J. Neuropathol. Exp. Neurol. 20:5, 1961. 53. Zhang, Q.-L., Lin, P.-X., Chang, Y., and Webster, H. deF.: Effects of nerve segment supernatants on cultured Schwann cell proliferation and laminin production. J. Neurosci. Res. 37:612, 1994.

54. Zhang, Q. L., Lin, P. X., Shi, D., et al.: Vasoactive intestinal peptide: mediator of laminin synthesis in cultured Schwann cells. J. Neurosci. Res. 43:496, 1996. 55. Zhang, Q.-L., Liu, J., Lin, P.-X., and Webster, H. deF.: Local administration of vasoactive intestinal peptide after nerve transection accelerates early myelination and growth of regenerating axons. J. Peripher. Nerv. Syst. 7:118, 2002.

2 Gross Anatomy of the Peripheral Nervous System ERNEST D. GARDNER AND RICHARD P. BUNGE

General Features Spinal and Peripheral Nerves Autonomic Nervous System

Gross Anatomy Head, Neck, and Upper Limb Thorax

By common definition, the peripheral nervous system includes the cranial nerves, the spinal nerves with their roots and rami, the peripheral nerves, and the peripheral components of the autonomic nervous system. This chapter presents the principles of organization and distribution of the peripheral nervous system, and outlines its major gross anatomic features; the cranial nerves and their autonomic components are considered in other chapters. Books and atlases that provide more detailed information, most of which have valuable bibliographies, are included in the references.18,26,28,53,56,57,69,72 Illustrations are limited to those that present concepts of the arrangement and distribution of spinal and peripheral nerves. The terminology used follows the Nomina Anatomica,49 translated into English where appropriate.

GENERAL FEATURES Spinal and Peripheral Nerves The dorsal and ventral roots are attached to the spinal cord by a series of filaments. After traversing the subarachnoid space, each root enters a dural pouch, or sac, and then a dural sheath, the sac being subdivided by a septum. Immediately peripheral to the spinal ganglion of the dorsal root, corresponding dorsal and ventral roots join to form a spinal nerve. The dural sheaths become confluent at the ganglion, and then merge with the epineurium of the spinal nerve. Each spinal nerve quickly

Abdomen, Pelvis, and Lower Limb Back

divides into a dorsal and a ventral (primary) ramus. The dorsal rami supply the back; the ventral rami supply the limbs and ventrolateral part of the body wall. In the cervical and lumbosacral regions, the ventral rami intermingle and form plexuses from which the major peripheral nerves emerge. Cranial nerves are also included in the term peripheral nerve. Although the funicular organization and connective tissue elements of nerves are described in other chapters, some general aspects are worth noting here. The sheaths convey the intrinsic blood and lymphatic vessels as well as the nervi nervorum, which supply the connective tissue and vessels with sensory and autonomic fibers. The sheaths also impart strength, especially against tensile stresses (see Chapter 3). Spinal roots have thinner and less well-defined sheaths and are therefore more fragile. The funicular arrangement of nerves is a matter of considerable importance, especially in connection with injury, surgical repair, and regeneration. Sunderland has described these arrangements in detail, especially with respect to the variation in number and size of funiculi, their intraneural course, and their redistribution through the formation of funicular plexuses.64 Cranial nerves differ significantly from spinal nerves, especially in their mode of embryologic development and their relation to the special senses, and because some cranial nerves supply branchial arch structures. They are attached to the brain at irregular rather than regular intervals; they are not formed of dorsal and ventral roots; some have more than one ganglion, whereas others have none; 11

12

Structure of the Peripheral Nervous System

and the optic nerve is a fiber tract of the central nervous system and not a peripheral nerve. Distribution of Spinal and Peripheral Nerves When the ventral ramus of a spinal nerve enters a plexus and joins other such rami, its component funiculi ultimately enter several of the peripheral nerves emerging from the plexus. Thus, as a general principle, each ramus entering a plexus contributes to several peripheral nerves, and each such peripheral nerve contains fibers derived from several ventral rami. Thus each spinal nerve has a pattern of ultimate distribution referred to as segmental or dermatomal, in contrast to that characteristic of peripheral nerves20,24,64 (Fig. 2–1). The term segmental refers to the fact that the longitudinal extent of spinal cord to which a right and left pair of spinal roots is attached constitutes a segment of the spinal cord. The term dermatome refers to the area of skin supplied by the sensory fibers of a single dorsal root through the dorsal and ventral rami of its spinal nerve. It is evident, then, that because most dorsal rami have cutaneous branches, the area of skin supplied by just a ventral ramus is usually not a complete dermatome. The mixture of nerve fibers in plexuses is such that it is difficult, if not impossible, to trace their course by dissection, and dermatomal distribution has been determined by physiologic experimentation and by studying disorders and

FIGURE 2–1 Schematic diagram of spinal and peripheral nerve distribution. Only sensory fibers to the skin are represented. Two nerve fibers of spinal nerve A are shown entering a plexus. One of the fibers joins peripheral nerve X, and the other joins peripheral nerve Y. Two fibers of spinal nerve B also join the two peripheral nerves. Thus the areas supplied by the two spinal nerves are different from the areas supplied by the two peripheral nerves, as shown in the subdivided rectangle. (From Gardner, E., Gray, D. J., and O’Rahilly, R.: Anatomy, 4th ed. Philadelphia, W. B. Saunders, 1975, with permission.)

surgical sections of spinal roots and nerves. The results of such studies have yielded complex maps, chiefly because of variation, overlap, and differences in method. Variation results from intrasegmental rootlet anastomoses adjacent to the cervical and lumbosacral portions of the spinal cord50 and from individual differences in plexus formation and peripheral nerve distribution. Overlap is such that section of a single root does not produce complete anesthesia in the area supplied by that root; at the most, some degree of hypalgesia may result. Overlap is greater for touch than for pain—hence the more common occurrence of hypalgesia rather than hypesthesia after section of a single root. Problems of interpretation also result from differences in method. As pointed out by Pearson and co-workers, the fundamental problem is one of segmentation and involves the question: Do certain dermatomes extend as a series of bands from the median plane of the back into the limbs, and are these bands of dermatomes arranged in an uninterrupted sequence?51 A full discussion of this problem and its embryologic basis is beyond the scope of this chapter. Nevertheless, it is important to note that the maps published by Keegan and Garrett indicate that all dermatomes, from C2 to S1, form an uninterrupted, simply arranged series extending from the median plane of the back.35 Their data were based largely on the detection of hypalgesia following compression of a single nerve root by a herniated nucleus pulposus. The degree of compression was usually not verified. Valuable as Keegan and Garrett’s maps might be as diagnostic aids in such disorders, it must be emphasized that their methods yield zones of hypalgesia that are not complete dermatomes. Moreover, as Pearson and colleagues51 pointed out, the discussion presented by Keegan and Garrett35 contains significant conceptual defects, and some of their data are at serious odds with anatomic arrangements. In contrast, Foerster published data on dermatomes that, while admittedly incomplete, derived from studies based on sound physiologic methods as follows13,14: 1. Method of residual sensitivity—a single root was left intact, as contiguous roots above and below were divided. This gave the total distribution of the intact dorsal root. 2. Constructive method—a series of contiguous roots was divided; consequently, the superior border of the resulting anesthesia represented the inferior border of the dermatome corresponding to the next higher intact root, and the inferior border of the anesthetic area represented the superior border of the next lower intact root. 3. Stimulation of dorsal roots—electrical stimulation yielded vasodilatation in areas of skin that were smaller than the dermatomes determined by the residual method, but that were similar in shape and

Gross Anatomy of the Peripheral Nervous System

general location. These areas resembled the zones reported by Head and Campbell in their classic study of the distribution of herpes zoster.25 The maps shown in Figures 2–2 and 2–3 are based on Foerster’s study of a large number of human patients in which he used combinations of the methods just outlined; they have been shown to be useful clinically.12–14 It must, however, be emphasized that the maps depict presumably normal dermatomes, not areas of sensory deficit. This

FIGURE 2–2 Front view of dermatomal distribution, based on Foerster’s data.13,14 The right and left halves show the total distribution of alternating dermatomes, thereby illustrating the degree of overlap. In some regions, however, overlap is very great and more than two dermatomes may be involved. Hence, additional figures (insets) are necessary to show the total distribution of certain dermatomes. (From Gardner, E., Gray, D. J., and O’Rahilly, R.: Anatomy, 4th ed. Philadelphia, W. B. Saunders, 1975, with permission.)

13

statement is qualified because spinal cord sensitivity may be altered by root sections in such a way as to affect the size of the dermatome being tested by peripheral stimulation. It is of interest that there is little correspondence between dermatomes and underlying muscles. Many tables have been published that list the segmental supply of muscles or of the segments controlling movements at joints. Schemes have been proposed that attempt to bring some logical proximodistal order into the segmental

14

Structure of the Peripheral Nervous System

enlargements supply the more proximal limb muscles (note, however, that segments C5 to C8 are involved in shoulder movements), and that the more caudal segments supply the more distal muscles (T1 is the chief motor supply of the intrinsic muscles of the hand). A muscle usually receives fibers from each of the spinal nerves that enter the peripheral nerve supplying it (although one spinal nerve may be its chief supply), and section of a single spinal nerve weakens but does not usually paralyze a muscle. Finally, published diagrams that illustrate the segmental sensory supply of bones and joints are based chiefly on dissections and clinical observations. Their clinical value is doubtful, especially in view of the diffuse and often referred nature of pain from these deep structures. The distribution of peripheral nerves in the limbs is fundamentally different from that of spinal nerves. These enable one to contrast cutaneous nerve supply (Figs. 2–4 and 2–5) with the dermatomes (Figs. 2–2 and 2–3). In the trunk, however, peripheral and spinal nerves are usually identical in their cutaneous distribution. Peripheral nerves to the limbs also overlap, but to a lesser extent than do spinal nerves. Thus, if a peripheral nerve is cut, the muscles supplied by the nerve are greatly weakened or completely paralyzed, autonomic dysfunction occurs, and sensation is lost in the central part of the area of distribution of the nerve and diminished at the edges of its area of supply. The last is due to overlap from adjacent peripheral nerves; the overlap is less than in the case of spinal nerves and is often less for touch than for pain. Peripheral nerves vary in their course and distribution, and adjacent nerves may communicate with each other.18,26 Such communications sometimes account for unexpected residual sensation or movement after section of a nerve. Whatever the pattern of distribution, major peripheral nerves usually have five general kinds of branches:

FIGURE 2–3 Back view of dermatomes, based on Foerster’s data13,14 as explained in legend for FIGURE 2–2. (From Gardner, E., Gray, D. J., and O’Rahilly, R.: Anatomy, 4th ed. Philadelphia, W. B. Saunders, 1975, with permission.)

1. Muscular—motor, sensory, and autonomic fibers; the sensory fibers are from muscle fibers and associated connective tissue and tendons and sometimes joints 2. Cutaneous or mucosal—each has sensory and autonomic fibers; the cutaneous branches often include fibers from subjacent joints, ligaments, and tendons, especially in the case of digital nerves 3. Articular—arising where the nerve crosses a joint and containing sensory and autonomic fibers 4. Vascular—sensory and autonomic to adjacent blood vessels 5. Terminal—one, several, or all of the foregoing

supply of limb muscles.44 These tables and schemes, however, tend to be limited in value because of variation, overlap, and incomplete information about segmental motor origin and distribution. The general arrangement is that the more rostral segments of the cervical and lumbosacral

Also of importance, in addition to peripheral distribution, are the order and site of origin of individual branches, including distances to the muscles they supply, as well as the order in which structures are innervated, and the nature and extent of variations.63,66–68 Moreover, in several instances peripheral nerves pass through osteofibrous

Gross Anatomy of the Peripheral Nervous System

15

FIGURE 2–4 Approximate areas of cutaneous nerve distribution illustrated in the right upper limb. Neither variation nor overlap is shown. (From Gardner, E., Gray, D. J., and O’Rahilly, R.: Anatomy, 4th ed. Philadelphia, W. B. Saunders, 1975, with permission.)

canals, where they may be subject to compression. The chief sites of compression and related literature are summarized by Smorto and Basmajian.61

Autonomic Nervous System Certain general principles of organization are presented here to clarify the relationships of spinal and peripheral nerves with the autonomic nervous system. By classic anatomic definition, the autonomic nervous system, sometimes called the visceral or vegetative nervous system, is that system of motor nerve fibers that supplies cardiac muscle, smooth muscle, and glands and consists, in its simplest form, of a pathway of two succeeding nerve cells. The first cell is located in the brain or spinal cord, the second in a ganglion outside the brain and spinal cord. Anatomically and functionally, however, the autonomic nervous system is much more complex than the simplistic definition just given would indicate, to a degree far beyond the scope of this chapter. The autonomic pathway or outflow begins with certain nerve cells in the brainstem and spinal cord. The axons of these cells, termed preganglionic fibers, leave the brainstem and spinal cord over certain cranial nerves and ventral roots and synapse in peripheral autonomic ganglia (including the suprarenal medullae and certain chromaffin cells). The axons of the ganglion cells are termed postganglionic fibers and are distributed to cardiac muscle, smooth muscle, and certain gland cells. The locations, arrangements, connections, and patterns of distribution of

the preganglionic and postganglionic autonomic fibers are grouped, on the basis of an anatomic classification of subdivisions of the pathways, into sympathetic, parasympathetic, and enteric systems or divisions. Most viscera are supplied by both sympathetic and parasympathetic divisions; the enteric system is limited to the wall of the bowel. Works cited in the references, several of which contain valuable bibliographies, provide additional information about the autonomic nervous system.9,28,38,40,46,54 Sympathetic (Thoracolumbar) System This part of the autonomic nervous system comprises the preganglionic fibers that issue from the thoracic and upper lumbar levels of the spinal cord. These fibers travel in ventral roots and spinal nerves to reach peripheral sympathetic ganglia, where they synapse with ganglion cells. The locations of these ganglia are related to the embryonic migration of cells from the neural tube and neural crest, which form the ganglia of the sympathetic trunk and prevertebral plexuses. Sympathetic Trunk and Ganglia. Most preganglionic fibers leave the spinal nerves or ventral rami and reach the adjacent sympathetic trunk and ganglia by way of rami communicantes (Fig. 2–6). The sympathetic trunks are long nerve strands, one on each side of the vertebral column, extending from the base of the skull to the coccyx. Each usually contains 21 to 25 ganglia of varying sizes, but broader ranges have been recorded.

16

Structure of the Peripheral Nervous System

FIGURE 2–5 Approximate areas of cutaneous nerve distribution to the right lower limb. Neither variation nor overlap is shown. (From Gardner, E., Gray, D. J., and O’Rahilly, R.:Anatomy, 4th ed. Philadelphia, W. B. Saunders, 1975, with permission.)

Many of the preganglionic fibers that enter the sympathetic trunks synapse in the ganglia of the trunks and in accessory ganglia; those that do not synapse continue through and reach ganglia of the prevertebral plexuses. Of the postganglionic fibers arising in the trunk ganglia, some go directly to adjacent viscera and blood vessels; the others return to spinal nerves and dorsal and ventral rami by way of rami communicantes. The fibers of both the ventral and dorsal6 rami eventually supply secretory fibers to sweat glands, motor fibers to the smooth muscle of the hair follicles (arrectores pilorum), and motor fibers to the smooth muscle of the blood vessels of the limbs and walls of the trunk. However, some of the

postganglionic fibers to the back and to the proximal parts of the limbs reach these parts by accompanying blood vessels. Each trunk ganglion has one to four rami communicantes, which connect it with the corresponding nerve and often with the nerve above or below. The ramus or rami containing the most postganglionic fibers tend to connect with the corresponding nerve (each spinal nerve or one of its rami receives such fibers). In the thorax, the rami containing the most preganglionic fibers are more oblique in direction (coming from the spinal nerve above or below), and they are more lateral in position, that is, farther from the spinal cord.55

Gross Anatomy of the Peripheral Nervous System

17

viscera or blood vessels. The preganglionic fibers that enter these ganglia are those that traverse the sympathetic trunks without synapsing. They reach the prevertebral ganglia by way of branches termed splanchnic nerves (the term splanchnic is also applied to certain visceral branches in the pelvis). These nerves may contain ganglia along their course. The postganglionic fibers from these various ganglia go directly to adjacent viscera and blood vessels. These ganglia, their preganglionic and postganglionic fibers, and the thoracic and abdominal branches of the vagus nerves form what are termed the prevertebral plexuses. Many sensory fibers from viscera (for pain as well as reflexes) traverse these plexuses and reach the spinal cord by way of splanchnic nerves, rami communicantes, and dorsal roots, or the brainstem by way of the vagus nerves. Some of the migrating embryonic cells form the chromaffin system, in particular the medullae of the adrenal glands. Correspondingly, some of the preganglionic fibers in the splanchnic nerves end in relation to these cells. Accessory Ganglia. Some of the migrating embryonic cells stop along spinal nerves, ventral rami, and rami communicantes, especially in the cervical, lower thoracic, and upper lumbar levels. Here they form scattered sympathetic cells, which often are collected into definite accessory (intermediate) ganglia.60,76 The postganglionic fibers from these cells continue in their associated nerves; hence sympathectomies in these regions may not be completely effective.48

FIGURE 2–6 The sympathetic trunks. Preganglionic rami communicantes are shown as interrupted lines, postganglionic rami as solid black. Based on studies by Pick and Sheehan.55 (From Gardner, E., Gray, D. J., and O’Rahilly, R.: Anatomy, 4th ed. Philadelphia, W. B. Saunders, 1975, with permission.)

Prevertebral Plexuses and Ganglia and Chromaffin System. Many embryonic ganglion cells migrate to a position in front of the vertebral column, where they form ganglionic masses that are named according to adjacent

Parasympathetic (Craniosacral) System This part of the autonomic system comprises the preganglionic fibers that issue from the brainstem by way of the 3rd, 7th, 9th, 10th, and 11th cranial nerves, and from the sacral cord by way of its second and third or third and fourth ventral roots. The ganglion cells are usually in or near the organ to be innervated; hence, the postganglionic fibers are short. Moreover, none seems to go to the blood vessels, smooth muscle, or glands of the limbs or body walls. The cranial nerves are described elsewhere, but it is worth noting that the parasympathetic ganglia of the third, seventh, and ninth cranial nerves form what are termed the cephalic or cranial parasympathetic ganglia (ciliary, pterygopalatine, otic, and submandibular). Their postganglionic fibers supply the eye, lacrimal and salivary glands, and mucous and serous glands of the oral and nasal cavities. The preganglionic fibers of the 11th nerve are distributed with the vagus nerves; the ganglion cells are near or in the walls of the viscera of the neck, thorax, and abdomen. As mentioned earlier, vagal branches contribute to the prevertebral plexuses.

18

Structure of the Peripheral Nervous System

The sacral preganglionic parasympathetic fibers leave the sacral plexus as pelvic splanchnic nerves, enter the inferior hypogastric plexus, and reach ganglion cells in the walls of pelvic organs. The Enteric System In his extensive treatise on the autonomic nervous system, Langley classified the enteric plexuses of the gut as a third autonomic division.43 Langley recognized that the enteric nervous system receives sympathetic and parasympathetic inputs, but he considered this extensive ganglionic network relatively autonomous in its control of gut function. More recent studies have established the remarkable extent and complexity of the enteric nervous system. Furness and Costa have estimated that there may be 108 ganglion cells in the gut wall of the human,15 and Gershon has reviewed evidence that the many types of neurons present employ at least three known neurotransmitters plus as many as nine putative transmitters or modulators.19 Physiologic aspects of visceral functional control are considered in Chapter 13. The enteric nervous system extends the length of the gastrointestinal tract from esophagus to rectum. It is composed of two major plexuses of ganglion cells, as well as interconnecting fibers. The more internally located neurons (Meissner’s plexus) are located in the submucosa of the gut wall; the more externally located neurons (Auerbach’s or the myenteric plexus) lie between the external longitudinal and internal circular smooth muscle layers of the muscularis externa. Neurons occur in these regions as small aggregates; these aggregates are interconnected by fascicles of unmyelinated nerve fibers, and extensive connections are also found between the submucosal neurons and the myenteric components. Certain histologic aspects of these plexuses are unusual. The small ganglia are entirely nonvascular, and the cytoarchitecture is more like that of the central nervous system than of peripheral sensory or autonomic ganglia.16,19 The similarity to the central nervous system derives from the lack of space and extracellular matrix material between the individual cells of the ganglia and the diversity of neuronal and synaptic types present. Details of the histology and cytology of this unique portion of the autonomic nervous system are included in Chapter 12.

GROSS ANATOMY The gross anatomy of the peripheral nervous system of each region of the body and the distribution and branching of nerves are outlined here, with emphasis on certain regional or topographic features. Additional special details are to be found in references dealing with peripheral nerve injuries.24,58,61,64

Head, Neck, and Upper Limb Cervical Plexus The ventral rami of the upper four cervical nerves unite to form the cervical plexus; those of the lower four, together with the greater part of that of the first thoracic, join to form the brachial plexus. (The distribution of the dorsal rami of the cervical nerves is described later.) Each cervical ramus receives one or more rami communicantes (containing postganglionic fibers) from a cervical sympathetic ganglion, and often from the vertebral nerve and plexus as well. The cervical plexus is arranged as an irregular series of loops located in front of the levator scapulae and scalenus medius muscles, under cover of the sternocleidomastoid muscle and internal jugular vein. The branches of the loops are superficial (to the skin of the back of the head, the neck, and the shoulder) and deep (to certain neck muscles and the diaphragm). Superficial Branches. These cutaneous branches emerge near the middle of the posterior border of the sternocleidomastoid. They include the lesser occipital nerve, which ascends behind the ear and supplies some of the skin on the side of the head and on the cranial surface of the ear; the great auricular nerve, which ascends obliquely across the sternocleidomastoid muscle to supply the skin over the parotid gland, over the mastoid process, and on both surfaces of the ear; the transverse cervical nerves, which supply the skin on the side and front of the neck; and the supraclavicular nerves, derived from a common trunk that divides into anterior, middle, and posterior supraclavicular nerves, which supply the skin over the shoulder and the front of the thorax. Deep Branches. These are the ansa cervicalis, the phrenic nerve, and the muscular branches to the sternocleidomastoid, trapezius, levator scapulae, scalene, and prevertebral muscles. There are also small communicating branches to the 10th, 11th, and 12th cranial nerves. The ansa cervicalis (ansa hypoglossi) is a loop formed by fibers of the first three (or the second and third) cervical nerves. It presents a superior root (the so-called descending branch of the hypoglossal nerve), which connects it with the hypoglossal nerve, but which consists of fibers from the second or first cervical nerve, and an inferior root (nervus descendens cervicalis), which connects it with branches from the second and third cervical nerves. The ansa cervicalis supplies the infrahyoid muscles. The thyrohyoid, however, receives its cervical fibers by way of the hypoglossal nerve. The phrenic nerve, which supplies the diaphragm, arises chiefly from the fourth cervical nerve, but commonly has a root from the fifth as well, and sometimes

Gross Anatomy of the Peripheral Nervous System

from the third. Rarely, it may arise entirely from the accessory phrenic nerve. It descends in front of the scalenus anterior muscle, enters the thorax, and descends between the pericardium and mediastinal pleura. The right phrenic nerve passes in front of the root of the right lung, pierces the diaphragm near the opening for the inferior vena cava (or traverses the opening for that vessel), and distributes most of its motor fibers from below. The left phrenic nerve passes in front of the root of the left lung and pierces the diaphragm immediately to the left of the pericardium. Each nerve carries motor fibers to the diaphragm and sensory and autonomic fibers for the diaphragm, pleura, and peritoneum. Referred pain from the area of supply of a phrenic nerve is commonly felt in the skin over the trapezius (C4 and C5). Pain is sometimes referred to the region of the ear; this is probably related to a contribution from the third cervical nerve. The accessory phrenic nerve is present in about one third of instances. It usually arises from the fifth cervical nerve through the nerve to the subclavius, but may arise from the cervical plexus or from a cardiac branch of a cervical sympathetic ganglion. The accessory nerve runs a variable course before joining the phrenic nerve in the thorax. If such a nerve is present, section of the phrenic nerve in the neck will not paralyze the corresponding half of the diaphragm completely. Brachial Plexus The brachial plexus, which is situated partly in the neck and partly in the axilla, is formed by the ventral rami of the lower four cervical nerves and the greater part of the ventral ramus of the first thoracic nerve. These rami lie first between the scalenus anterior and the scalenus medius, and then in the posterior triangle of the neck. Here the plexus is situated above the clavicle, posterior and lateral to the sternocleidomastoid. It lies above and behind the third part of the subclavian artery and is crossed by the inferior belly of the omohyoid. In this situation, the plexus may be injected with a local anesthetic. The general features of the plexus are discussed in greater detail elsewhere (see Chapter 55).11,22,23,36 The plexus descends behind the concavity of the medial two thirds of the clavicle, and accompanies the axillary artery under cover of the pectoralis major. It is enclosed with the axillary vessels in the axillary sheath, and its cords are arranged around the second part of the axillary artery behind the pectoralis minor. The terminal branches of the plexus arise at the inferolateral border of the pectoralis minor. The brachial plexus can be marked on the surface by a line from the posterior margin of the sternocleidomastoid at the level of the cricoid cartilage to the midpoint of the clavicle. The plexus can be palpated both above and below the omohyoid, in the angle between the clavical and the sternocleidomastoid.

19

The common, though not invariable, arrangement of branches is as follows. The ventral rami of the fifth and sixth cervical nerves unite to form the upper trunk, that of the seventh remains single as the middle trunk, and the eighth cervical and first thoracic rami form the lower trunk. Each trunk then divides into an anterior and a posterior division (for the front and back of the limb, respectively). The anterior divisions of the upper and middle trunks unite to form the lateral cord, the anterior division of the lower trunk forms the medial cord, and the three posterior divisions form the posterior cord. The terminal branches arise from the three cords. The brachial plexus is thus composed successively of (1) ventral rami and trunks that lie in the neck in relation to the subclavian artery (the lowest trunk lies on the first rib behind the subclavian artery), (2) divisions that lie behind the clavicle, and (3) cords and branches that lie in the axilla in relation to the axillary artery. Variations. The brachial plexus frequently receives contributions from the fourth cervical or the second thoracic nerve also, or from both. When the contribution from the fourth cervical is large and that from the first thoracic small, the plexus is described as being prefixed in relation to the vertebral column. When the contributions from the first and second thoracic nerves are large, the plexus is termed postfixed; this latter situation is prominent when the first rib is rudimentary.3,10 It is, however, uncommon to find prefixation or postfixation in the sense of a complete shift in which a full nerve is gained at one end while one is completely lost at the other end. Other variations (in gross form, component arrangements, and branching) are common. Branches of the Ventral Rami. These are the dorsal scapular and long thoracic nerves, and twigs to the scalene and longus colli muscles. The dorsal scapular nerve (chiefly C5) sometimes supplies the levator scapulae but is distributed chiefly to the rhomboids. The long thoracic nerve arises by three roots from C5 to C7, descends behind the brachial plexus, and enters the external surface of the serratus anterior. Branches of the Trunks. These are the nerves to the subclavius and the suprascapular nerve and occasionally medial and lateral pectoral nerves also. The nerve to the subclavius (C5) descends behind the clavicle, in front of the brachial plexus to supply the subclavius muscle and the sternoclavicular joints. It often sends fibers to the phrenic nerve by a communicating branch, the accessory phrenic nerve described earlier. The suprascapular nerve (C5, C6) passes through the scapular notch, supplies the acromioclavicular and shoulder joints and the supraspinatus muscle, and then passes through the spinoglenoid notch to end in the infraspinatus.

20

Structure of the Peripheral Nervous System

Branches of the Cords. These include a number of cutaneous and muscular branches and the important terminal branches, namely the median, ulnar, radial, musculocutaneous, and axillary nerves. Several lateral pectoral nerves (C5 to C7) arise from the lateral cord (or upper and middle trunks) and supply both pectoral muscles as well as the acromioclavicular and shoulder joints. Medial pectoral nerves (C8, T1) from the medial cord (or lower trunk) supply both pectoral muscles. Also arising from the medial cord is the medial antebrachial cutaneous nerve (C8, T1), which descends with the brachial artery to the lower part of the arm, where it becomes cutaneous. It divides into anterior and ulnar branches, which supply the skin of the medial half of the forearm. The medial brachial cutaneous nerve (chiefly T1), also a branch of the medial cord, is a small nerve that supplies the medial and posterior aspects of the arm and communicates with the intercostobrachial nerve. From the posterior cord, there arise the upper subscapular nerve (or nerves) (C5), for the subscapularis; the thoracodorsal nerve (C7, C8), for the latissimus dorsi; and the lower subscapular nerve (or nerves) (C5, C6), for the subscapularis and teres major. The median nerve (C5, C6 to C8, T1) arises from the medial and lateral cords by medial and lateral roots, which unite in a variable fashion (Fig. 2–7). Descending as a part of the neurovascular bundle, it enters the cubital fossa, where it lies under cover of the bicipital aponeurosis and supplies the elbow joint. It then passes between the two heads of the pronator teres and descends on the deep surface of the flexor digitorum superficialis. It is indicated on the surface by a line down the middle of the forearm to the midpoint between the styloid processes. Just above the flexor retinaculum, it is quite superficial in the interval between the flexor carpi radialis and palmaris longus tendons, and completely so if the latter muscle is absent. The median nerve enters the hand by passing through the carpal canal, behind the flexor retinaculum. It then spreads out in an enlargement and divides into its terminal branches under cover of the palmar aponeurosis and the superficial palmar arch. Usually it divides first into lateral and medial portions or divisions. The median nerve has no muscular branches in the upper arm (i.e., above the elbow). In the forearm, it supplies all the muscles of the front of the forearm except the flexor carpiulnaris and the medial half of the flexor digitorum profundus. In the cubital fossa, a bundle of muscular branches is given to the pronator teres, flexor carpi radialis, palmaris longus, and flexor digitorum superficialis. The anterior interosseous nerve also arises in the cubital fossa. It descends on the front of the interosseous membrane, supplies the flexor pollicis longus and flexor digitorum profundus (lateral part), passes behind and supplies the pronator quadratus, and ends in twigs to the wrist and intercarpal joints. In the lower part of the

FIGURE 2–7 Schematic representation of the sequences of muscular and cutaneous branches of the musculocutaneous and median nerves. The black dot in each instance represents a muscle or a muscle group. The sequences of muscular branches are the more common ones. Based on studies by Sunderland and Ray.68 (From Gardner, E., Gray, D. J., and O’Rahilly, R.: Anatomy, 4th ed. Philadelphia, W. B. Saunders, 1975, with permission.)

forearm, the median nerve gives off an inconstant palmar branch for the supply of a small area of the skin of the palm. The median and ulnar nerves may communicate in the forearm, and the anterior interosseous nerve may communicate with the ulnar nerve. In the hand, the lateral division gives off an important muscular branch (recurrent branch) for the abductor pollicis, flexor pollicis brevis, and opponens pollicis, which anastomoses with the deep branch of the ulnar nerve.21 The lateral division then divides in such a way as to furnish three palmar digital nerves for both sides of the thumb and the lateral aspect of the index finger and the first lumbrical. The medial division divides in such a way as to furnish four palmar digital nerves for the adjacent sides of the index and middle fingers, middle and ring fingers, and second lumbrical. All digital nerves, near their terminations, send branches dorsally to the backs of the distal parts of the fingers, and all digital nerves, in both hand and foot

Gross Anatomy of the Peripheral Nervous System

supply ligaments, joints, and tendons of the digits, as well as skin. Rarely the median nerve supplies the first dorsal interosseous or the adductor pollicis. As a general rule, the median nerve tends to supply the thenar muscles, and the ulnar nerve the remainder. However, the dividing line between the two distributions is variable, and either nerve may invade the territory of the other. The ulnar nerve (C7, C8, T1) arises from the medial cord, but usually has a lateral root also from the lateral cord, which carries the C7 fibers (Fig. 2–8). It descends with the neurovascular bundle, pierces the medial intermuscular septum, and descends behind the medial epicondyle, between the two heads of the flexor carpi ulnaris. Here it supplies the elbow joint. It then descends on the flexor digitorum profundus to the middle of the forearm and then along the lateral side of the ulnar artery. Both enter the hand by passing in front of the flexor retinaculum, a slip of which covers them, between the pisiform and the hook of the hamate. The nerve then divides into its

FIGURE 2–8 Schematic representation of the sequences of muscular and cutaneous branches of the ulnar nerve. The sequences of muscular branches are the more common ones. Based on studies by Sunderland and Hughes.67 (From Gardner, E., Gray, D. J., and O’Rahilly, R.: Anatomy, 4th ed. Philadelphia, W. B. Saunders, 1975, with permission.)

21

superficial and deep terminal branches. The course of the nerve in the forearm may be indicated on the surface by a line from the front of the medial epicondyle to the lateral margin of the pisiform. There are no muscular branches in the upper arm (i.e., above the elbow). In the upper forearm, muscular branches are given to the flexor carpi ulnaris and the medial part of the flexor digitorum profundus. In the middle of the forearm, the large cutaneous dorsal branch arises and then descends to the back of the hand. In the lower part of the forearm, a variable palmar branch is given to the skin on the medial side of the palm. In the hand, the dorsal branch gives twigs to the skin of the back of the hand and then divides into three dorsal digital nerves that supply the dorsal aspects of the medial side of the little finger and the adjacent sides of the ring and middle fingers (the most lateral of these digital nerves communicates with the adjoining digital branch of the superficial radial nerve). The dorsal digital nerves (including those from the radial nerve) do not reach the tips of the fingers. They extend as far as the nail in the first and fifth fingers, but only to the middle or proximal phalanges in the second, third, and fourth fingers.5 The innervation is completed distally in all fingers by palmar digital nerves. It should be emphasized that there is considerable variation in the number of digits supplied by the median, ulnar, and radial nerves32,34 and that an increase in the number of digits supplied by one nerve is accompanied by a decrease in the number of digits supplied by an adjacent nerve.62 Also, the area of skin supplied on the dorsum of the hand may be invaded by one of the antebrachial cutaneous nerves.62 In the palm of the hand, the superficial branch of the ulnar nerve supplies the palmaris brevis and then divides in such a way as to provide palmar digital nerves for the medial side of the little finger and the adjacent sides of the little and ring fingers. The deep branch, after supplying the hypothenar muscles, extends laterally with the deep palmar arch. In its course it supplies all the interossei, the third and fourth lumbricals, and the adductor pollicis and ends in the flexor pollicis brevis (deep head). The musculocutaneous nerve (C5 to C7) arises from the lateral cord, pierces the coracobrachialis, and descends between the biceps and brachialis, supplying all three muscles and the elbow joint (see Fig. 2–7). On reaching the lateral side of the arm, it continues as the lateral antebrachial cutaneous nerve. This nerve divides into anterior and posterior branches, which supply the skin on the lateral part of the forearm as far as the thumb. Either branch may reach the dorsum of the hand. The radial nerve (C5, C6 to C8, T1), a continuation of the posterior cord, is the largest branch of the brachial plexus (Fig. 2–9). It begins at the medial margin of the biceps, opposite the posterior axillary fold, and descends

22

Structure of the Peripheral Nervous System

FIGURE 2–9 Schematic representation of the sequence of muscular branches of the axillary and radial nerves. The sequences of muscular branches are the more common ones. Based on studies by Sunderland.63 (From Gardner, E., Gray, D. J., and O’Rahilly, R.: Anatomy, 4th ed. Philadelphia, W. B. Saunders, 1975, with permission.)

behind the axillary and brachial arteries. It then runs obliquely across the back of the arm, winding around the humerus under cover of the lateral head of the triceps. Lying first on the medial head, and then in the groove on the humerus, it pierces the lateral intermuscular septum at the superior point of trisection of a line between the deltoid insertion and the lateral epicondyle, and comes forward into the cubital fossa. Here it is deeply placed, between the brachialis and brachioradialis. At about the level of the lateral epicondyle, it divides into superficial and deep branches. In thin persons, the radial nerve may be palpated as it winds around the humerus, especially about 1 to 2 cm below the deltoid insertion, and also in the interval between the brachialis and brachioradialis. In the arm, the posterior brachial cutaneous nerve arises from the radial nerve in the axilla. It supplies the skin on the back of the arm. Several muscular branches are given to the three heads of the triceps; that to the medial head accompanies the ulnar nerve and is known as the ulnar

collateral nerve. A branch is also given to the anconeus, and this branch also supplies the elbow joint. The lower lateral brachial cutaneous nerve arises in the arm (sometimes in common with the next branch) and supplies the lateral surface of the lower part of the arm. The posterior antebrachial cutaneous nerve arises in the groove, pierces the lateral head of the triceps, and supplies the skin of the back of the forearm as far as the wrist. Muscular branches of the radial nerve supply the brachioradialis, extensor carpi radialis longus and brevis, and brachialis (probably sensory), and small twigs are given to the elbow joint. The superficial branch is the continuation of the radial nerve. It descends under cover of the brachioradialis, lateral to the radial artery. It then turns dorsally and becomes subcutaneous. It supplies the lateral side of the dorsum of the hand and then divides into a number of dorsal digital nerves for the thumb and index and middle fingers that cross the anatomic snuffbox. The superficial branch may be palpable before entering the snuffbox, and its terminal digital branches can be felt as they cross the tendon of the extensor pollicis longus. A distribution to all five digits has been recorded.45 The deep branch of the radial nerve winds laterally around the radius between the superficial and deep layers of the supinator, where it is often in contact with a bare area of the radius.8 On the back of the forearm, it lies between the superficial and deep extensors. After supplying the supinator and superficial group (the extensor digitorum, extensor digitiminimi, extensor carpi ulnaris, and often extensor carpiradialis brevis), it continues as the posterior interosseous nerve (the entire deep branch is sometimes given this name). This nerve descends on the interosseous membrane and ends on the back of the carpus, where it supplies the wrist and intercarpal joints. During its course, it supplies the abductor pollicis longus, the extensor pollicis brevis and longus, and the extensor indicis. The axillary nerve (C5, C6) is a branch of the posterior cord (see Fig. 2–9). It lies behind the axillary artery, lateral to the radial nerve. At the lower border of the subscapularis, it turns posteriorly, near the joint capsule, and passes through the quadrangular space, at a level indicated by a horizontal plane through the middle of the deltoid. As it does so, it supplies the shoulder joint. It passes medial to the surgical neck of the humerus and divides into anterior and posterior branches. The anterior branch supplies the deltoid. The posterior branch supplies the deltoid and teres minor and then turns around the lower border of the deltoid to supply the skin on the back of the arm as the upper lateral brachial cutaneous nerve. Autonomic Nerve Supply The parasympathetic supply to the head and neck is by way of the cranial nerves mentioned previously and is described with the individual nerves.

23

Gross Anatomy of the Peripheral Nervous System

Sympathetic Trunk and Ganglia. The preganglionic fibers for the head and neck arise chiefly from the first and second thoracic segments of the spinal cord (the range is C8 to T4) (see Fig. 2–6). The fibers reach the thoracic part of the sympathetic trunk, in which they ascend to the cervical part, where they synapse and from which postganglionic fibers are distributed to the blood vessels, smooth muscle, and glands of the head and neck. The preganglionic fibers for the upper limb arise from approximately the 2nd to the 9th or 10th thoracic segments of the spinal cord. They enter the sympathetic trunk and ascend to synapse in the stellate and adjacent ganglia. Postganglionic fibers reach the upper limb by way of fibers that accompany the subclavian and axillary arteries and their branches. The cervical part of the sympathetic trunk consists of three or four ganglia connected by an intervening cord or cords. Postganglionic fibers leave by rami communicantes (one or more to each cervical ventral ramus) by branches that accompany adjacent blood vessels and by branches that go directly to certain cranial nerves and to viscera of the neck and thorax. The superior cervical ganglion lies behind the internal carotid artery, in front of the longus capitis muscle, and extends from the first to the second or third cervical vertebra. The ganglion gives rami communicantes to the upper cervical ventral rami and the last four cranial nerves, twigs to the carotid body and sinus and to the pharyngeal plexus, and cervical cardiac nerves to the heart. Several branches form a plexus along the external carotid artery,17 with some of the fibers reaching the salivary glands. One or more large branches form an internal carotid nerve,39 which ascends with the internal carotid artery to supply the eye, orbit, and intracranial structures by forming first an internal carotid plexus and then subsidiary plexuses along the branches of the artery and by giving twigs to various nerves (tympanic, greater petrosal, and third, fourth, fifth, and sixth cranial) and to the ciliary ganglion. The middle cervical ganglion is quite variable and is often fused with either the superior or the vertebral ganglion. It usually lies just above the arch formed by the inferior thyroid artery (along which twigs form a plexus), at the level of the sixth cervical vertebra. Rami are given to cervical nerves, usually C4 to C6, as well as to the heart. The vertebral ganglion usually lies in front of the vertebral artery, just below the arch of the inferior thyroid artery, and about at the level of the seventh cervical vertebra. As the sympathetic trunk extends downward from the vertebral to the stellate (or inferior cervical) ganglion, it forms two or more cords that pass on each side of the vertebral artery. Rami from the ganglion reach some of the lower cervical nerves and thereby enter the branchial plexus. Twigs are given to the vertebral plexus (discussed later), and a cord termed the ansasubclavia loops in front of and below the first part of the subclavian

artery to the stellate (or inferior cervical) ganglion. Branches of the ansa contribute to a plexus along the subclavian artery. The cervicothoracic (stellate) ganglion has two parts: the inferior cervical and the first thoracic (occasionally a second and third also). These ganglia may be completely fused.29 The ganglionic mass lies usually at the level of the seventh cervical and first thoracic vertebrae, in front of the eight cervical and first thoracic nerves and the seventh cervical transverse process and neck of the first rib, and behind the vertebral artery. The stellate ganglion receives preganglionic fibers from the first, or first and second, thoracic nerves. Its postganglionic fibers supply chiefly the upper limb by way of rami to the lower cervical and upper thoracic nerves and branches to the subclavian and vertebral arteries. Branches of the vertebral and stellate ganglia form a vertebral plexus, which accompanies the vertebral artery into the posterior cranial fossa. During its course, postganglionic fibers are given to the lower cervical nerves.65 Some fibers ascend separately, behind the artery, as a distinct vertebral nerve, which extends to the level of the axis or atlas. Postganglionic fibers in it are given to the cervical nerves and spinal meninges.

Thorax Thoracic Nerves Each of the 12 thoracic nerves gives off a meningeal branch and then, after emerging from an intervertebral foramen, divides into a dorsal and a ventral ramus. The meningeal branches and dorsal rami are described with the back (see later). Each ventral ramus is connected to the sympathetic trunk by a variable number of rami communicantes, and each runs a separate course forward, supplying the skin, muscles, and serous membranes of the thoracic and abdominal walls. The ventral rami of the first 11 nerves are called intercostal nerves; that of the 12th is the subcostal nerve. Typical Intercostal Nerves. The fourth, fifth, and sixth intercostal nerves are typical intercostal nerves and supply only the thoracic wall (Fig. 2–10). Each passes below the neck of the numerically corresponding rib and enters the costal groove below the posterior intercostal vessels. At the anterior end of the intercostal spaces, the nerves turn forward through the overlying muscles and, as anterior cutaneous branches, are distributed to the skin of the front of the thorax. Here they give off medial mammary branches. At the angle of the rib, each nerve supplies the external intercostal muscle and gives off a collateral and a lateral cutaneous branch. The collateral branch passes forward in the intercostal space and ends anteriorly as a lower anterior cutaneous nerve. The lateral cutaneous branch pierces the overlying muscles and divides into

24

Structure of the Peripheral Nervous System

FIGURE 2–10 Diagrammatic representation of the nerves of the thoracic wall. The thickness of the intercostal muscles is exaggerated. (From Gardner, F., Gray, D. J., and O’Rahilly, R.: Anatomy, 4th ed. Philadelphia, W. B. Saunders, 1975, with permission.)

anterior and posterior branches that supply the skin of the thorax. Some of the anterior branches form lateral mammary branches. The intercostal nerves supply the intercostal, subcostal, serratus posterior superior, and transversus thoracis muscles. Special Intercostal Nerves. The first, second, and third intercostal nerves are special in that they supply the arm as well as the thorax. The first thoracic nerve is the largest of the thoracic spinal nerves. It divides into a larger upper and a smaller lower part. The upper joins the brachial plexus, and the lower becomes the first intercostal nerve. Its distribution is like that of a typical intercostal nerve,3 except that its lateral cutaneous branch supplies the skin of the axilla and may communicate with the intercostobrachial nerve. The second intercostal nerve, which often contributes to the brachial plexus, also has a distribution like that of a typical intercostal nerve, except that its lateral cutaneous branch passes into the arm as the intercostobrachial nerve. This nerve pierces the overlying muscles and supplies the skin on the back and medial side of the arm as far as the elbow. It usually anastomoses with the posterior and medial brachial cutaneous nerves and supplies the axillary arch when that muscle is present. The third intercostal nerve similarly has a distribution like that of a typical intercostal nerve, but its lateral cutaneous branch often gives a twig to the medial side of the arm. The 7th to 11th intercostal nerves are also special in that they supply the abdominal wall as well as the thorax (Fig. 2–11). Known as thoracoabdominal nerves, they course

FIGURE 2–11 The cutaneous distribution of the thoracoabdominal nerves. (From Gardner, E., Gray, D. J., and O’Rahilly, R.: Anatomy, 4th ed. Philadelphia, W. B. Saunders, 1975, with permission.)

forward and downward to the anterior ends of the intercostal spaces. They continue between the transversus and internal oblique muscles and then between the rectus abdominis and the posterior wall of its sheath. Here, each nerve divides into two branches. The larger one forms a plexiform arrangement that gives twigs to the rectus and from which an anterior cutaneous branch pierces the rectus to supply the overlying skin. The smaller branch also supplies the rectus and may pierce the rectus and become cutaneous. During its course, each thoracoabdominal nerve gives off a lateral cutaneous branch that pierces the external oblique muscle and divides into anterior and posterior branches. These supply the back, side, and front, of the abdominal wall. The thoracoabdominal nerves also supply the intercostal, subcostal, serratus posterior inferior, transversus abdominis, external and internal oblique, and rectus abdominis muscles, and give sensory twigs to adjacent diaphragm, pleura, and peritoneum. The ventral ramus of the 12th thoracic nerve is special in that it is subcostal rather than intercostal in position and is known as the subcostal nerve. It enters the abdomen, courses downward and laterally behind the kidney, pierces the transversus abdominis, and passes between this muscle

Gross Anatomy of the Peripheral Nervous System

and the internal oblique. It then enters the sheath of the rectus, turns forward, and becomes cutaneous between the umbilicus and the public symphysis. Its lateral cutaneous branch supplies the skin of the gluteal region and upper thigh as far down as the greater trochanter. Muscular branches are given to the transversus, oblique, and rectus muscles. Sensory twigs are given to the adjacent peritoneum. Autonomic Nerve Supply The parasympathetic supply of the thoracic viscera is furnished by the vagus nerves (see Chapter 52). Sympathetic Trunk and Ganglia. The sympathetic trunks enter the thorax from the neck, descend in front of the heads of the ribs and the posterior intercostal vessels and accompanying nerves, and enter the abdomen by piercing the crura of the diaphragm or by passing behind the medial arcuate ligaments (see Fig. 2–6). In the thorax, each trunk usually has 11 or 12 (occasionally 10 or 13) separate ganglia of varying size. As described earlier, the first thoracic ganglion is often fused with the inferior cervical to form the cervicothoracic or stellate ganglion. The trunk itself may be very slender between two adjacent ganglia; sometimes it is double. Each ganglion has one to four rami communicantes whose distinguishing characteristics were described under General Features. Preganglionic fibers for the thoracic and abdominal walls and back arise from all levels of the thoracic spinal cord and reach the sympathetic trunk by way of rami communicantes. The postganglionic fibers return to the spinal nerves by way of rami communicantes and are distributed by way of dorsal and ventral rami and the meningeal branches of the spinal nerves. The postganglionic fibers for thoracic and abdominal viscera are distributed by visceral branches. These include cardiac branches of the ansa subclavia, stellate ganglion, and upper four or five thoracic ganglia and pulmonary branches of the upper four or five thoracic ganglia. Scattered filaments of the trunk are also given to the aortic and esophageal plexuses. The major visceral branches, however, are the three splanchnic nerves. The greater splanchnic nerve is formed by three or four large roots and an inconstant number of smaller ones, all of which arise in the variable fashion from the trunk and ganglia; the upper and lower limits are usually the 5th and 10th. The nerve pierces the diaphragm or passes through its aortic opening and ends in the celiac ganglion and plexuses. It and the other splanchnic nerves contain sensory as well as preganglionic fibers, and relatively large splanchnic ganglion is usually present along the nerve near the diaphragm. The lesser splanchnic nerve, which may be absent, usually arises from the lower thoracic ganglia by one to the three rootlets. Descending slightly lateral to the greater

25

splanchnic nerve, it pierces the diaphragm and joins the aorticorenal ganglion and celiac plexus. The lowest splanchnic nerve, which also may be absent, usually arises from the lowest thoracic ganglion. Descending medial to the sympathetic trunk, it enters the abdomen and joins the aorticorenal ganglion and adjacent plexus. More detailed information about the sympathetic trunks, ganglia, and branches is available in various other sources.30,46,55 (See also references cited under General Features.) Prevertebral Plexuses. These are formed by the branches of the vagus nerves and sympathetic trunks that supply the thoracic viscera and blood vessels. The cardiac plexus may be considered as comprising inter-connected aortic arch, right and left coronary, and right and left atrial plexuses, all lying in adventitia or under epicardium.47 The sympathetic components reach the plexuses by way of cervical, cervicothoracic, and thoracic cardiac nerves (the preganglionic fibers arise from the upper four to five or six segments of the thoracic spinal cord). A variable number of cardiac ganglia occur along the cervicothoracic nerves. The pulmonary plexuses, which are usually described as anterior and posterior (with respect to the root of the lung), are essentially continuous with the cardiac plexus. Lying mostly posterior to the root of the lung and formed chiefly by the vagus nerves, the plexuses receive branches directly from the upper four to five or six segments of the thoracic spinal cord. The esophageal plexus is formed in a variable fashion by the vagus nerves after they leave the pulmonary plexuses. Lying in the fibrous wall of the esophagus, the plexus receives filaments from the sympathetic trunks and greater splanchnic nerves; the preganglionic fibers arise mainly from the lower segments of the thoracic spinal cord. The thoracic aortic plexus receives filaments from the vagus nerves and sympathetic trunks (the preganglionic fibers arise from the upper thoracic segments) that ramify in its adventitia and form a delicate plexus that continues for a short distance along the branches of the aorta. The plexus is also continuous through the aortic opening of the diaphragm with the abdominal aortic and celiac plexuses.

Abdomen, Pelvis, and Lower Limb Lumbar Plexus The dorsal rami of the lumbar nerves, which provide part of the nerve supply of the back, are described later. The ventral rami enter the psoas major muscle, where they combine in a variable fashion to form the lumber plexus (a division into anterior and posterior divisions that then combine, as occurs in the trunks of the brachial plexus, has been described but is difficult to demonstrate).

26

Structure of the Peripheral Nervous System

Within the muscle they are connected to the lumbar sympathetic trunk by rami communicates. Strictly speaking, the second to fourth nerves are usually (in about three fourths of instances) described as forming the lumbar plexus proper. However, the lower part of the fourth lumbar nerve and all of the fifth enter the sacral plexus (the combined trunk is known as the lumbosacral trunk or furcal nerve), and the two plexuses are commonly known as the lumbosacral plexus; the fourth lumbar nerve is then the one ventral ramus common to both plexuses. Moreover, the branches of the first lumbar nerve also are usually described with the lumbar plexus. The general features of the lumbar plexus are discussed further elsewhere (see Chapter 56).1,27,71 Variations. As in the case of the brachial plexus, prefixation and postfixation of the lumbrosacral plexus in the sense of complete shifts upward or downward are uncommon. Nevertheless, the plexus is often spoken of as prefixed when the upper level is the 11th or 12th thoracic nerve and postfixed when the lower border is the fifth sacral or first coccygeal nerve. The total range may therefore be from the 11th thoracic to the first coccygeal. The rami supplying the limbs, exclusive of the cutaneous branches of T12 and L1, can range from the first lumbar to the third sacral. Moreover, minor variations in pattern are common, and the plexuses are often asymmetrical with respect to right and left. Branches. Direct branches (L1 to L4) are given to the quadratus lumborum, psoas major, and psoas minor muscles. The first lumbar nerve has variable connections with the subcostal nerve and with the second lumbar and gives twigs to adjacent muscles. It resembles an intercostal nerve in giving off a collateral branch, the ilioninguinal nerve, and then continuing as the iliohypogastric nerve, which has lateral and anterior cutaneous branches.7 The terminal branches of the lumbar plexus are the lateral femoral cutaneous and the femoral nerves, which emerge laterally from the psoas major, the genitofemoral nerve, which emerges from its front, and the obturator and sometimes the accessory obturator nerves, which emerge medially. The iliohypogastric nerve emerges from the lateral side of the psoas major, runs behind the kidney, pierces the transversus muscle above the iliac crest, and divides into lateral and anterior cutaneous branches. The lateral branch supplies the skin over the side of the buttock, and the anterior branch the skin above the pubis (see Fig. 2–11). Muscular branches, if any, are probably sensory. A motor branch is occasionally given to the pyramidalis. The ilioinguinal nerve runs a similar course to the iliac crest. Here it pierces the transversus and internal oblique, accompanies the spermatic cord through the

inguinal canal, emerges from the superficial ring, and gives cutaneous branches to the thigh and anterior scrotal or anterior labial branches. The lateral femoral cutaneous nerve (L2, or L2 and L3, or L1 and L2) runs obliquely across the iliacus toward the anterior superior iliac spine and enters the thigh by passing behind the inguinal ligament. It divides into anterior and posterior branches that supply the skin of the anterior and lateral aspects of the thigh. The femoral nerve (chiefly L4, plus L2 and L3) is the largest branch of the lumbar plexus (Fig. 2–12). It descends between the psoas and iliacus and enters the thigh behind the middle of the inguinal ligament in the muscular compartment lateral to the femoral vessels. Entering the femoral triangle, it breaks up into a number of terminal branches. In the iliac fossa it supplies the iliacus and the femoral artery. The nerve to the pectineus arises here (or in the femoral triangle) and passes behind the femoral sheath to supply the pectineus and the hip joint. The terminal branches of the femoral nerve are sometimes classified into an anterior division (anterior cutaneous and a branch to the sartorius) and a posterior division (muscular and saphenous). The muscular branch of the anterior division goes directly to the sartorius. The anterior cutaneous branches of the same division are subdivided into the intermediate and medial cutaneous nerves. The intermediate nerve, which is usually double, gives branches to the sartorius and supplies the skin on the front of the thigh; distally it contributes to the patellar plexus. The medial nerves supply the skin on the medial side of the thigh and contribute to the subsartorial and patellar plexuses. The muscular branches of the posterior division supply the rectus femoris (a twig from this is given to the hip joint), the vastus medialis (a branch continues to the knee joint), the vastus intermedius (branches continue to the articularis genus and the knee joint), and the vastus lateralis (a branch continues to the knee joint). The saphenous nerve can be regarded as the termination of the femoral nerve. It descends with the femoral vessels through the femoral triangle and subsartorial canal, crosses the femoral artery from lateral to medial, and then becomes cutaneous. It descends in the leg with the large saphenous vein and supplies the skin on the medial side of the leg and foot. It gives a branch to the knee joint and contributes to the subsartorial and patellar plexuses. The saphenous nerve may be joined by filaments constituting an accessory femoral nerve. This is a common variant that arises from the lumbar plexus, runs a separate course into the thigh, and usually ends by joining one of the cutaneous branches of the femoral nerve. The subsartorial plexus, which lies in the adductor canal deep to the sartorius, is formed by branches from the medial femoral cutaneous, saphenous, and obturator nerves. The patellar plexus, which lies in front of the knee,

Gross Anatomy of the Peripheral Nervous System

27

FIGURE 2–12 The sequences of branches of the sciatic, tibial, and common peroneal nerves, based on studies by Sunderland and Hughes. The sequences of branches of the femoral and obturator nerves are based on studies by Pitres and Testut. (From Gardner, E., Gray, D. J., and O’Rahilly, R.: Anatomy, 4th ed. Philadelphia, W. B. Saunders, 1975, with permission.)

is formed by branches of the lateral, intermediate, and medial femoral cutaneous and saphenous nerves. The genitofemoral nerve (L2, or L1 and L2, occasionally L3) descends on the front of the psoas major and divides into genital and femoral branches. The genital branch enters the inguinal canal through the deep ring, supplies the cremaster, and continues to supply the scrotum or labium majus and the adjacent part of the thigh. The femoral branch enters the femoral sheath, lateral to the artery, and turns forward to supply the skin of the femoral triangle.

The obturator nerve (L3, L4, sometimes L2 or L5 also) emerges from the medial side of the psoas major at the inlet of the pelvis (see Fig. 2–12). It runs downward and forward on the lateral wall of the pelvis to the obturator groove, where it divides into anterior and posterior branches. These pass through the obturator foramen into the thigh, where they are separated by the adductor brevis. Branches to the hip joint arise from the trunk and from one or both branches. The anterior branch lies in front of the adductor brevis, behind the pectineus and adductor longus. It descends along the latter muscle;

28

Structure of the Peripheral Nervous System

supplies it and the gracilis, adductor brevis, and sometimes pectineus also; and ends as a filament to the femoral artery and subsartorial plexus; twigs from it supply the overlying skin and occasionally the knee joint. The posterior branch pierces the obturator externus and descends behind the adductor brevis, in front of the adductor magnus. It then pierces the adductor magnus and accompanies the popliteal artery to the back of the knee joint. It supplies the obturator externus, adductor magnus, and sometimes the adductor brevis. The accessory obturator nerve (L3, L4), which is persent in nearly 10% of cases, descends medial to the psoas and enters the thigh deep to the pectineus.75 It supplies the hip joint and pectineus. Sacral Plexus The dorsal rami of the sacral nerves are described with the back. The ventral rami of the first four sacral nerves emerge from the sacral canal though the pelvic sacral foramina. The fourth then divides into upper and lower divisions; the upper division and the first three ventral rami combine with the lumbosacral trunk to form the sacral plexus. Anterior and posterior divisions of the first three rami have been described, but, as in the case of the lumbar plexus, they are difficult to demonstrate. Each ramus contributing to the sacral plexus is connected to a single ganglion of the sacral sympathetic trunk by one or more rami communicates. The sacral plexus lies in front of the piriformis muscle, separated from the internal iliac vessels and ureter anteriorly by the parietal pelvic fascia. From the complex and variable union of the ventral rami, 12 named branches arise. Five supply pelvic strucures; the remainder are distributed to the buttock and lower limb. Branches to Pelvic Structures. The chief branch is the pudendal nerve (S2, S3, S4), which supplies most of the perineum. It passes through the greater sciatic notch below the piriformis and, after crossing the back of the ischial spine, enters the perineum through the lesser sciatic notch, accompanied by the internal pudendal artery. Running forward in the pudendal canal in the ischiorectal fossa, it gives off the inferior rectal nerve and then divides into the perineal nerve and the dorsal nerve and then divides into the perineal nerve and the dorsal nerve of the penis or clitoris. The inferior rectal nerve (which may arise separately from S3 and S4 of the sacral plexus) pierces the medial wall of the pudendal canal and divides into branches that traverse the ischiorectal fossa. They supply the skin around the anus, the sphincter ani externus, and the anal mucosa as far upward as the pectinate line. The perineal nerve divides into superficial and deep branches. The superficial branch gives off two posterior scrotal or labial nerves (medial and lateral), which pierce the superficial and deep perineal fasciae and

run forward to supply the scrotum or labium majus. The deep branch of the perineal nerve gives twigs to the sphincter ani externus and levator ani and then enters the superficial perineal space to supply the bulbospongiosus, ischiocavernosus, superficial transversus perinei, and bulb of the penis. The dorsal nerve of the penis or clitoris, the other terminal branch of the pudendal nerve, pierces the posterior edge of the urogenital diaphragm. As it runs forward, it supplies the deep transversus perinei and sphincter urethrae muscles. It then pierces the inferior fascia of the diaphragm, gives a branch to the corpus cavernosum penis or clitoris, and then runs forward on the dorsum of the penis or clitoris to supply the skin, prepuce, and glans. In addition to the pudendal nerve, the sacral plexus gives the following additional branches to pelvic structures: the nerve to the piriformis (S1, S2), the nerve to the levator ani and coccugeus (S3, S4), the nerve to the sphincter ani externus (perineal branch of S4), and the pelvid splanchnic nerves (S2, S3 and S4, S5), which contain parasympathetic preganglionic as well as sensory fibers and which enter the inferior hypogastric plexus. Branches to the Buttock and Lower Limb. These are seven in number. The superior gluteal nerve (L4, L5, S1) passes backward through the greater sciatic notch, above the piriformis. An upper branch supplies the gluteus medius, and a lower branch supplies the gluteus minimus, tensor fasciae latae, and hip joint. The inferior gluteal nerve (L5, S1, S2) passes through the greater sciatic foramen below the piriformis and supplies the gluteus maximus. The nerve to the obturator internus (L5, S1, S2) leaves the pelvis below the piriformis, supplies the superior gemellus, and then passes through the lesser sciatic foramen to the obturator internus. The nerve to the quadratus femoris (L4, L5, S1) leaves the pelvis below the piriformis in front of the sciatic nerve. It supplies the inferior gemellus, the quadratus femoris, and the hip joint. The posterior femoral cutaneous nerve (S1 to S3) leaves the pelvis below the piriformis. It descends in company with the sciatic nerve, becomes superficial near the popliteal fossa, and accompanies the small saphenous vein to the middle of the calf. Its branches are inferior clunial nerves (gluteal branches) to the skin of the buttock, perineal branches to the skin of the genitalia, and femoral and sural branches to the skin on the back of the thigh and calf. The perforating cutaneous (inferior medial clunial) nerve (S2, S3) pierces the sacrotuberous ligament and supplies the skin over the lower part of the buttock. The sciatic nerve (L4, L5, S1 to S3), the largest nerve in the body, consists of peroneal and tibial parts, which are usually bound together, leaving the pelvis through the greater sciatic foramen, below the piriformis (see Fig. 2–12).2 Sometimes they leave separately, the peroneal portion piercing the piriformis and the tibial portion passing

Gross Anatomy of the Peripheral Nervous System

below it. The exit of the sciatic nerve from the pelvis is indicated by the superior point of trisection of a line from the posterior superior iliac spine to the ischial tuberosity. Its downward course is indicated by a line down the middle of the back of the thigh (from the midpoint of a line between the greater trochanter and ischial tuberosity). The sciatic nerve descends under cover of the gluteus maximus, between the greater trochanter and ischial tuberosity. In the thigh it lies anteriorly on the adductor magnus and is accompanied by the posterior femoral cutaneous nerve and the companion artery. In the lower third of the thigh the nerve separates into its two components, the tibial and common peroneal nerves (the separation may occur at any level in the gluteal region or thigh). Its branches arise mostly on the medial side and supply the semitendinosus, semimembranosus, long head of the biceps, adductor magnus (all by the tibial nerve), and short head of the biceps (by the common peroneal nerve). The tibial (medial popliteal) nerve (L4 to S3) descends separately through the popliteal fossa (see Fig. 2–12). It then lies on the popliteus muscle, under cover of the gastrocnemius, and at the lower border of the popliteus passes deep to the fibrous arch of the soleus to reach the back of the leg. Here it descends first on the tibialis posterior and flexor digitorum longus and then on the tibia. Then, becoming more superficial and crossing the posterior tibial artery posteriorly to gain its lateral side, it ends by dividing into medial and lateral plantar nerves under cover of the flexor retinaculum. In the thigh, muscular branches arise as listed with the sciatic nerve. In the popliteal fossa, branches are given to the knee joint and muscular branches to the gastrocnemius, soleus, plantaris, popliteus, and tibialis posterior. A branch of the nerve to the popliteus, the interosseous nerve of the leg, descends on the interosseous membrane. The medial sural cutaneous nerve joins the peroneal communicating branch of the common peroneal nerve to form the sural nerve. The sural nerve descends in company with the small saphenous vein, gives lateral calcaneal branches to the skin of the back of the leg and lateral aspect of the foot and heel, gives twigs to the ankle joints, and continues forward along the lateral side of the little toe as the lateral dorsal cutaneous nerve. In the leg the tibial nerve gives muscular branches to the soleus, tibialis posterior, flexor hallucis longus, and flexor digitorum longus. Medial calcaneal branches supply the skin of the heel and sole, and a twig is given to the ankle joint. The course of the tibial nerve in the leg is indicated on the surface by a line from about the level of the tibial tuberosity downward to the midpoint between the medial malleolus and the heel. The medial plantar nerve, the larger of the two terminal branches of the tibial nerve, at first lies deep to the abductor hallucis. It then runs forward in the sole between the abductor and the flexor digitorum brevis. It supplies these muscles and the skin on the medial side of the sole.

29

Its terminal branches are four plantar digital nerves for muscles (flexor hallucis and first lumbrical) and for the medial side of the big toe and the adjacent sides of the first and second, second and third, and third and fourth toes. The nerves extend on to the dorsum and supply the nail beds and tips of the toes. The lateral plantar nerve runs forward and laterally between the quadratus plantae and the flexor digitorum brevis and divides into superficial and deep branches. During its course it supplies the quadratus plantae and abductor digiti minimi and the skin of the lateral side of the sole. The superficial branch supplies the flexor digiti minimi brevis and the lateral side of the sole and little toe and, by plantar digital nerves, the adjacent sides of the fourth and fifth toes. The deep branch turns medially; supplies the interossei, the second, third, and fourth lumbricals, and the adductor hallucis; and gives off articular twigs. The common peroneal (lateral popliteal) nerve (L4 to S2) descends through the popliteal fossa, following the medial edge of the biceps closely (see Fig. 2–12). It crosses the lateral head of the biceps, gains the back of the head of the fibula, and winds around the neck of that bone (where it is often palpable, and where it is susceptible to injury) under cover of the peroneus longus. Here it divides into its terminal branches, the superficial and deep peroneal nerves. While a part of the sciatic nerve, it supplies the short head of the biceps and sometimes the knee joint also. In the popliteal fossa, it supplies the knee joint and gives rise to a branch that divides into the lateral sural cutaneous nerve (for the skin on the lateral side of the leg) and the peroneal communicating branch (which joins the medial sural cutaneous nerve to form the sural nerve). At the neck of the fibula, it gives off a small recurrent branch that supplies the knee and tibiofibular joints and the tibialis anterior. The common peroneal nerve sometimes supplies the peroneus longus or extensor digitorum longus or both. The superficial peroneal (musculocutaneous) nerve, one of the two terminal branches of the common peroneal nerve, descends in front of the fibula, between the peronei and the extensor digitorum longus. Muscular branches supply the peroneus longus and peroneus brevis; the branch to the latter is often prolonged to the extensor digitorum brevis and adjacent joints and is termed the accessory deep peroneal nerve.42,73 In the lower part of the leg, the superficial peroneal nerve divides into medial and intermediate dorsal cutaneous nerves. These pass in front of the extensor retinacula, the medial branch supplying the skin and joints of the medial side of the big toe and (by dorsal digital nerves) the adjacent sides of the second and third toes; the intermediate nerve (by dorsal digital nerves) supplies the adjacent sides of the third and fourth and the fourth and fifth toes. As in the hand, the territories of distribution of the cutaneous nerves of the foot show considerable variation in size and overlap and reciprocal changes in size.33

30

Structure of the Peripheral Nervous System

The deep peroneal nerve continues the winding course of the common peroneal nerve around the neck of the fibula, then pierces the anterior intermuscular septum and extensor digitorum longus, and descends on the interosseous membrane. It meets the anterior tibial artery, and both pass deep to the extensor retinacula. Branches are given to the tibialis anterior, extensor hallucis longus, extensor digitorum longus, peroneus tertius, and ankle joint. In the foot, where it lies about midway between the malleoli, the nerve divides into its terminal branches, medial and lateral. The medial branch gives dorsal digital nerves for the adjacent sides of the first and second toes, and the lateral branch supplies the extensor digitorum brevis and adjacent joints. It may also send twigs (probably afferent) to the first three dorsal interossei. Coccygeal Plexus The ventral ramus of the fifth sacral nerve enters the pelvis between the sacrum and coccyx; that of the coccygeal nerve passes forward below the rudimentary transverse process of the first piece of the coccyx. The coccygeal (or sacrococcygeal) plexus is formed by these two ventral rami, together with the lower division of the ventral ramus of the fourth sacral nerve.59 The plexus supplies the coccyx, the sacrococcygeal joint, and the skin over the coccyx. Autonomic Nerve Supply The parasympathetic supply of the abdominal viscera is furnished chiefly by the vagus nerves. The descending and sigmoid portions of the colon and the pelvic viscera are supplied by the sacral parasympathetics. Details of the relationship between the vagal nerve fibers, the sympathetic nerve fibers, and the nerve cell bodies and fibers of the enteric nerve plexuses are given in Chapter 12. Sympathetic Trunk and Ganglia. The two trunks enter the abdomen by piercing the diaphragm or by passing behind the medial arcuate ligaments (see Fig. 2–6). In descending on the vertebral column, adjacent to the psoas major muscles, the right trunk lies behind the inferior vena cava, the left one beside the aorta. The trunks continue into the pelvis, where they lie on the pelvic surface of the sacrum, medial to the upper three pelvic sacral foramina and usually in front of the fourth. They end by uniting in front of the coccyx to form an enlargement, the ganglion impar. In the lumbar region, the two trunks are seldom symmetrical and the ganglia are irregular in size, number, and position. There are usually four or five ganglia (three or four, according to Webber70), but there may be from two to six, and occasionally a trunk is an elongated ganglionic mass. The variations are such as to make identification of the proper level of a specific ganglion very difficult. Nor can counting from the highest lumbar ganglion be depended upon. When the first lumbar ganglion is

present, it lies between the crus of the diaphragm and the vertebral column, is difficult to reach, and is often overlooked. Rami communicantes are a better means of identification. In the pelvis, the number of ganglia is variable, but there are usually three or four. Each lumbar ganglion has two or more rami communicantes to two or more spinal nerves. The lowest ramus at upper levels contains the most preganglionic fibers and is the clue to the identification of a ganglion. For example, the second lumbar ganglion is connected to the first and second lumbar nerves; hence the lowest ramus (from the second nerve) leads to the second ganglion. Nevertheless, the identification of ganglia during surgical procedures remains an uncertain and difficult task. The second lumbar nerve is usually the lowest one to contain preganglionic sympathetic fibers. Each sacral ganglion tends to be connected by rami communicantes with only one spinal nerve. The postganglionic fibers in the lumbar and sacral rami communicantes are for the supply of the lower part of the abdominal wall and back; the anal canal, perineum, and external genitalia; and the lower limb by way of the lumbosacral and coccygeal plexuses. The preganglionic fibers for the lower limbs arise from the lower thoracic and upper lumbar levels of the spinal cord and descend in the trunks to the lumbar and sacral ganglia, where they synapse. The preganglionic fibers for abdominal viscera descend mostly in the thoracic splanchnic nerves. Some, however, descend in the sympathetic trunks and leave by way of visceral branches. These consist of four or more lumbar splanchnic nerves of variable size, which, depending upon level, join the celiac or intermesenteric and adjacent plexuses or the superior hypogastric plexus.39 These nerves also contain preganglionic and sensory fibers, and some of their postganglionic fibers reach pelvic viscera by way of the hypogastric plexuses, and the iliac fossa and upper thigh by way of the aortic plexus. The preganglionic fibers for pelvic viscera arise in the upper lumbar levels of the spinal cord. They descend and synapse in lumbar and sacral ganglia. The postganglionic fibers reach pelvic viscera by way of rami communicantes or lumbar splanchnic nerves (just described) and by way of sacral splanchnic nerves. These last are a variable number of fine visceral branches of the sacral sympathetic trunks that join the inferior hypogastric plexus. More detailed information about the sympathetic trunks and ganglia is available in the references cited under General Features and elsewhere.41,55 Prevertebral Plexuses. In the abdomen, these are best considered as a single great plexus formed by the splanchnic nerves, branches of both vagus nerves, and masses of ganglion cells, with various parts named

31

Gross Anatomy of the Peripheral Nervous System

according to the arteries with which they are associated. The plexus lies in front of the upper part of the abdominal aorta. Its chief ganglia are the irregularly shaped celiac ganglia at the origin of the celiac trunk, each lying on the corresponding crus of the diaphragm. The inferolateral extensions of the ganglia are termed the aorticorenal ganglia. The superior mesenteric ganglion (or ganglia) is usually fused with the celiac ganglia. Smaller ganglia, for example, phrenic and renal, are often found along the smaller subsidiary plexuses. The celiac and superior mesenteric plexuses lie on the front and sides of the aorta at the origins of the celiac trunk and the superior mesenteric and renal arteries. They contain the celiac, aorticorenal, and superior mesenteric ganglia and many smaller unnamed masses. Branches of the plexuses extend along arteries and are named accordingly—for example, hepatic, gastric, phrenic, suprarenal, renal, ureteric, and testicular or ovarian. Downward continuations of the prevertebral plexus form the aortic and inferior mesenteric plexuses. The aortic plexus consists of a variable number of interconnected strands that, as they descend along the aorta, receive branches from the lumbar splanchnic nerves. The part of the plexus between the origins of the superior and inferior mesenteric arteries is also known as the intermesenteric plexus, and below the bifurcation of the aorta the aortic plexus becomes the superior hypogastric plexus. Some filaments from the aortic plexus accompany the lumbar arteries and provide a pathway for postganglionic fibers to the abdominal wall and back. Other filaments, together with branches of the lumbar splanchnic nerves, form a plexus along the common and external iliac arteries. This plexus is reinforced by a branch of the femoral nerve and continues into the thigh along the femoral artery. The inferior mesenteric plexus is a continuation of the aortic plexus along the inferior mesenteric plexus. It contains one or more inferior mesenteric ganglia and it forms the superior rectal plexus, which supplies the rectum with sympathetic and sensory fibers. In the pelvis, in front of the fifth lumbar vertebra, the aortic plexus becomes the superior hypogastric plexus (or presacral nerve). This then divides in front of the sacrum into two elongated narrow networks termed the right and left hypogastric nerves. Each of these nerves descends on the side of the rectum (or rectum and vagina) and is joined by the pelvic splanchnic nerves to form the right and left inferior hypogastric (or pelvic) plexuses. Each plexus contains small, scattered pelvic ganglia, and each is joined by the sacral splanchnic nerves from the sympathetic trunk. Many branches of the inferior hypogastric plexus supply the rectum, with some of them forming a middle rectal plexus. Large portions of the inferior hypogastric plexuses form either the prostatic plexus (which continues forward as the cavernous nerves of the penis and from

which is derived the vesical plexus) or the uterovaginal plexus (fibers from the lowermost part of this plexus continue as the cavernous nerves of the clitoris). The inferior hypogastric plexuses carry sensory, postganglionic sympathetic, and preganglionic parasympathetic fibers. Some of the preganglionic parasympathetic fibers supply the descending and sigmoid colonic segments, usually by a single ascending branch of the hypogastric nerve.74 Some of the sensory fibers are pain fibers that reach the spinal cord by way of the superior hypogastric plexuses and lumbar splanchnic nerves as well as by way of pelvic splanchnic nerves. The other sensory fibers, concerned with reflexes and visceral sensations, reach the spinal cord by way of the pelvic splanchnic nerves.

Back The nerve supply of the back is provided by the meningeal branches and dorsal rami of spinal nerves. Each spinal nerve gives off a meningeal branch (or sinuvertebral nerve), which is often connected with adjacent sympathetic ganglia and which reenters the vertebral canal and supplies dura mater, posterior longitudinal ligament, periosteum, and epidural and intraosseous blood vessels.52 The meningeal branches of the upper three cervical nerves give off branches that ascend through the foramen magnum and supply the dura mater of the floor of the posterior cranial fossa.37 The dorsal rami arise after the spinal nerves emerge from the intervertebral foramina (sacral nerves divide within the sacrum), and the dorsal rami pass backward to supply the muscles, bones, joints, and skin of the back. Most divide into medial and lateral branches, which descend as they run dorsally, each supplying muscles and each anastomosing with nerves above and below to form a plexus in the muscles. In the upper half of the trunk the medial branches supply skin, whereas in the lower half the lateral branches become cutaneous. However, the level of shift is variable (see Thoracic Dorsal Rami below). Cervical Dorsal Rami The connecting loops between the dorsal rami of the first three or four cervical nerves form what is termed the posterior cervical plexus,4 which gives branches to adjacent muscles. The dorsal ramus of the first cervical nerve, which usually lacks cutaneous fibers, is known as the suboccipital nerve. It emerges above the posterior arch of the atlas, below the vertebral artery, and supplies the semispinalis capitis and the muscles of the suboccipital triangle. The dorsal ramus of the second cervical nerve supplies the obliquus capitis inferior and then divides into medial and lateral branches. The medial branch, termed the greater occipital nerve, pierces the semispinalis capitis and trapezius and then accompanies the occipital artery and supplies the skin of the scalp as far forward as the vertex.

32

Structure of the Peripheral Nervous System

The medial branch of the dorsal ramus of the third cervical nerve continues as the third occipital nerve, piercing the trapezius and supplying the skin on the back of the head. The dorsal rami of C6, C7, and C8 usually have no cutaneous branches4,31,51; hence the C5 dermatome is adjacent to T1 and, with overlap, C4 meets T2. This junction marks what might be called the dorsal axial line (from about the spine of C7 to near the deltoid insertion). Thoracic Dorsal Rami Each ramus passes backward, supplying the deeply placed muscles, and divides into a medial and a lateral cutaneous branch, which are separated by slips of the longissimus thoracis muscle. The medial branches pass backward and downward, supplying the erector spinae and its divisions. The medial branches of the upper thoracic nerves (T1 to T3) also become cutaneous. The lateral branches supply the levatores costarum, the longissimus thoracis, and the iliocostalis thoracis. They have a long downward course,31 the lower ones (T9 to T12) piercing the latissimus dorsi and supplying the skin of the back as far down as the gluteal region. In what might be termed a transitional zone, both medial and lateral branches of T4 to T8 give cutaneous twigs. Lumbar, Sacral, and Coccygeal Dorsal Rami The lateral branches of the upper lumbar rami give rise to the superior clunial nerves, which supply the skin of the buttock. The lateral branches of the lower lumbar rami, together with those of the sacral dorsal rami, form the dorsal sacral plexus. The dorsal rami of the first four sacral nerves pass backward through the dorsal sacral foramina, whereas those of the fifth sacral and the coccygeal nerves emerge through the sacral hiatus. The medial branches of the first dorsal rami supply the erector spinae. The lateral branches, together with those of the lower lumbar and a contribution from the fifth sacral, form the dorsal sacral plexus immediately behind the sacrum and coccyx.27 Loops of this plexus give off two or three middle clunial nerves (though not invariably), which pierce the overlying gluteus maximus and supply the skin of the buttock. The dorsal rami of the fifth sacral and coccygeal nerves lack medial and lateral branches; they communicate (often forming a single nerve) and supply adjacent ligaments and overlying skin.

REFERENCES 1. Bardeen, C. R., and Elting, A. W.: A statistical study of the variations in the formation and position of the lumbo-sacral plexus in man. Anat. Anz. 19:124, 209, 1901. 2. Beaton, L. E., and Anson, B. J.: The relation of the sciatic nerve and of its subdivisions to the piriformis muscle. Anat. Rec. 70:1, 1937.

3. Cave, A. J. E.: The distribution of the first intercostal nerve and its relation to the first rib. J. Anat. 63:367, 1929. 4. Cave, A. J. E.: The innervation and morphology of the cervical intertransverse muscles. J. Anat. 71:497, 1937. 5. Dankmeijer, J., and Waltman, J. M.: Sur I’innervation de la face dorsale des doigts humains. Acta Anat. 10:377, 1950. 6. Dass, R.: Sympathetic components of the dorsal primary divisions of human spinal nerves. Anat. Rec. 113:493, 1952. 7. Davies, F.: A note on the first lumbar nerve (anterior ramus). J. Anat. 70:177, 1935. 8. Davies, F., and Laird, M.: The supinator muscle and the deep radial (posterior interosseous) nerve. Anat. Rec. 101:243, 1948. 9. Delmas, J., and Laux, G.: Système Nerveux Sympathique. Paris, Masson et Cie, 1952. 10. Dow, D. R.: The anatomy of rudimentary first thoracic ribs with special reference to the arrangement of the brachial plexus. J. Anat. 59:166, 1925. 11. Fenart, R.: La morphogénése du plexus brachial, ses rapports avec la formation du cou et du membre supérieur. Acta Anat. 32:322, 1958. 12. Fender, F. A.: Foerster’s scheme of the dermatomes. Arch. Neurol. Psychiatry 41:688, 1939. 13. Foerster, O.: The dermatomes in man. Brain 56:1, 1933. 14. Foerster, O.: Symptomatologie der Erkankungen des Rücken-marks and seiner Wurzeln. In Bumke, I., and Foerster, O. (eds.): Handbuch der Neurologie, Vol. 5. Berlin, Springer, 1936. 15. Furness, J. B., and Costa, M.: Types of nerves in the enteric nervous system. Neuroscience 5:1, 1980. 16. Gabella, G.: Structure of the Autonomic Nervous System. London, Chapman and Hall, 1976. 17. Gardner, E.: Surgical anatomy of the external carotid plexus. Arch. Surg. 46:238, 1943. 18. Gardner, E., Gray, D. J., and O’Rahilly, R.: Anatomy, 4th ed. Philadelphia, W. B. Saunders, 1975. 19. Gershon, M. D.: The enteric nervous system. Annu. Rev. Neurosci. 4:227, 1981. 20. Hansen, K., and Schliack, H.: Segmentale Innervation. Stuttgart, Thieme, 1962. 21. Harness, D., and Sekeles, E.: The double anastomotic innervation of thenar muscles. J. Anat. 109:461, 1971. 22. Harris, W.: The true form of the brachial plexus, and its motor distribution. J. Anat. 38:399, 1904. 23. Harris, W.: The Morphology of the Brachial Plexus. London, Humphrey Milford, 1939. 24. Haymaker, W., and Woodhall, B.: Peripheral Nerve Injuries, 2nd ed. Philadelphia, W. B. Saunders, 1953. 25. Head, H., and Campbell, A. W.: The pathology of herpes zoster and its bearing on sensory localisation. Brain 23:353, 1900. 26. Hollinshead, H.: Anatomy for Surgeons, 2nd ed. New York, Hoeber Medical Division, Harper & Row, 1968, 1969, 1971. 27. Horwitz, M. T.: The anatomy of (A), the lumbosacral nerve plexus—its relation to variations of vertebral segmentation, and (B), the posterior sacral nerve plexus. Anat. Rec. 74:91, 1939. 28. Hovelacque, A.: Anatomie des Nerfs Craniens et Rachidiens et du Système Grande Sympathique Chez L’homme. Paris, Gaston Doin et Cie, 1927.

Gross Anatomy of the Peripheral Nervous System 29. Jamieson, R. W., Smith, D. B., and Anson, B. J.: The cervical sympathetic ganglia. Bull. Northwest. Univ. Med. Sch. 26:219, 1952. 30. Jit, I., and Mukerjee, R. N.: Observations on the anatomy of the human thoracic sympathetic chain and its ganglia; with an anatomical assessment of operations for hypertension. J. Anat. Soc. India 9:55, 1960. 31. Johnston, H. M.: The cutaneous branches of the posterior primary divisions of the spinal nerves, and their distribution in the skin. J. Anat. 43:80, 1908. 32. Jones, F. W.: The Principles of Anatomy as Seen in the Hand, 2nd ed. London, Baillière, Tindall, & Cox, 1942. 33. Jones, F. W.: Structure and Function as Seen in the Foot, 2nd ed. London, Baillière, Tindall, & Cox, 1942. 34. Kaplan, E. B.: Functional and Surgical Anatomy of the Hand, 2nd ed. Philadelphia, J. B. Lippincott, 1965. 35. Keegan, J. J., and Garrett, F. D.: The segmental distribution of the cutaneous nerves in the limbs of man. Anat. Rec. 102:409, 1948. 36. Kerr, A. T.: The brachial plexus of nerves in man, the variations in its formation and branches. Am. J. Anat. 23:285, 1918. 37. Kimmel, D. L.: Innervation of spinal dura mater and dura mater of the posterior cranial fossa. Neurology (Minneap.) 11:800, 1961. 38. Kuntz, A.: The Autonomic Nervous System, 4th ed. Philadelphia, Lea & Febiger, 1953. 39. Kuntz, A.: Components of splanchnic and intermesenteric nerves. J. Comp. Neurol. 105:251, 1956. 40. Kuntz, A., Hoffman, H. H., and Napolitano, L. M.: Cephalic sympathetic nerves. Arch. Surg. 75:108, 1957. 41. Labbok, A.: Anatomische Untersuchungen and Typen des Kreuzabschnittes der Trunci sympathici. Anat. Anz. 85:14, 1937. 42. Lambert, E. H.: The accessory deep peroneal nerve. Neurology (Minneap.) 19:1169, 1969. 43. Langley, J. N.: The Autonomic Nervous System. Part 1. Cambridge, UK, Heffer, 1926. 44. Last, R. J.: Innervation of the limbs. J. Bone Joint Surg. Br. 31:452, 1949. 45. Learmonth, J. R.: A variation in the distribution of the radial branch of the musculo-spinal nerve. J. Anat. 53:371, 1919. 46. Mitchell, G. A. G.: Anatomy of the Autonomic Nervous System. Edinburgh, E. & S. Livingstone, 1953. 47. Mizeres, N. J.: The cardiac plexus in man. Am. J. Anat. 112:141, 1963. 48. Monro, P. A. G.: Sympathectomy. London, Oxford University Press, 1959. 49. Nomina Anatomica, 4th ed. Amsterdam, Excerpta Medica Foundation, 1977. 50. Pallie, W.: The intersegmental anastomoses of posterior spinal rootlets and their significance. J. Neurosurg. 16:188, 1959. 51. Pearson, A. A, Sauter, R. W., and Buckley, T. F.: Further observations on the cutaneous branches of the dorsal primary rami of the spinal nerves. Am. J. Anat. 118:891, 1966. 52. Pedersen, H. E., Blunck, C. F. J., and Gardner, E.: The anatomy of lumbosacral posterior rami and meningeal

53.

54. 55. 56. 57. 58. 59. 60. 61. 62. 63.

64. 65.

66.

67.

68.

69. 70. 71. 72. 73. 74. 75. 76.

33

branches of spinal nerves (sinu-vertebral nerves). J. Bone Joint Surg. Am. 38:377, 1956. Pernkopf, E.: Atlas of Topographical and Applied Human Anatomy (ed., H. Ferner; transl., H. Monsen). Philadelphia, W. B. Saunders, 1963. Pick, J.: The Autonomic Nervous System. Philadelphia, J. B. Lippincott, 1970. Pick, J., and Sheehan, D.: Sympathetic rami in man. J. Anat. 80:12, 1946. Pitres, A., and Testut, L.: Les Nerfs en Schémas. Paris, Gaston Doin et Cie, 1952. Schadé, J. P.: The Peripheral Nervous System. Amsterdam, Elsevier, 1966. Seddon, H.: Surgical Disorders of the Peripheral Nerves. Edinburgh, Churchill Livingstone, 1972. Sicard, A., and Bruézière, J.: Le plexus sacro-coccygien. Arch. Anat. Histol. Embryol. 33:43, 1950. Skoog, T.: Ganglia in the communicating rami of the cervical sympathetic trunk. Lancet 2:457, 1947. Smorto, M. P., and Basmajian, J. V.: Clinical Electroneurography. Baltimore, Williams & Wilkins, 1972. Stopford, J. S. B.: The variation in distribution of the cutaneous nerves of the hand and digits. J. Anat. 53:14, 1918. Sunderland, S.: Metrical and non-metrical features of the muscular branches of the radial nerve. J. Comp. Neurol. 85:93, 1946. Sunderland, S.: Nerves and Nerve Injuries, 2nd ed. New York, Churchill Livingstone, 1978. Sunderland, S., and Bedbrook, G. M.: The relative sympathetic contribution to individual roots of the brachial plexus in man. Brain 72:297, 1949. Sunderland, S., and Hughes, E. S. R.: Metrical and nonmetrical features of the muscular branches of the sciatic nerve and its medial and lateral popliteal divisions. J. Comp. Neurol. 85:205, 1946. Sunderland, S., and Hughes, E. S. R.: Metrical and nonmetrical features of the muscular branches of the ulnar nerve. J. Comp. Neurol. 85:113, 1946. Sunderland, S., and Ray, L. J.: Metrical and non-metrical features of the muscular branches of the median nerve. J. Comp. Neurol. 85:191, 1946. von Lanz, T., and Wachsmuth, W.: Praktische Anatomie. (Some parts in 2nd edition.) Berlin, J. Springer, 1953–1972. Webber, R. H.: A contribution on the sympathetic nerves in the lumbar region. Anat. Rec. 130:581, 1958. Webber, R. H.: Some variations in the lumbar plexus of nerves in man. Acta Anat. 44:336, 1961. Williams, P. L, and Warwick, R. (eds.): Gray’s Anatomy, 36th ed. Philadelphia, W. B. Saunders, 1980. Winckler, G.: Le nerf péronier accessoire profond. Arch. Anat. Histol. Embryol. 18:181, 1934. Woodburne, R. T.: The sacral parasympathetic innervation of the colon. Anat. Rec. 124:67, 1956. Woodburne, R. T.: The accessory obturator nerve and the innervation of the pectineus muscle. Anat. Rec. 136:367, 1960. Wrete, M.: Die intermediären vegetativen Ganglien der Lum-balregion beim Menschen. Z. Mikrosk. Anat. Forsch. 53:122, 1943.

3 Microscopic Anatomy of the Peripheral Nervous System C.-H. BERTHOLD, JOHN P. FRAHER, R. H. M. KING, AND MARTIN RYDMARK

Nerve Trunks and Spinal Roots Connective Tissues Endoneurium Perineurium Epineurium

Node of Ranvier The Constricted Axon Segment The Node–Paranodal Apparatus The Schwann Cell The Myelin Sheath

Myelinated Nerve Fibers

Unmyelinated Nerve Fibers

General Organization of Myelinated Axons Periaxonal Space Axolemma Axoplasm The Paranode-Node-Paranode Region Internodal End Region

Historical Background Development General Organization Relationship of Unmyelinated Axons with Schwann Cells Spatial Relationships between Axons, Schwann Cells, and Collagen

Nerve Trunks and Spinal Roots R. H. M. King The nerve fibers comprising the peripheral nerves are collected together in bundles or fascicles. A nerve fiber is defined as an axon plus its associated Schwann cells. Each fascicle is delineated by a perineurial sheath composed of flattened squamous cells connected together by tight junctions. The individual fascicles are embedded in a collagenous matrix or epineurium. The connective tissues of peripheral nerves were described by Ranvier332 and Key and Retzius.216 The latter suggested that the nerve was divided into perifascicular connective tissues, lamellated sheath, and intrafascicular tissue, which they named epineurium, perineurium, and endoneurium, respectively (Fig. 3–1). This remains the most satisfactory and convenient subdivision and is in universal use today. Unlike epineurium and endoneurium, however, the term perineurium does not just

Morphometry Regeneration

The CNS-PNS Transitional Zone General Form Fiber Transition Tissue Composition Historical Background Development Comparative Anatomy Blood Supply Mechanical Features

refer to the collagenous connective tissue but also includes the cellular component. The perineurium is part of the membranous covering that ensheaths the whole of the nervous system (Fig. 3–2). Having said that, many authors use the term endoneurium to refer to the contents of the fascicle within the perineurium, including Schwann cells and axons, and not just to the connective tissue (e.g., Dyck et al.,99 Midroni et al.270). The nerve fascicles may branch so that any particular nerve fiber may change from one fascicle to another and may change association with fibers of other types. Electrophysiologic studies of sensory median nerve fascicles have shown that there is intrafascicular segregation by modality of both myelinated and unmyelinated fibers. These studies suggested that some Schwann cells contain only afferent C fibers while others contain efferent sympathetic fibers.177 This may result from alterations of Schwann cell proteins to reflect the type of axon with which they are associated. This segregation may improve the efficiency of regeneration because axons could make useful contact with similar end organs even if they were not identical to the 35

36

Structure of the Peripheral Nervous System

original one. Earlier studies, however, showed that some axons in a Remak fiber (unit of Schwann cell and unmyelinated axons) contain dense-cored vesicles while others do not. This observation implies that adrenergic axons can share the same Schwann cell as nonadrenergic ones.389 Axons conduct electrical impulses at rates varying between 1 and 100 m/s. Saltatory conduction in myelinated fibers, whereby the nerve impulse jumps from one node of Ranvier (see Node of Ranvier later) to the next, results in the faster transmission speeds, whereas continuous transmission of the nerve impulse, found in unmyelinated axons, occurs at the lowest velocities. In addition to propagation of electrical impulses, axons also have chemical conduction mechanisms. There are fast axonal anterograde and retrograde transport (200 to 400 mm/day) systems that depend on the tubulin and intermediate filament cytoskeleton. In addition, there are much slower transport systems whereby the components of this network are transported at rates of about 1 to 2 mm/day. These are discussed in more detail later (see Axoplasm). FIGURE 3–1 Transverse section of normal human sural nerve branch showing epineurium (ep), perineurium (arrowhead), and endoneurium (en). There are fat globules (arrows) and blood vessels (asterisk) in the epineurium. Lymph vessels are not identifiable at this magnification. Fleming-fixed, Kultschitsky-stained paraffin section. Bar: 100 ␮m. See Color Plate

CONNECTIVE TISSUES Endoneurium Very approximately 50% of the space enclosed by the perineurium is occupied by axons and their Schwann cells. The endoneurial fluid and fibrous collagen takes up another

FIGURE 3–2 Portion of transverse section through part of the brachial plexus of rabbit, showing disposition of epineurium (Ep), perineurium (P), and endoneurium (En). The structure labeled endoneurium is an intrafascicular partition or septum (see text for explanation). (From Key, A., and Retzius, G.: Studien in der Anatomie des Nervensystems und des Bindegewebes. Stockholm, Samson & Wallin, 1876.)

Microscopic Anatomy of the Peripheral Nervous System

20% to 30% of the space in peripheral nerves; there is very little collagen in spinal roots. Fibrous collagen provides mechanical strength to the fascicle and supports the nerve fibers, both myelinated and unmyelinated, and their associated Schwann cells. Spinal roots are in a protected environment, so mechanical strength is less important than in peripheral nerves, which are subject to stresses by skeletal movements. The fascicle also contains fibroblasts, occasional macrophages and mast cells, and small blood vessels. Approximately 10% of the nuclei within the perineurium belong to fibroblasts,297 and another 2% to 9% to endogenous macrophages,170 and there are very small numbers of mast cells; the remainder are Schwann cell nuclei. Lymphocytes positive for CD4 or CD8 antigen are not usually found in the endoneurium of normal nerves, although they may occasionally be present in the epineurium. Fibroblast counts show that they make up a relatively greater proportion of cell nuclei in nerves containing a higher percentage of large myelinated fibers compared to those in which most fibers are small or unmyelinated, possibly because there are more Schwann cell nuclei per unit area in the latter situation. There are no lymph vessels in the endoneurium. The small blood vessels are mostly capillaries but may be slightly larger, with a variable number of pericytes. Arterioles are occasionally present.16 Nerve biopsies for diagnostic purposes are most commonly taken from the sural nerve because this is relatively easily accessible at the ankle, where it lies between the lateral malleolus and the tendo calcaneus. The sural nerve is a sensory branch of the sciatic nerve innervating the dorsum of the foot. Having been the subject of most attention, this nerve is the best characterized. The numbers of myelinated and unmyelinated fibers have been counted by several authors and are often used as an aid to diagnosis. The use of total myelinated fiber counts is quite rare because this requires a biopsy of the whole sural nerve (6 to 14 fascicles) rather than just a small number of fascicles. More commonly only a proportion of fascicles are removed at biopsy in order to minimize any resultant sensory deficit. Total fiber counts are also very time consuming because there may be as many as 10,000 in the sural nerve at the level of the lateral malleolus.17 The more frequently used myelinated fiber density is, however, susceptible to sampling artifact, and the results are dependent on shrinkage resulting from fixation and on other processing artifacts. The density found therefore varies slightly among different authors but is between 7000 and 9000/mm2 for myelinated fibers and 30,000 and 40,000/mm2 for unmyelinated fibers in adults.207 Myelinated fiber densities may not be helpful in interpretation, however, especially in situations in which there are amyloid deposits or Renaut bodies (see later) or excessive numbers of Schwann cells, as in hereditary motor and sensory neuropathy type 1a. Unmyelinated axons are particularly difficult to quantify because of identification problems, patchy distribution within the fascicle, and their susceptibility to fixation artifacts.

37

At birth the density of nerve fibers is approximately 26,000/mm2, and this reduces as the fibers enlarge and the quantity of fibrous collagen increases.207 The total number of fibers reaches a peak at about age 5 years, but the density continues to fall as a result of increases in myelin thickness and in the quantity of fibrous collagen. The process of maturation is quite slow; although the axon reaches its final diameter around the age of 5 years, the myelin sheath does not reach its maximum thickness until 15 years of age.368 The proportion of myelinated fibers of different calibers also changes with maturation and is unimodal in early childhood, becoming bimodal around the age of 5 years.207 Endoneurial collagen is mainly composed of collagens I and III forming long fibrils with a characteristic crossbanding every 67 nm and a mean diameter of 51 nm.375 This fibrillar material occupies much of the spaces between the nerve fibers and other components of the nerve trunk. In addition, there are other less conspicuous collagenous sheaths of the nerve fibers themselves. In large fibers two sheaths may be identified. The inner one, immediately adjacent to the Schwann cell plasma membrane, is the sheath of Plenk and Laidlaw. This is a very fine sheath of collagen fibers with a circular and oblique orientation that is continuous across the nodes of Ranvier but inflected to follow the Schwann cell basal lamina.227,319,320 By light microscopy this is indistinguishable from the Schwann cell basal lamina. Adjacent to this is another sheath of longitudinal collagen fibrils described by Key and Retzius that is not inflected at the nodes.216 The electron microscopic appearances of these structures were described by Thomas in very early electron microscopic studies of peripheral nerves.419 They have also been studied by scanning electron microscopy, which confirmed the transmission electron microscopic findings.142,439 In small and medium-sized nerve fibers, it is not possible to distinguish two separate sheaths. Perineurial collagen has a similar composition and also has a diameter of 51 nm. This forms wavy bundles that are oriented predominantly along the length of the fascicle, although some may be oriented circumferentially or obliquely. Epineurial collagen, in contrast, forms fibrils of a noticeably larger diameter (80 to 100 nm) and may have a different chemical composition, with a higher proportion of collagen I.375 True elastic tissue is not found in the endoneurium, only the oxytalan component. This is seen by electron microscopy as bundles of nonbranching fibrils with a diameter of 10 to 12.5 nm and an axial periodicity of 17.5 nm and is probably the scaffold on which amorphous elastin is deposited to make elastic fibers.157,264 The amorphous component, elaunin, is not usually seen in the endoneurium or perineurium but may be encountered in the epineurium. Fibrous long-spacing collagen (FLSC) may be occasionally encountered in the perineurium (Fig. 3–3B) or endoneurium, sometimes merging with Schwann cell basal laminae in the latter situation. In addition to nerve trunk perineurium, it may also be present in between the layers of perineurial cells that form the capsules of Meissner and pacinian corpuscles.

38

Structure of the Peripheral Nervous System

FIGURE 3–3 A, Electron micrograph showing transverse section of perineurium of normal sural nerve. Where adjacent perineurial cells overlap, they are connected by tight junctions (arrow) and there are localized regions of densely packed fibrils (asterisk). There are pinocytotic vesicles on both surfaces of the cells (arrowheads). The basal laminal layer (BL) is thick and does not possess a lamina lucida. Collagen fibrils (C) are unstained. Bar: 0.5 ␮m. B, There are bundles of fibrous long-spacing collagen (FLS) lying adjacent to the basal lamina of some cells (asterisk). The bundles of normal collagen are mostly cut in cross section, but some run obliquely or radially (arrowheads). Regions of cell overlap are indicated by an arrow. Bar: 0.5 ␮m.

FLSC forms the Luse bodies described first in schwannomas253 but has also been reported in many other situations, particularly neurofibromas. Luse bodies are well-delineated, usually fusiform structures containing a proteoglycan, probably dermatan sulfate proteoglycan,87 and have a periodicity of 100 to 150 nm. Although some immunochemical studies suggest that they are a version of collagen VI,59,60 others have failed to confirm this.87,359 These conflicting identifications may imply that FLSC can be formed from the breakdown or turnover of either collagen I, III, or VI87 or a reaction to an abnormal extracellular environment. The incidence increases with age, and it is not found in the perineurium of children. FLSC has not been reported in animal nerves. The basal laminal sheaths of Schwann cells have two regions, a dense region (the lamina densa) and a clear space

(the lamina lucida) separating it from the cell membrane. These result from the arrangement of laminin, collagen IV glycosaminoglycans, and other components. Perineurial cell basal lamina often lacks the lamina lucida and also may be considerably thicker than Schwann cell or endoneurial cell basal laminae (see Fig. 3–3A). Renaut bodies may occasionally be encountered in the endoneurium, especially at sites of compression. These are fusiform bodies mainly consisting of fibrous material, predominantly oxytalan with some fibrous collagen and fine cell processes resembling fibroblasts. These are often epithelial membrane antigen positive, suggesting that they are derived from perineurial cells.455 Renaut bodies are typically 200 to 300 ␮m in diameter and rather longer and lie along the long axis of the fascicle. They do not contain

Microscopic Anatomy of the Peripheral Nervous System

axons or Schwann cells. They most commonly lie adjacent to the inner surface of the perineurium but may be deeper in the endoneurium (Fig. 3–4). In the past they have been sometimes misinterpreted as infarcts or other abnormal structures, but they are now accepted as components of normal nerves.

Perineurium The perineurium is composed of modified fibroblasts that are connected together by tight (zonula occludens) and gap (zonula adherens) junctions to form concentric cylindrical sheaths separated by fibrous collagen (see Fig. 3–3). In early development, the nerve fibers are separated into bundles by a loose arrangement of fibroblasts. The mature arrangement with development of tight junctions and a basal laminal coating does not appear until after the commencement of myelination. In the rat, myelination commences at birth but the diffusion barrier properties of the perineurium are not detectable until 30 days after birth.35,380 This may render the immature nerve susceptible to toxins and infectious agents that would be excluded by the perineurial diffusion barrier in the adult.223 It is unknown at what age these barrier properties develop in humans.

FIGURE 3–4 Renaut bodies (asterisks) are most commonly found adjacent to the perineurium but may also be deeper in the endoneurium (human sural nerve at the ankle). Resin section stained with thionin and acridine orange.379 Bar: 100 ␮m. See Color Plate

39

Overlapping perineurial cells do not possess basal laminae between the adjacent membranes. They are linked in these regions by tight junctions; this is the mechanical arrangement underlying their diffusion barrier properties. The multiple interconnecting network formed by tight junctions is well shown by freeze-fracture investigations15,336 (Fig. 3–5). These studies showed that these junctions are far more widespread than suggested by electron microscopy of ultrathin sections. The pinocytotic vesicles (caveolae) seen by electron microscopy take up substances to be transported across the endoneurium. Gap junctions are much more frequently seen in developing nerve and are rare in mature nerve.15 These junctions allow small molecules to pass directly between perineurial cells. There are high levels of various enzymes in perineurial cells, such as phosphorylating enzymes, ATPase, and creatine kinase,372 consistent with the perineurium’s function as an active diffusion barrier. The endoneurial fluid is hypertonic compared with interstitial fluid of general connective tissues.225,280 Presumably the blood-nerve barrier created by blood vessel epithelial cells has a role in regulating this together with the barrier created by the perineurium. If the perineurium is damaged, the contents bulge outward into the epineurium and the larger myelinated fibers are demyelinated in the underlying region.399,421 During recovery after perineurial injury, fibroblasts first encircle the damaged fibers, and these then form small perineurial septa giving rise to minifascicles rather like those seen in neuromas.411 The size of the perineurial window is important in determining the degree of injury. Traumatic nerve injury in which sharp objects penetrate the perineurium may result in persistent neuropathy caused by a persisting defect in the perineurium.411 Perineurial cells also possess cytoplasmic actin-containing structures similar to the filamentous aggregates seen in smooth muscle cells. By analogy with the contractile sheath of seminiferous tubules,77 it has been suggested that these may give the whole structure contractile properties,348 but this has not been confirmed.425 This could explain the ability of the nerve fascicle to maintain its shape and adapt to changes in the volume of the perineurium. Peristaltic-like movements could also move endoneurial fluid along the fascicle. Although perineurial cells were originally thought to derive from the neural crest,373 regeneration experiments and developmental studies suggested that they were derived from fibroblasts.422 This has been confirmed by the use of a retroviral marker.62 Transgenic mice deficient in the signaling protein desert hedgehog have been used to show that its production by Schwann cells converts loosely arranged fibroblasts into mature perineurium.312 Once this has occurred, they can be maintained in vitro and retain their typical characteristics despite repeat passaging and lack of contact with any other cell type.315

40

Structure of the Peripheral Nervous System

FIGURE 3–5 Zonula occludens between overlapping borders of perineurial cells from rabbit sciatic nerve showing multiple interconnecting ridges with intervening areas containing caveolae (p). (A ⫻19,000; B ⫻63,000.) (Courtesy of Dr. G. Gabriel.)

As already discussed, the perineurium is a metabolically active structure that maintains a constant microenvironment in the endoneurial space and forms a barrier to diffusion between the epineurium and endoneurium. This barrier is only breached at the distal termination and where blood vessels traverse the perineurium. Inflammatory cells can use transperineurial vessels as a route into the endoneurium. Although these vessels are accompanied by a sleeve of perineurial cells, this sleeve terminates within the endoneurium and the cells do not fit tightly against the vascular endoneurial cells. Despite this, the combination of collagen fibrils and perineurial cells forms a sufficiently tight seal to maintain a pressure differential between the epineurium and endoneurium. The perineurium may also perform a filtering function by selectively taking up and transporting some substances by pinocytosis305 (see Fig. 3–3A). In axonal degeneration, the perineurial cells take up lipid droplets from myelin breakdown.81 There may also be lipid-laden macrophages between the cellular laminae, implying a movement of these cells from endoneurium to epineurium. The perineurium connects to the pia-arachnoid at the proximal end of the nerve trunk. In sensory nerves the perineurium extends over the dorsal root ganglion, forming the inner layers of its capsule (Fig. 3–6).177 There is an extensive meshwork of tight junctions in this location similar to that seen in the nerve trunk.337 The termination

at the distal end depends on the type of fiber. In encapsulated sensory endings the perineurium appears continuous with the outer capsule layers.371 This is confirmed by immunohistochemical studies showing that the outer layers and capsule of the pacinian corpuscle (Fig. 3–7) are derived from perineurial cells, whereas the inner laminae immunostain for S100 and Leu-7 antigen and are therefore derived from Schwann cells.442 Pacinian corpuscles have frequently been noted in the epineurium and even occasionally in the endoneurium.175 Timofeew’s corpuscles are smaller encapsulated end organs only found in late fetal and early postnatal life. These have been reported in association with autonomic nerves and ganglia in the pelvis.13 In contrast to the situation in which the perineurium is continuous with a capsule surrounding the ending, and hence the axons are always enclosed, in free sensory endings the perineurium terminates before the ending, leaving the axon exposed. Little work has been done on the occurrence of nerve endings in the nerve trunks themselves, but free endings have been identified in the epineurium, perineurium, and endoneurium.201 Investigating the distal regions of the dental pulp shows that the perineurium progressively loses basal lamina and tight junctions, commencing with the outer aspects of the sheath, until the nerve fibers are only surrounded by a loose collection of fibroblasts.295 Similarly, at the

Microscopic Anatomy of the Peripheral Nervous System

41

FIGURE 3–6 Diagram illustrating relationships of the peripheral nerve sheaths to the meningeal ensheathment of the spinal cord. The epineurium (EP) is in continuity with the dura mater (DM). The endoneurium (EN) persists from the peripheral nerves through the spinal roots to their junction with the spinal cord. At the subarachnoid angle (SA), the greater portion of the perineurium (P) passes between the dura and the arachnoid (A), but a few layers appear to continue over the roots as the inner layer of the root sheath (RS). The arachnoid is reflected over the roots at the subarachnoid angle and becomes continuous with the outer layers of the root sheath. At the junction with the spinal cord, the outer layers become continuous with the pia mater (PM). (From Haller, F. R., and Low, F. N.: The fine structure of the peripheral nerve root sheath in the subarachnoid space in the rat and other laboratory animals. Am. J. Anat. 131:1, 1971, with permission.)

motor end plate, the perineurium terminates before the formation of the end plate proper and forms a bellshaped covering that does not make direct contact with the nerve or the muscle.353 A gap of about 1.5 ␮m separates the perineurium from the muscle basal lamina, also rendering the nerve fiber susceptible to external influences at this point, as is also the case with free sensory endings.

Epineurium The quantity of connective tissues surrounding the nerve fascicles varies considerably depending on the location of the nerve. There is only a loose connection with the

surrounding connective tissues, so that the nerve trunks are relatively mobile. This presumably reduces the likelihood of damage by entrapment during movement of joints. In general there is a more extensive epineurium at joints and in trunks with greater numbers of fascicles than in smaller nerves with few fascicles. The epineurium can be separated into two layers, the epifascicular epineurium and the interfascicular epineurium. Mice deficient in desert hedgehog not only lack a true perineurium but have very sparse epineurial collagen.312 This suggests that the development of well-defined nerve fascicles is required for the formation of an epineurium. There is no epineurium in spinal roots, where the root sheath is a looser, less compact layer.

42

Structure of the Peripheral Nervous System

protect the nerve fascicles from compression damage.412 The loss of these fat deposits in generalized wasting may be a factor in the development of pressure palsies in wasted, bedridden patients.

Myelinated Nerve Fibers C.-H. Berthold, R. H. M. King, and M. Rydmark

FIGURE 3–7 Transverse section through a pacinian corpuscle within the epineurium of a fascicle in human sural nerve. A large myelinated fiber is just visible at the center of the concentric layers of cell processes (arrow). Resin section stained with thionin and acridine orange. Bar: 100 ␮m. See Color Plate

Nerve trunks may be distinguished from other longitudinal structures by naked eye observation as a result of the white spiral bands seen on their surface. These are termed the spiral bands of Fontana.115 Stretching or damaging the nerve changes or destroys these bands and results in a homogeneous gray appearance. Fontana deduced that these bands resulted from the zigzag course of the nerve fibers within the nerve trunk.115 This allows stretching of the nerve by up to 20% without damaging the nerve fibers.76 More recent work, however, showed that the wavy pattern actually originates in the arrangement of the collagen fibrils in the inner layer of the epineurium, which lie in wavy bundles.403 These waves may be necessary to allow the nerve to straighten without stretching the nerve fibers, because collagen fibrils are not extensible. A longitudinal anastomotic network of arterioles and venules provides the blood supply to the nerve fascicles. Epineurial capillaries may be fenestrated, unlike those in the endoneurium. In addition, there is a lymphatic capillary network in the epineurium that accompanies the arteries and is connected to the regional lymph nodes but without connections to the endoneurium.413 There may also be considerable quantities of adipose tissue in the outer layer of the epineurium, especially in the sciatic nerve in the buttocks and thigh. This distribution has led to the suggestion that this fat may act as a buffer to

The microscopic and molecular anatomy of peripheral myelinated nerve fibers has been reviewed several times during the past decade.18,20,34,156,228,229,322,360–362,425,452 The following description is based mainly on the microscopic organization as found in peripheral myelinated mammalian (rat, guinea pig, rabbit, cat, human, and other) nerve fibers more than 4 to 5 ␮m in diameter. Particular emphasis is given to the node of Ranvier and its close internodal neighborhood. The peripheral myelinated nerve fiber (Fig. 3–8) consists of a single continuous neuronal process, the axon, and a set of Schwann cells arranged serially along the outside of the axon. Each Schwann cell enwraps the axon segment with which it is associated with an electrically insulating, multilayered cell membrane specialization, the myelin sheath. The meeting points of consecutive Schwann cells appear as short (approximately 1 ␮m long) constricted and myelin-free fiber segments, the nodes of Ranvier. The intervening 300- to 2000-␮m-long fiber segments, each corresponding to the extension of one Schwann cell with a length proportional to fiber size, are the internodes (see Reynolds and Heath339). An internode can be separated into three parts: a main internodal region that constitutes about 95% of the internodal length, holds the Schwann cell nucleus at its midpoint, and displays the biconical myelin sheath clefts referred to as Schmidt-Lanterman incisures; and a proximal and a distal end region, each corresponding to 2% to 3% of the internode (“proximal” and “distal” refer to the cell body). The transition from the main internodal region to the end regions is gradual and indicated by an increasing amount of mitochondrion-rich Schwann cell cytoplasm outside a more or less crenated myelin sheath (Fig. 3–9A to E). As a rule, the end regions are dilated, the distal one slightly more than the proximal one, forming the paranodal bulbs of classic light and electron microscopy.250,460–462 In transverse section, a myelinated nerve fiber shows two main concentric zones: an outer Schwann cell zone and an inner axonal zone. The two zones are separated by the periaxonal space (see Fig. 3–10C later). Internodally the Schwann cell zone can be divided into four concentric regions: (1) an outermost basal lamina that sharply

Microscopic Anatomy of the Peripheral Nervous System

43

FIGURE 3–8 Schematic drawing to show the major components of a large myelinated nerve fiber and the terminology used. ASN ⫽ axon–Schwann cell network; B ⫽ mitochondrion bag; CON ⫽ constricted axon segment; IE Reg. ⫽ internodal end region; IN ⫽ internode; JPS ⫽ juxtaparanodal segment; NoR ⫽ node of Ranvier; PS ⫽ paranodal segment; PNP ⫽ paranode-node-paranode region; SCN ⫽ Schwann cell nuclear region, which is situated more or less at the midpoint of an internode. According to this view, an internode consists of a proximal and a distal end region (IE Reg.) and between these the main internodal region. An internodal end region is separated in a juxtaparanodal and a paranodal segment (IE Reg. ⫽ JPS ⫹ PS). A node of Ranvier and its two bordering internodal end regions make up a PNP region.

demarcates the fiber from the endoneurial space and runs uninterrupted all the way between the central nervous system (CNS) and the fiber terminations, (2) the outer (abaxonal) cytoplasmic compartment, (3) the myelin sheath, and (4) the inner (adaxonal) cytoplasmic compartment. At a node of Ranvier, the Schwann cell zone consists of the basal lamina and nodal constituents derived from the outer cytoplasmic compartment. The transverse contour of the myelin sheath varies according to transverse level as a result of the distribution of the Schwann cell cytoplasm outside the sheath. Thus the contour of the myelin sheath is more or less indented by the Schwann cell perikaryon, equipped with shallow notches (see Fig. 3–9A) along the stretch between the perikaryon and the internodal end regions and crenated in the latter (see Figs. 3–9C to E). The axonal contour adapts to that of the myelin sheath except at the nodes of Ranvier, where it is approximately circular (see Fig. 3–9D to F). A close-to-circular myelin sheath/axonal contour is also noted at sites where a Schmidt-Lanterman incisure enters the myelin from the outer Schwann cell compartment (see Fig. 3–9B). Three morphometric variables are currently used in order to define a myelinated nerve fiber, assuming a circular transverse fiber section: (1) d ⫽ the diameter of

the axon measured in the main internodal region, avoiding the Schwann cell perikaryon and Schmidt-Lanterman incisures; (2) nl ⫽ number of myelin sheath lamellae; and (3) il ⫽ internodal length. From these three a number of secondary variables can be calculated, including fiber diameter (D ⫽ d ⫹ 2 ⫻ nl ⫻ lamellar thickness), myelin sheath transverse area, and internodal myelin volume.32,287 The size of mammalian peripheral myelinated axons, given as a d value, is in the range 1 to 20 ␮m; few exceed 15 ␮m. A rough estimate of il is obtained from il ⫽ d ⫻ 100. Local variations in fiber variables such as axon diameter, myelin sheath thickness, internodal distance, branching, and size of branches all influence axon function and impart integrative capacities to the axon.448 The microscopic organization of internodal main regions is relatively simple compared with that of a node of Ranvier and its two bordering internodal end regions. These parts of myelinated nerve fibers—the paranode-node-paranode (PNP) regions—constitute, structurally as well as functionally, the most spectacular parts of a myelinated nerve fiber. The PNP regions are responsible for the generation and propagation of the action potential.342 They interfere with axoplasmic transport,21,29,327 and they are reactive centers in

44

Structure of the Peripheral Nervous System

A

B

C

D

FIGURE 3–9 Electron micrographs of transversely sectioned alpha motor nerve fibers; L7 ventral spinal root, adult cat. A, Main internodal region. The sectioning plane runs outside the Schwann cell perikaryon and in between Schmidt-Lanterman incisures. Strands of the outer Schwann cell compartment, containing a few mitochondria, are at the arrowheads. At this level the transverse contour of the myelin sheath is slightly undulated. Ax ⫽ axon. Bar: 2 ␮m. B, Main internodal region. The sectioning plane runs through a level where a Schmidt-Lanterman incisure reaches the outer Schwann cell compartment (arrowheads). At such levels the myelin contour is approximately circular and surrounded by a substantial layer of the outer Schwann cell compartment; in this way there is a thin cytoplasmic girdle that cross-connects the longitudinal strands of cytoplasm (arrowheads in A) that run between the perikaryon and the node of Ranvier. Note the more or less angulated contours of adjacent fibers without Schmidt-Lanterman incisures. Ax ⫽ axon. Bar: 2 ␮m. C, Juxtaparanodal segment about 10 ␮m from the node. The arrow indicates one of six mitochondrion bags. Ax ⫽ axon. Bar: 2 ␮m. D, Transition level between the juxtaparanodal and paranodal segments, about 5 ␮m from the node of Ranvier. The innermost myelin lamellae have formed the “first” turns of the terminal cytoplasmic spiral and attach to a constricted and rather circular axonal contour. Asterisks indicate three of the six extensive axon–Schwann cell networks situated inside the myelin crests. Ax ⫽ axon. Bar: 2 ␮m. Figure continued on opposite page

45

Microscopic Anatomy of the Peripheral Nervous System

F

E

FIGURE 3–9 Continued E, Paranodal segment about 3 ␮m from the node of Ranvier. Arrow indicates a mitochondrion bag. Ax ⫽ axon. Bar: 2 ␮m. F, Node of Ranvier. The nodal axon (Ax) is surrounded by a nodal gap (Ng) filled with radiating Schwann cell microvilli. Note the tuftlike arrangements of the microvilli at the lower left and the upper right part of the gap. PN ⫽ perinodal space. Bar: 2 ␮m.

a large number of neuropathies (see later),50,51,187,205,277,381 as well as during the early phase of wallerian degeneration,453 and in collateral sprouting.166,199,200,263,461 The structural organization of the PNP regions differs according to fiber size. Fibers of a diameter less than 4 to 5 ␮m have relatively simple PNP regions that are referred to as “nodes of Ranvier type I” (see Fig. 3–11A later). Larger fibers, the ones treated here, have “nodes of Ranvier type II” (see Fig. 3–11B later).318,328 In the present context, “paranode” has been used (1) in a broad sense, according to traditional concepts229,332,454,463 that can be traced back to Ranvier himself, referring to the whole “internodal end region” of Figure 3–8 and used here only in the notion of the “PNP region”; and (2) in the more usual contemporary, restricted sense, referring to those segments of a myelinated nerve fiber that include the attachment of the myelin sheath to the axon.

GENERAL ORGANIZATION OF MYELINATED AXONS The axon consists of a relatively firm gelatinous cord of neuronal cytoplasm, the axoplasm,174,249 enclosed by the axolemma (i.e., by the cell membrane of the axon). The axon is separated from the adaxonal Schwann cell membrane by the periaxonal space.

Periaxonal Space According to the preparative procedure used, the periaxonal space has been reported to measure from about 30 nm down to just a few nanometers (Fig. 3–10C). At some sites the two facing cell membranes appear to fuse and form five-layered membrane complexes, reminiscent of tight junctions (Fig. 3–10C). Freeze-fracture studies, however, have not shown tight junctions between the axolemma and the adaxonal Schwann cell membrane. In contrast, an extensive tight junctional complex has been demonstrated in the mouth of the inner mesaxon (see later).365 The Schwann cell membrane that faces the periaxonal space throughout an internode shows a strong myelin-associated glycoprotein (MAG) positivity.324 The periaxonal space is probably sealed off from the mesaxonal space of the compact myelin, being open only at the nodes, where it apparently communicates more or less freely with the endoneurial space via the nodal gaps as indicated by the accessibility of different tracer substances.176 In large fibers, however, not even lanthanum enters the periaxonal space from the nodes.427

Axolemma The axolemma appears as a conventional three-layered unit membrane about 8 nm thick when examined in standard electron microscopic preparations. Freeze-fracturing demonstrates that the outside of the inner leaflet of the

FIGURE 3–10 A, Electron micrograph of an alpha motor nerve fiber sectioned transversely about 15 ␮m from the node of Ranvier; L7 ventral spinal root, adult cat. Microtubular (MT) and neurofilament (NF) domains are well separated. Arrows point at axoplasmic reticulum profiles; arrowheads at vesiculotubular membranous profiles. m ⫽ mitochondrion. Bar: 0.2 ␮m. B, Freeze-etched stereo pair electron micrographs of saponin-treated frog axon showing 11-nm neurofilaments linked by a ladder-like arrangement of 4- to 6-nm cross-bridges 20 to 30 nm in length. Bar: 0.1 ␮m. (From Hirokawa, N.: Cross-linker system between neurofilaments, microtubules, and membranous organelles in frog axons revealed by the quick-freeze, deep-etching method. J. Cell Biol. 94:129, 1982, with permission.) C, Electron micrograph of an alpha motor nerve fiber sectioned transversely through the main internodal region; L7 ventral spinal root, adult cat. The specimen was fixed in glutaraldehyde ⫹ tannic acid followed by OsO4 ⫹ potassium ferrocyanide. With this method, organelles appear coated by a diffuse layer of electron-dense material. The lipid (intermediate) layer of the cytomembranes appears like a “white” (clear) line, approximately 3 nm in width. There is a thick axoplasmic cortex (arrowheads) inside the axolemma (white arrow). Neurofilaments (NF) are interconnected by diffuse electron-dense bridges. One of several microtubules is encircled. Asterisks indicate axoplasmic reticulum profiles associated with the axoplasmic cortex. Periaxonal space is marked with one dot (•) and the inner adaxonal-cytoplasmic Schwann cell compartment by two dots (••). The axolemma and the Schwann cell membrane seem to fuse at several sites (black arrows). My ⫽ myelin sheath; m ⫽ mitochondria. Bar: 0.1 ␮m.

Microscopic Anatomy of the Peripheral Nervous System

axolemma, the so-called P face, is comparatively rich in intramembranous particles (IMPs) (see Fig. 3–12C later). The inside of the outer leaflet, the so called E face, displays, except at nodes of Ranvier (see later), relatively few IMPs.145,272,405 Many E-face IMPs are about 10 nm in size and may represent voltage-sensitive Na⫹ channels.224,344,404,449 The axolemma conveys signals between the neuron and its Schwann cells that control the proliferative and myelinproducing functions of the Schwann cells and partly regulate axon size.46,49,84,85,171,172,194,239,355,429 The additional presence of receptor and transducer proteins involved in axon–Schwann cell communication has been postulated by several authors.108,109,209,244 The axolemma is stabilized by the immediate subjacent part of the axoplasm, which here forms the axoplasmic cortex (see Fig. 3–10C), an outer condensed part of the cytoskeletal microtrabecular matrix containing ankyrin, fodrin, actin, and A-60.112,151,184,189,192,194,241,335,366,438

Axoplasm The axoplasm consists of a cytosol and formed elements suspended in it. The formed elements are (1) the axoplasmic organelles (i.e., mitochondria, the axoplasmic [endoplasmic] reticulum, dense lamellar bodies, multivesicular bodies, vesiculotubular profiles, and membranous cisterns); (2) the cytoskeleton; and (3) axoplasmic inclusions. Organelles and inclusions show their highest occurrence and their most elaborate arrangements in the PNP regions of large myelinated axons. Ribosomes and Golgi apparatuses are absent, though stray ribosome-like granules have been reported (see Alvarez3). Organelles Numerical data referring to the frequency of various axoplasmic organelles in the main internodal region are meager.20,475 Mitochondria. Axonal mitochondria are 0.1 to 0.3 ␮m in diameter and 0.5 to more than 10 ␮m in length. Usually they lack electron-dense granules and have flattened or tubelike, often longitudinally oriented, cristae. The concentration of axonal mitochondria decreases with increasing axon diameter from about 1/␮m2 transverse area of axoplasm in the thinnest myelinated axons to about 0.1/␮m2 in the thickest ones.20 A close relationship between mitochondria and the axoplasmic reticulum (AR) has been demonstrated ultrastructurally, and formation of axonal mitochondria in relation with the AR has been suggested.396 Mitochondria are transported both anterogradely and retrogradely in the axon. Transportation is saltatory: Short periods of movement at a velocity similar to that of other organelles are interrupted by long periods of rest.116,117 The outer membrane of the mitochondrion is considered multivalent in terms of sites capable of interacting with the force-generating transport mechanism.167,256

47

The Axoplasmic Reticulum. The AR is to be distinguished from the smooth membranous vesicular, tubular, and vesiculotubular profiles that occur scattered in small amounts all along the axon and accumulate in large numbers proximal to a transportation block.168,246,431 The AR forms a system of delicate varicose membranous tubules about 20 to 60 nm in diameter and flattened sacs, up to 100 nm in width, that branch and interconnect frequently as they extend lengthwise throughout the axon from the hillock into the terminals.246,330,331 The rough endoplasmic reticulum of the soma connects via a “transitional endoplasmic reticulum” to the AR in the proximal part of the axon hillock.246 A connection between the AR and the Golgi apparatus has also been demonstrated.325,402 The AR can be subdivided into a minor, outer, subaxolemmal and a major, inner, central compartment. The former is situated just inside the axolemma, where, embedded in the axoplasmic cortex, it forms a single layer of tubes and sacs that interconnects with the central AR.21,331 In view of abundant branching and a minimum tubular caliber of less than 20 nm, the proposition that the AR extends uninterrupted throughout the axon has been hard to contradict. The results of cold-block experiments indicate that the AR itself is not rapidly transported either anterogradely or retrogradely.104 Most likely, the AR is a vector in slow anterograde transport and is not involved in retrograde transport.246 It acts as a reservoir for Ca2⫹ ions.97 Dense Lamellar Bodies and Multivesicular Bodies. These bodies vary in shape and size from short tubes 0.1 ⫻ 0.2 to 0.5 ␮m to rounded entities about 0.5 ␮m in diameter (Fig. 3–11B; see also Figs. 3–15A and 3–15D later). Dense lamellar bodies (DLBs) and multivesicular bodies (MVBs) are rare outside the PNP regions.28 The ultrastructural appearance of the DLBs and the MVBs and their occasional content of acid hydrolases indicate that they belong to a heterogeneous group of organelles that includes secondary lysosomes and residual bodies.148–150 They are considered to be transported retrogradely.104,432 Vesiculotubular Profiles. These elements are 0.05 to 0.1 ␮m in diameter and up to 0.5 ␮m (the tubes) in length (see Fig. 3–11B; see also Fig. 3–15A and C later). Some seem empty, whereas others contain granular electron-dense material. As a rule, the tubes appear varicose or even beaded. Tubes and vesicles lie close together in the PNP regions, forming rows or strands together with AR profiles in association with bundles of microtubules (MTs). 106,431 These organelles are transported anterogradely in the axon. 104,432 They probably derive from the Golgi apparatus of the soma and act as fast-transport vectors for newly synthesized protein and lipid.106,179,208,246,406,428

48

Structure of the Peripheral Nervous System

B B PN

Ax

C

B B

Ax

Microscopic Anatomy of the Peripheral Nervous System

Membranous Cisterns. All membrane-demarcated and empty-looking entities more than 0.15 ␮m in size are here classified as cisterns. This is a heterogeneous group that includes vacuoles, of which some are probably large endosomes and some preparatory artifacts. The Cytoskeleton This, the most conspicuous of the axoplasmic components,64 consists of the MTs, the neurofilaments (NFs), and the microtrabecular matrix, including the microfilaments (Fig. 3–12B; see also Fig. 3–10).101,105,191,435 It defines the size, shape, stability, and growth pattern of the axon. It contains the machinery necessary for bidirectional axoplasmic transport, that is, the interaction with the anterograde transport “motor” (kinesin) and the retrograde “motor” (dynein).167,374 Axoplasmic transport and its molecular correlates differ in several aspects from that in the cell body and the dendrites.61,197 There are also differences in the composition of the axonal cytoskeleton with respect to type of peripheral nerve, distance from cell body, and age.291,417,447 Microtubules extend longitudinally in the axon and form the tracks along which the fast bidirectional axoplasmic transport of membranous organelles takes place.456 They are approximately 25 nm thick, are unbranched, and vary from a few to more than 1000 ␮m in length.45,65,433 They lie solitary in the axoplasm or form small bundles.243 The number of MTs varies inversely with axon size from 30 to 40/␮m2 cross-sectioned axoplasm in small axons to 10 to 15/␮m2 in large axons.20,107,195 The MT packing varies along the course of an axon354 as well as between afferent and efferent ones.309,354,475–477 MTs are fragile structures, and about 50% of the MT mass is particularly labile and disassembles easily.11,12 They are, however, well preserved after glutaraldehyde–tannic acid fixation.

49

Neurofilaments run longitudinally in the axon, often with a slight spiraling course. They are approximately 10 nm thick, of undefined length, and unbranched. There are about 150 to 200 NFs/␮m2 cross-sectioned axoplasm, their density normally being relatively unaffected by the size of the axon. NFs are considered to be the “major determinant” of axon size.193–195,311,435,444 Three kinds of NF monomers polymerize together to form a neurofilament: NF-L (68 kDa), NF-M (150 kDa), and NF-H (200 kDa), with all three existing in many different isoforms depending on the degree of phosphorylation.257,289 The less phosphorylated, the higher is the probability that the NF units exist as mobile mono- or oligomers and the more labile are polymers of such units. NF density and degree of phosphorylation increase with distance from the cell body and with age. The microtrabecular matrix is only faintly indicated in electron micrographs of axons submitted to conventional preparative protocols. It then appears as thin strings of a wispy material that radiate from and interconnect the NFs and some MTs (see Fig. 3–10A). Other preparatory procedures, including quick freeze in combination with deepetching and rotatory shading (see Fig. 3–10B), examination of unfixed tissue in the high-voltage electron microscope, and the use of tannic acid as a mordant in combination with glutaraldehyde fixation (see Figs. 3–10C and 3–12B), have all added new features to an otherwise rather emptylooking axoplasm.21,105,261,434 The most spectacular of these is the plethora of microtrabecules (cross-linkers) that radiate from the shafts of the NFs and the MTs, linking them into a dense lattice that extends throughout the axoplasm (see Fig. 3–10).189,192,266,366 The axoplasmic organelles are suspended in the NF-MT lattice and cross-linked to it by microtrabecules. Three main types of microtrabecules, all 4 to 6 nm thick and 20 to 150 nm long, have been

FIGURE 3–11 A, Electron micrograph of longitudinal section from a human sural nerve showing a small myelinated fiber with a type I node of Ranvier. The paranodal segments (Para) are smoothly tapered. Few nodal microvilli are distinguishable. There is an extensive collection of mitochondria (asterisk) together with glycogen deposits (G). The paranodal myelin lamellae are connected together by a strip of dense desmosome-like structures (arrow). Bar: 1 ␮m. B, Electron micrograph of longitudinal section through a highly segregated CON segment and its proximal (P) and distal (D) internodal end regions; alpha motor nerve fiber, L7 ventral spinal root, adult cat. Dense lamellar bodies (DLBs), multivesicular bodies (MVBs), filled vesiculotubular profiles, and some mitochondria occupy the clear axoplasm distal to the nodal midlevel. Large numbers of empty-looking vesiculotubular profiles and some mitochondria are present in the finely granular axoplasm proximal to the nodal midlevel. Note the bulging contour of the proximal paranodal segment (Para). ASN ⫽ axon–Schwann cell network. Bar: 2 ␮m. C, Electron micrograph of longitudinal section through a PNP region; alpha motor nerve fiber, L7 ventral spinal root, adult cat (sectioning plane is outside the nodal gap). PN ⫽ perinodal space between the nodal ends of the two adjacent paranodes. The myelin sheaths are crenated and equipped with longitudinal crests and intervening furrows. The latter are filled with mitochondrion-rich Schwann cell cytoplasm forming the mitochondrion bags (B). Note electron-dense glycogen granules dispersed among the mitochondria. Ax ⫽ axoplasm inside a myelin crest. Bar: 2 ␮m. D, Electron micrograph showing detail from a longitudinally sectioned mitochondrion bag of an internodal end region; alpha motor nerve fiber, L7 ventral spinal root, adult cat. The Schwann cell cytoplasm contains masses of densely packed mitochondria (Mi) and clusters of glycogen particles (G). My ⫽ myelin sheath; E ⫽ endoneurial space. Bar: 0.5 ␮m.

B

C

FIGURE 3–12 A, Electron micrograph of longitudinal section through a nodal gap (NG) containing densely packed microvilli emanating from the Schwann cell nodal collars (SC); alpha motor nerve fiber, L7 ventral spinal root, adult cat. The nodal collars separate the nodal gap from the perinodal space (P). Asterisks in the myelin sheath indicate the tips of “ear-of-barley”–like arrangements of the terminal cytoplasmic pockets. The nodal axolemma extends between the two bars (note thick electron-dense coating inside the axolemma in this GMA embedding), while the node gap proper extends between the two arrowheads. A nodal recess extends between a bar and an arrowhead. The nodal gap walls (W) overlie each recess and delineate the node gap from the myelin sheath. Ax ⫽ axon. Bar: 0.5 ␮m. (From Berthold, C. H., and Rydmark, M.: Electron microscopic serial section analysis of nodes of Ranvier in lumbosacral spinal roots of the cat: ultrastructural organization of nodal compartments in fibres of different sizes. J. Neurocytol. 12:475, 1983, with permission.) B, Electron micrograph of transversely sectioned nodal axon segment; alpha motor nerve fiber, L7 ventral spinal root, adult cat. Two arrowheads indicate the axolemma. The axoplasmic cortex is particularly well developed (white bar) and measures 60 to 100 nm. Matrix substance coats all formed elements and connects some of them directly with the axoplasmic cortex. Arrows mark out crosscut AR profiles. MTs are encircled. Asterisks are in organelles of the vesiculotubular type. Ng ⫽ node gap. Glutaraldehyde and tannic acid followed by osmium tetroxide and potassium ferrocyanide. Bar: 0.1 ␮m. Legend continued on opposite page

Microscopic Anatomy of the Peripheral Nervous System

identified in frog myelinated axons.189,192,232,241,243 After tannic acid treatment, all membranous organelles, the NFs, the MTs, and the inside of the axolemma appear coated by a 3- to 30-nm thick, uneven, at times coarsely granular layer of material from which the microtrabecules seem to emanate (see Fig. 3–10C). The matrix layer coating the inside of the axolemma (see Figs. 3–10C and 3–12B) is referred to as the axoplasmic cortex.105 Previous observations have suggested that the cytoskeleton is a rather immobile element in the axoplasm. The idea formerly accepted—that the whole NF-MT lattice slides slowly down the axon235,236—was seriously questioned,300,302,303 as it had been already from some earlier experimental studies.397 The slow component-a (SCa) of axoplasmic transport that contains tubulin and the NF appeared to be gradually incorporated in a “stationary” polymer lattice and renewed it by local disassembly and reassembly.301 More recent observations have shown that MTs and NFs are, indeed, actively transported along axons by fast motors, although in an infrequent and asynchronous way interrupted by long pauses (see Baas,10 Miller et al.,271 and Wang and Brown446). A similar situation may also apply to members of the other slowly transported component (SCb). In addition to proteins heading for the terminals, such as synapsin and clathrin-uncoating protein complexes,38 this component contains proteins that belong to the microtrabecular matrix, such as actin (during transport probably complexed with an actin depolymerizing factor47), fodrin (brain spectrin), myosin, tau,221 tropomyosin, and calmodulin.39,112,236,237,243 The axoplasm can, according to the distribution of its cytoskeletal components, be separated into three principal domains: (1) a subaxolemmal domain, (2) domains characterized by NFs, and (3) domains characterized by MTs and diffuse granular material.366 AR is present in all three domains. The subaxolemmal domain, mainly studied in squid giant axons, where it is several micrometers thick,44 should in vertebrate axons correspond to the axoplasmic cortex and the outer aspect of the subjacent axoplasmic zone. The domain interconnects the axolemma and the cytoskeleton and contains ankyrin, fodrin (brain spectrin), actin, and A-60.184,189,192,241,268,366,434 It includes the outer compartment of the AR, which covers 10% to 20% of the inner aspect of the axolemma.21 Fast anterograde axoplasmic transport seems to have a predilection for the subaxolemmal domain.331,393 The NF domain constitutes the shape- and

51

size-supporting framework of the axon. It is, as seen in transverse sections, pierced and separated in interconnecting subdomains by the “channels,” “tunnels,” or “streets” formed by the MT domains (see Fig. 3–10A), along which rapid anterograde and retrograde transport of membranous organelles takes place. Granular Material The presence of granular material (in the present context this refers to electron-dense granules as seen after conventional preparative procedures) is sporadic and is usually restricted to the PNP regions. There are two main types of granules, coarse ones of high electron density and minute ones that aggregate in areas with a pepper-like appearance.32 Granules with high electron density measure from about 25 nm to greater than 100 nm. Of these, the smaller are classified as glycogen granules.20,318,327,474 They usually form clusters and become more common with increasing age.31,424 Tiny granules that form territories of a pepper-like texture seem to be present only in the PNP regions, where they, as a rule, occupy either the proximal or the distal half (see Fig. 3–11B). They are less than 10 nm in diameter and lie packed closely together, obscuring MTs and NFs. Most likely, the accumulations of these tiny granules give rise to the diffuse staining seen on light microscopy in the PNP regions of some axons.28,273,328 These may represent specific proteins, but the precise identity of the granular material is unknown.

THE PARANODE-NODE-PARANODE REGION A PNP region consists of a central node of Ranvier and two bordering internodal end regions. The latter can, with reference to the node, be separated in two parts: (1) a short (irrespective of fiber size) 3- to 5-␮m long paranodal segment situated adjacent to the node; and (2) more internodally, a juxtaparanodal segment the length of which is proportional to fiber size. The juxtaparanodal segment (see Fig. 3–9C) is defined by the changing shape and size of the Schwann cell, the myelin sheath, and the axon as compared with their appearance in the main internodal region. The paranodal segment is defined by the termination of the myelin sheath and its attachment to the axon (see Fig. 3–9E). The transition between the juxtaparanodal and the paranodal segment is, in fibers more than 4 to 5 ␮m

FIGURE 3–12 Continued C, Electron micrograph of freeze-fracture preparation of rodent sciatic nerve exposing the P face of the axolemma of the node (N) and paranode region of a myelinated fiber. There is a dense population of intramembraneous particles on the nodal axolemma. The paranodal region is characterized by the presence of alternating ridges (arrows) and troughs in the axolemma related to the attachment of the helically arranged terminal myelin loops (TL). The parallel transverse bands associated with their attachment are visible on the axolemma. Direction of platinum shadowing is indicated by the arrowhead in the lower left corner. Bar: 0.2 ␮m. (Courtesy of N. G. Beamish.).

52

Structure of the Peripheral Nervous System

in diameter, indicated by a reduction of the axon transverse area by 75% or more350 and by the axon adopting an almost circular transverse contour (see Fig. 3–9D). The axon keeps a circular transverse outline as it passes through the proximal and distal paranodal myelin sheath cuffs and the intervening nodal gap. In this way there is at the middle of a PNP region a 7- to 10-␮m-long constricted, fairly cylindrical axon segment: the CON segment. The CON segments and their radial Schwann cell surroundings form together the functionally and structurally most complex and also the most vulnerable parts of a myelinated nerve fiber. The molecular machinery

necessary to generate trains of action potentials is localized here (for review and references, see Waxman and Ritchie452). Here also are the conspicuous zones of myelin sheath attachment to the axon. The CON segments furthermore have to maintain axoplasmic flow through a tube whose transverse area is reduced by 75% to 90% as compared with that of the main internodal region, a reduction nevertheless shown to be advantageous for conduction velocity.178,262 The CON segment is considered separately later, after descriptions of the internodal end region and the node of Ranvier. Morphometric data are presented in Table 3–1.

Table 3–1. Paranode and Node of a Large Cat Alpha Axon: A Morphometric Description General fiber characteristics Diameter (D) ⫽ 17.5 ␮m Axon diameter (d) ⫽ 12.5 ␮m Number of myelin lamellae ⫽ 140 Myelin sheath thickness ⫽ 2.5 ␮m Internodal length ⫽ 1800 ␮m Internodal End Region Juxtaparanodal segment (values refer to 1 region) Length Axon mean cross-sectional area Axon maximum circumference Axon mean circumference Axon membrane area Myelin sheath maximum cross-sectional area Myelin sheath maximum circumference Schwann cell adaxonal membrane area Schwann cell adaxonal cytoplasmic maximum cross-sectional area Schwann cell adaxonal cytoplasmic volume Schwann cell endoneurial membrane area Schwann cell outer cytoplasmic compartment maximum cross-sectional area Schwann cell outer cytoplasmic compartment volume Schwann cell mitochondria, maximum number noted in a cross section Schwann cell mitochondria, calculated total number (assumed mitochondrion size: 0.15 ⫻ 0.5 ␮m) Paranodal segment (values refer to 1 region) Length Axon diameter Axon cross-sectional area Axon circumference Axon membrane area Axon volume Terminal cytoplasmic spiral, number of turns Terminal cytoplasmic spiral, number of turns attached to axolemma Terminal cytoplasmic spiral, length Terminal cytoplasmic spiral, diameter of cord Terminal cytoplasmic spiral, membrane area Terminal cytoplasmic spiral, volume Schwann cell outer cytoplasmic compartment cross-sectional area Schwann cell outer cytoplasmic compartment volume Schwann cell endoneurial membrane area

75 ␮m 102 ␮m2 63 ␮m 49 ␮m 3680 ␮m2 188 ␮m2 88 ␮m 3680 ␮m2 0.25 ␮m2 15 ␮m3 4330 ␮m2 28 ␮m2 1214 ␮m3 450 10.000 4 ␮m 4.7 ␮m 17.7 ␮m2 14.9 ␮m 60 ␮m2 71 ␮m3 140 26 2530 ␮m 0.1 ␮m 795 ␮m2 20 ␮m3 8 ␮m2 32 ␮m3 200 ␮m2

4%* 83% 160% 125% 5% 160% 160% 5% 160% 5% 5% 230% 6% ⬎2000% — 0.2% 38% 14% 38% 0.08% 0.03% — — — — — — 65% 0.2% 0.3%

Table continued on opposite page

53

Microscopic Anatomy of the Peripheral Nervous System

Table 3–1. Paranode and Node of a Large Cat Alpha Axon: A Morphometric Description—Continued Node of Ranvier Axon compartment Axon length Axon diameter Axon cross-sectional area Axon circumference Axon membrane area Axon volume Schwann cell compartment Extracellular interconnection area between endoneurial space and node gap Node gap width Node gap height Node gap extracellular volume Node gap Schwann cell microvilli number mean length mean diameter total membrane area total volume Node gap walls, membrane area Node gap recesses, ceiling area Paranode-Node-Paranode axon region Mitochondria Multivesicular bodies Dense lamellar bodies Vesiculotubular membraneous profiles

1 ␮m 5 ␮m 20 ␮m2 16 ␮m 24 ␮m2 19 ␮m3

0.006% 40% 16% 40% 0.03% 0.009%

0.2 ␮m2 0.3 ␮m 1 ␮m 0.7 ␮m3

— — — —

800 1 ␮m 84 nm 210 ␮m2 4.4 ␮m3 43 ␮m2 11 ␮m2

— — — — — —

— — — —

2.6% 100% 92.6% 97.0%

*Percentage values represent parts of the corresponding internodal total. Data are taken from Rydmark and Berthold,351 Rydmark et al.,352 and Berthold et al.28 and are compensated for preparative dimensional changes.27

Internodal End Region The juxtaparanodal segment is characterized by an irregular myelin sheath contour and a comparatively voluminous outer Schwann cell cytoplasmic compartment. In many mammalian species, including humans, this cytoplasm is concentrated in longitudinal cords that extend with an increasing volume from the main internodal region to the node of Ranvier (see Figs. 3–9C and D). The cords, one or two in small fibers and up to seven in the largest ones, “indent” the underlying myelin and axon. This gives rise to conspicuous myelin sheath irregularities, often like longitudinal crests, that in transverse sections produce the typical appearance of a deeply crenated myelin sheath that surrounds a correspondingly fluted axon.20,231,327,463 In fibers more than 5 to 6 ␮m in diameter, the cords of Schwann cell cytoplasm become extraordinarily rich in mitochondria as the node is approached and form the “mitochondrion bags” of the internodal end region (see Figs. 3–9C and 3–11C and D; see also Table 3–1).352 The number of mitochondria reaches a maximum value at a level about 10 ␮m from the node. More nodally, it decreases rapidly and becomes close to nil a few micrometers from the node. Mitochondrion bags are rich in glycogen, many contain lipid droplets, and

some hold Marchi-positive myelinoid bodies and crystalline lamellar bodies (Reich granules; see Figs. 3–17 and 3–18 later). The Juxtaparanodal Segment The juxtaparanodal axon segment is a cast of the surrounding myelin sheath and characterized by an irregular, usually fluted, shape. The slender longitudinal axon ridges (or bulges) cease at the transition between the juxtaparanodal and the paranodal segment, that is, at the level where the myelin sheath turns to the axon in order to terminate along the paranodal part of the CON segment (see Fig. 3–9D). In many juxtaparanodes, particularly in those of large alpha motor axons in old animals,424 the more nodal parts of the axon ridges break up into thin, winding, irregular processes that are embedded in the adaxonal Schwann cell compartment (Fig. 3–13A; see also Fig. 3–9A). This complex of interwoven axonal and Schwann cell profiles is the axon–Schwann cell network (ASN).20,151,398 Both kinds of profiles contain DLBs and MVBs, some of which are acid phosphatase positive and should be classified as secondary lysosomes and residual bodies (see Fig. 3–13A, inset).149,155

54

Structure of the Peripheral Nervous System

Ax A

FIGURE 3–13 A, Electron micrograph of transverse section close to the transition between the juxtaparanodal and the paranodal segments; alpha motor nerve fiber, L7 ventral spinal root, adult cat. The axon crests inside the myelin crests have fragmented and form, together with the inner cytoplasmic Schwann cell compartment, typical axon–Schwann cell networks (asterisks). Arrows point at myelin crests covered by a substantial layer of Schwann cell cytoplasm; arrowheads, at mitochondrion bags. Bar: 2 ␮m. Inset, Detail from axon–Schwann cell network after incubation for demonstration of acid phosphatase. One of the larger axon profiles contains a multivesicular body filled with the reaction product (arrow). Bar: 0.5 ␮m. (Inset micrograph courtesy of Dr. K. Gatzinsky.) B, Electron micrograph of longitudinal section through a paranodal segment that illustrates the attachment of the myelin sheath on the paranodal axon (Ax); alpha motor nerve fiber, L7 ventral spinal root, adult cat. Asterisks identify two “ear-of-barley” arrangements in the terminal cytoplasmic spiral of the Schwann cell. Some of the crosscut cords in the spiral are indicated with asterisks. G ⫽ nodal gap; W ⫽ nodal gap wall. Bar: 0.2 ␮m.

The ASN has the ability to trap, sequester, and probably disintegrate effete axonal constituents and may reflect the role of the PNP region as a site where local lysosomal degradation of transported materials can take place.148,149,153,398 The network appears strikingly hypertrophic in a number of neuropathies36,398 and just the opposite in CNTF knockout mice, in which it develops early during the postnatal period and then withers away.152 The major changes in the organization of the axon, as seen when tracing along the

juxtaparanodal segment toward the node, are the gradual reduction of the transverse axon area, the increasing delicacy and final disintegration of the axon ridges, and the displacement of MTs from the periphery to the core of the axon. As a result, the MT concentration increases in proportion to the reduction of the transverse area of the axon.20,433 At its nodal end the juxtaparanodal axolemma contains voltage-gated (Kv) potassium channels of the delayed-rectifier type (␣ subunits Kv1.1 and Kv1.2 and the

Microscopic Anatomy of the Peripheral Nervous System

cytoplasmic ␤ subunit Kv␤2) in complex with contactinassociated protein-2 (Caspr-2), so far of unknown significance in the mature nerve.8,56,269,321,333,346 The Paranodal Segment This contains the termination—the nodal end—of the outer, abaxonal cytoplasmic Schwann cell compartment. The abaxonal Schwann cell plasma membrane contains Kv1.5 potassium channels.269 The abaxonal Schwann cell cytoplasm coats the inturning myelin sheath, faces the perinodal space, and extends into the nodal gap, where it forms the Schwann cell brush border. Here the Schwann cell compartment is virtually devoid of mitochondria but rich in moderately electron-dense round or tubelike “juxtanodal Schwann cell bodies” 50 to 200 nm in size and of unknown significance (see Fig. 3–11C).32 Their localization to a cytoplasmic domain between numerous mitochondria and a brush border, as well as their general ultrastructural appearance, recall the “dense apical tubules” that appear in proximal tubule cells after oil infusion.78 The paranodal segment is defined by the termination of the myelin sheath. Viewed in a median longitudinal section, each terminating myelin lamella splits into two leaflets that enclose a drop-shaped portion of Schwann cell cytoplasm: the terminal cytoplasmic pocket (see Figs. 3–11B, 3–12A, and 3–13B). The whole set of pockets thus displayed in a longitudinal section is the sagittal representation of a single, very thin and continuous cord of Schwann cell cytoplasm that encircles the paranodal axon cylinder in a complex helical manner. Each turn of the spiral corresponds to the termination of one myelin lamella. The total length of this terminal cytoplasmic spiral (TCS) is about 2500 ␮m in a large cat alpha motor fiber and provides a membranous mantle area of about 800 ␮m2 (see Table 3–1). The TCS represents the nodal margin (the lateral belt) of the hypothetically unrolled myelin sheath (see Fig. 3–16 later). It connects the inner, adaxonal Schwann cell compartment with the outer, abaxonal, one. In this respect the TCS is equivalent to the cytoplasmic spiral (Golgi-Rezzonico spiral) associated with an incisure of Schmidt-Lanterman, although it is less than half the length of the latter. The inherent stability of the TCS, as well as that of the Schmidt-Lanterman cytoplasmic spiral, depends on E-cadherin and MAG.111,430 The turn of the TCS that belongs to the innermost myelin lamella attaches to the paranodal axon segment farthest from the node. Subsequent lamellae then terminate in consecutive order up to the node. Only some 10% to 20% of the lamellae of a thick fiber attach directly to the axolemma. The remaining 80% to 90% coil on top of one another in sets of 10 to 25 turns. In a longitudinal section this gives the picture of “ear-of-barley”–like aggregations of terminal cytoplasmic pockets (see Figs. 3–12A and 3–13B) that penetrate the surrounding compact myelin for a distance of 0.5 to 1.5 ␮m. The specific staining properties of the TCS

55

explain the light microscopic appearance of the so-called spinous cuffs or spiny bracelets of Nageotte.190,470 The TCS is claimed to be rich in Na⫹ and Ca2⫹.102 The turns of the TCS that join the axolemma indent it and give it a serrated outline. They attach to it by gap junction–like membrane complexes that, when viewed in freeze-fracture preparations, form the so-called transverse bands188,248,345,459 and complex patterns of IMPs in both P and E faces. The various types of membrane connections noted between the paranodal axon segment and the TCS are as a whole referred to as the glia-axon-junction (GAJ) complex,204 the largest of the mammalian cell adhesion complexes.56,314 The GAJ complex includes a periaxonal space, 3 to 5 nm in width, that, at least in its most nodal aspect, opens into a node gap (see below). The turns of the TCS are joined together by a system of more or less continuous tight junctions.365 The GAJ complex and the TCS react strongly with antibodies against connexin 32 and GM1,275,400 both substances known to take part in cell adhesion and intercellular communication. In addition, the GAJ complex is endowed with Caspr-1 (paranodin) and contactin on its axonal side and neurofascin-155 on its Schwann cell side.42,98,267,430 The demonstration of Na⫹-K⫹-Cl⫺ cotransporter protein in the TCS, in nodal compartments of the Schwann cell, and in association with the Schmidt-Lanterman cytoplasmic spiral suggests involvement in K⫹ redistribution during and after impulse activity.4 Disordered and/or disrupted GAJ complexes correlate strongly with axonal dysfunction in various neuropathies.57,158,206,367,418,420 Those turns of the TCS that belong to the outermost two to five myelin lamellae do not attach to the axolemma but are separated from it by a space 20 to 50 nm in height. They form, in this way, the “ceiling” of a nodal gap recess. The most nodal turn of the TCS (i.e., the cytoplasmic pocket of the outermost myelin lamella) is high and voluminous and constitutes the nodal gap wall. The subaxolemmal part of the AR is particularly well developed in the paranodal axon segment. Whether it is involved in the impulse-generating mechanism is not known, although the GAJ complex and its axonal and Schwann cell environments apparently release Ca2⫹ during impulse activity.102,242,441,468,469

Node of Ranvier The nodes of Ranvier are the only sites along a myelinated nerve fiber that support the fast de- and repolarization process necessary for generation of action potentials.449 This ability is mainly a result of the high concentration of voltage-sensitive Na⫹ channels in the nodal axolemma. The ultrastructural study of nodes of Ranvier is hampered by their vulnerability to manipulation and chemical influences.176 Serious artifacts develop when the preparatory protocol includes dehydration, as it usually does.27,438,468,469 In particular, the nodal gaps and the paranodal myelin

56

Structure of the Peripheral Nervous System

become distorted as the two myelin sheaths that meet at the node shrink, are torn apart, and split in relation to their TCSs. As a result, the nodal gap appears as a welldefined and relatively large and clear opening in the myelin sheath, a picture not at all consistent with the in vivo or pre-dehydration appearance, where nodal gaps are hard to define.27,154 The thicker the myelin sheath, the more serious are the preparatory artifacts. These difficulties probably explain why there are few systematic studies dealing with quantitative aspects of nodal ultrastructure in large fibers.20,21,274,318,327,351,438 The nodes of Ranvier can, just like other parts of a myelinated fiber, be described as consisting of two concentrically arranged zones: an outer Schwann cell compartment (here without myelin) and an inner axonal compartment. Cell adhesion molecules (CAMs) both of the neuron-glia and of the neuronal type coexist specifically in the node of Ranvier, where they appear in the Schwann cell compartment as well as in the axon. This localization of both CAMs in the adult myelinated nerve fiber emphasizes the importance of nodal structural integrity.220,340 The endoneurial territory outside a node of Ranvier and at the meeting point of two adjacent internodes is denoted the “perinodal space.” The Schwann Cell Compartment This consists of three regions separated from the perinodal space by the basal lamina: (1) the outer demarcation, (2) the nodal gap, and (3) the nodal gap walls (see Figs. 3–9F and 3–12A). The outer demarcation is defined by more or less overlapping extensions of the outer cytoplasmic compartments of the two adjoining internodes. The extensions are referred to as nodal collars. The nodal collars send a conventional brush border of radially arranged microvilli into the nodal gap. The brush border, as a rule, is particularly well developed in those sectors of the outer demarcation where the nodal collars are direct continuations of a mitochondrion bag. In some sectors of the outer demarcation, the nodal collars of the paranodal segments overlap and join, forming five-layered membrane complexes reminiscent of tight junctions. At some points the collars are joined by clumps of lamellar material.33 In other sectors of the same demarcation region, the nodal collars are minimal and lie 0.1 to 0.2 ␮m apart. Here the nodal gap is separated from the perinodal space only by the Schwann cell basal lamina. Solitary tufts of Schwann cell microvilli are common outside the nodal gap facing the perinodal space. Similar tufts are occasionally observed elsewhere on the outer cytoplasmic Schwann cell compartment (see Pannese et al.310). The nodal gap, when viewed in a median longitudinal section, is shaped like an isosceles trapezoid (see Fig. 3–12A), the longer and shorter sides of which correspond to the nodal axolemma and the plane of the outer demarcation, respectively. The slanting sides of the trapezoid are formed

by the nodal gap walls.33,351 The nodal gap contains the Schwann cell brush border and the gap matrix substance. The matrix substance, containing glycosaminoglycans, occupies the extracellular space of the gap. Its role as a cation exchanger and/or buffer has been much discussed but remains elusive.230,233 Its strong and capricious affinity for heavy cations such as Ag2⫹, Cu2⫹, Fe2⫹, and Pb2⫹ is noteworthy in connection with electron histochemistry based on the precipitation of metal salts.5,183,222,229,233,234,451,471–473 The metallophilia of the nodal gap substance lies behind the appearance of both the so-called nodal cementing disc and the shorter bar of Ranvier’s cross.186 The latter develops when the CON segment together with the nodal gap display metallophilia, such as after incubation in a solution of AgNO3. The nodal gap contains NG2, a chondroitin sulfate proteoglycan, presumably delivered into the gap from the outside by “perineural fibroblasts.”255 Speculation has it that NG2 might restrain an otherwise sprouting-happy nodal axon. The microvilli of the brush border end less than 5 nm from the axolemma. In a large fiber (D ⫽ ~15 ␮m), the height of the nodal gap varies in different sectors of the node from 0.1 ␮m to as much as 2 ␮m. There are about 800 to 1000 microvilli, each 70 to 80 nm in diameter and on the average 1 ␮m long, in a nodal gap (see Table 3–1). The seven or eight microfilaments (F-actin) in a microvillus and the lack of a terminal web in the nodal collars suggest structural kinship with the brush border of the proximal tubule cell of the kidney (see Kenny and Booth215). The actin filaments connect to the cell membrane of a microvillus with three ERM proteins: ezrin, radixin, and moesin.265,363 Freeze-fractured microvilli show P and E faces in which 45% of the intramembranous particles are about 10 nm, a size considered to correspond to that of voltage-sensitive Na⫹ channels.224,344,450 The nodal Schwann cell microvilli are furnished with inwardrectifying K⫹ channels of the IRK family269 and possibly are able to perform “K⫹-buffering” (see Orkand et al.307). The nodal gap forms low recesses that extend for a distance of 0.1 to 0.4 ␮m both proximally and distally between the axolemma and the most nodal turns of the TCS of the paranodal segment. Calculations based on morphometric data show that, in a large cat alpha fiber, the Schwann cell faces the nodal gap with a cell membrane area of about 250 ␮m2. Approximately 75%, 20%, and 5% of this area are provided by the brush border, the nodal gap walls, and the ceiling of the gap recess, respectively (see Table 3–1). The use of watersoluble embedding media after glutaraldehyde–osmium tetroxide fixation or the use of glutaraldehyde–tannic acid fixation followed by fixation in osmium tetroxide and potassium ferricyanide produces more or less closed nodal gaps with a maximal close packing of the microvilli (see Fig. 3–12A).27,32,33 The Schwann cell plasma membrane of the villi and related to the nodal gap (and to the paranodal axon segment) may contain, in addition to acetylcholine receptors, Na⫹ and fast and slow K⫹ channels.72,74,449,452 It has been suggested that some of the Na⫹ channels integrated

Microscopic Anatomy of the Peripheral Nervous System

57

in the nodal axolemma may have been synthesized in the Schwann cell and delivered to the nodal axolemma via the node gap microvilli169 (however, see Brophy56). The nodal gap walls are formed by the comparatively high and voluminous cytoplasmic pocket of the outermost myelin lamella of the two meeting internodes (see Figs. 3–12A and 3–13B). They are in direct communication with the paranodal outer Schwann cell compartment and contain juxtanodal Schwann cell bodies and occasional mitochondria. The Axonal Compartment This consists of the nodal axon segment, which corresponds to the midregion of the CON segment (see Figs. 3–9F and 3–11B). The segment is about 1 to 1.5 ␮m long regardless of fiber size.351 It extends between the most nodally and closely attached turn of the TCS of each of the two adjoining paranodes. The segment is usually barrel shaped with a diameter a few percent larger than that of the adjacent paranodal segments. The size of the mantle area of the nodal axon segment increases linearly with increasing axon size. In cat spinal roots, the nodal membrane area varies from about 4 ␮m2 in the smallest fibers to about 30 ␮m2 in the largest ones.351 The nodal axolemma (review by Salzer357) is characterized by a high content of P-face particles (about 1500/␮m2)345 (see Fig. 3–12) and an inside coating of electron-dense material. The inside coating, or the nodal axoplasmic cortex, is about 30 nm thick after conventional electron microscopic preparation (see Fig. 3–9F) and up to 100 nm after application of unconventional preparative methods (see Figs. 3–12A and B). The undercoating contains actin,478 ␤IV spectrin,19 and the 480- and 270-kDa isoforms of ankyrin G, which probably bind directly to the voltagegated Na⫹ channels459 and anchor them to the cytoskeleton by way of neurofascin, NrCAM, and L1.18,80,204,254,356 After staining with ferric ions and ferrocyanide, the nodal axolemma develops a thick, electron-dense precipitate, a reaction taken to indicate the presence of Na⫹ ion channels and their anchoring support.325,451 The thickness of the nodal axoplasmic cortex as shown in Figure 3–12B corresponds well with the assumed size of the ankyrin-neurofascin complex.204 In very small fibers (D ⫽ 1 to 2 ␮m), in which the nodal gap contains just a few microvilli, the inside coating is patchy and restricted only to sites related to the tip of a microvillus. In some nodes the axolemma projects spine- or crestlike outgrowths.33,438 The crests run across the longitudinal fiber axis. Both kinds of projections contain AR profiles. Coated pits are common in the axolemma at the bases of spines and crests and in relation to the nodal recesses. The presence of coated invaginations also in the nodal collars may suggest an interplay between axon and Schwann cell mediated via the nodal gap.33,116,238 The Schwann cell:axon membrane ratio of the node of Ranvier is plotted against axon diameter in Figure 3–14, which illustrates that the Schwann cell:axon membrane

FIGURE 3–14 Ratio between nodal Schwann cell membrane area and nodal axon membrane area in the cat and plotted against axon diameter as estimated in internodal main regions. Open circles ⫽ L7 ventral root axons; filled circles ⫽ L7 dorsal root axons. (From Rydmark, M., and Berthold, C. H.: Electron microscopic serial section analysis of nodes of Ranvier in lumbar spinal roots of the cat: a morphometric study of nodal compartments in fibres of different sizes. J. Neurocytol. 12:537, 1983, with permission.)

ratio reaches its highest values of 10 to 15 in the d interval of 5 to 10 ␮m. Provided that a high ratio reflects the ability of the Schwann cell to control the ionic milieu outside the axolemma in a way favorable for impulse conduction, nerve fibers with the highest conduction stamina would appear in the 5- to 10-␮m interval. This is the size of large gamma axons, small to medium-sized alpha axons, and type II afferents, which are all known to fire tonically (i.e., have high conduction stamina). This and the suggested accumulation of K⫹ channels in the paranodal axolemma recall the role of the Schwann cells as potassium clearing units outside squid giant axons.1,58

The Constricted Axon Segment The length of the CON segment (Fig. 3–15A; see also Fig. 3–8) is about 6 to 10 ␮m and increases little with fiber size.351 The axolemma of the CON segment is formed by a median nodal field, flanked symmetrically by a proximal and a distal paranodal field. Voltage-clamp experiments on normal mammalian myelinated fibers have shown that fast-depolarizing Na⫹ currents are generated at the nodes during impulse conduction, whereas fast-repolarizing K⫹ currents, well known to be present at active amphibian nodes, are lacking.48,49,73,218 A fast K⫹ current appears, however, at the active mammalian node if the myelin is loosened

58

Structure of the Peripheral Nervous System

P

I N

D

E

2 μm

Microscopic Anatomy of the Peripheral Nervous System

from the paranodal axon segments.73 These observations, in combination with results from freeze-fracturing and immunocytochemistry, have given the picture of a CON segment in which the nodal axolemma is rich in Na⫹ channels, slow tetraethylammonium-blockable K⫹ channels, and Na⫹,K⫹-ATPase5,6,36,103,370,378,410,466 but devoid of fast 4-aminopyridine–blockable K⫹ channels.14,37,75,96,219,341,343 The paranodal axolemma, in contrast, seems to be rich in fast K⫹ channels but poor in Na⫹ channels and slow K⫹ channels, and may contain chloride channels.407 In addition to its numerous large P-face particles and thick axoplasmic cortex, the nodal field of the axolemma of the CON segment is characterized by ␣-bungarotoxin binding sites140 and numerous filipinsterol complexes.40 The axolemma of the paranodal fields and the cytoplasm of the adjoining TCS contain Ca2⫹-activated ATPase.258 As a whole, the axolemma of the CON segment is 5'-nucleotidase positive, becomes impregnated when treated with bismuth iodide, and binds ruthenium red and lectins.88–92 The magnitude of the Schwann cell membrane area provided by the TCS that faces the extracellular space outside the CON segment (the periaxonal and the nodal gap spaces) is about 10 times that of the axolemma of the CON segment. Elsewhere along the fiber, this Schwann cell:axon membrane ratio is 1:1.33 The Schwann cell membrane domains that directly face the axolemma of the CON segment (i.e., some or all turns of the TCS and the ends of nodal microvilli) appear to be anchored to the axonal cytoskeleton by filamentous transmembrane linkers.204

59

The appearance of the axoplasm in the CON segments and their immediate proximal and distal juxtaparanodal extensions is highly variable with regard to both the amount and the distribution of membranous organelles and granular material.20,21,28,29,328 The various patterns run the gamut from an appearance that, except for the high concentration of MTs, barely deviates from that noted in the main internodal region, to that of an axoplasm crammed with organelles and granular material (see Figs. 3–11B and 3–15). As a rule, DLBs and MVBs (i.e., retrogradely transported organelles) occupy the axoplasm in the distal half of the CON segment and the adjacent part of the juxtaparanodal axon. Organelles transported anterogradely (vesiculotubular profiles) occupy the proximal half of the CON segment and adjacent parts of the juxtaparanodal axon. They are also numerous in the nodal segment, and some appear in the distal paranode. Such a proximodistal segregation of retrogradely and anterogradely transported elements is easily understood in terms of a local reduction of the axonal stream at the node. This is similar to the pattern noted at experimentally induced blocks of axoplasmic transport.79,104,390,392,432 The obvious variations noted in the segregation pattern, the presence of lysosomes, and the low occurrence of DLBs and MVBs outside the PNP regions (see Table 3–1)26,29,149,150 are, however, difficult to understand as passive rheologic phenomena. One possible explanation could be that these variations indicate differences in the functional/metabolic state of a neuron in general and of a PNP region in particular. The various patterns may reflect different operative

FIGURE 3–15 Electron micrographs of longitudinally sectioned nodes of Ranvier; alpha motor nerve fibers, L7 ventral spinal roots (A–D) and medial gastrocnemius nerve (E), adult cat. A, Two consecutive digitized electron micrographs are aligned and fused with an arithmetic minimum filter. The strandlike arrangements of amassed vesiculotubular organelles in the axoplasm of the proximal internodal end region (thin arrows) are particularly distinct. CON ⫽ constricted part of the nodal region; M ⫽ myelin sheath; N ⫽ nodal region. Thick arrow points towards the cell body. Bar: 2 ␮m. (From Pascher, R., Berthold, C. H., and Rydmark, M.: Computer-assisted simulation of high-voltage electron microscopy using serial images recorded by conventional transmission electron microscopy. J. Electron Microsc. [Tokyo] 51:113, 2002, with permission.) B, Detail from a proximal paranodal segment. Arrows point to axoplasmic reticulum profiles. VT ⫽ vesiculotubular profiles; D ⫽ distal; P ⫽ proximal. Bar: 1 ␮m. C, Detail from a proximal paranodal segment. The axoplasm is dominated by longitudinal strands of vesicotubular profiles (asterisks). D ⫽ distal; P ⫽ proximal. Bar: 1 ␮m. D, Detail from the transition between the distal paranodal and juxtaparanodal segments. The axoplasm is rich in microtubular profiles (MT) with associated vesicles and mitochondria (M). Asterisk indicates a strand of vesicotubular profiles. D ⫽ distal; P ⫽ proximal. Bar: 1 ␮m. E, Stereo view of a node of Ranvier from a HRP-transporting medial gastrocnemius nerve. The stereo pair is generated from a digitized image series of eight consecutive sections by fusing with an arithmetic minimum filter. The axoplasmic HRP activity appears as black rounded bodies of different sizes. The axoplasm proximal to the constricted segment contains very few HRP-positive bodies. The stereo view illustrates the continuous strings of HRP-positive bodies that extend from the distal internodal end region (D), where they are comparatively large, through the constricted axon segment at the node (N), where their size decreases and they become incorporated in the vesiculotubular strands. From here, small HRP-positive bodies continue to the proximal level (P) of the constricted axon segment, where they seem to aggregate into a thin, transverse, disclike configuration (white vertical bar). Uncontrasted sample. The viewing angle is 7 degrees. Bar: 2 ␮m. (From Pascher, R., Berthold, C. H., and Rydmark, M.: Computer-assisted simulation of high-voltage electron microscopy using serial images recorded by conventional transmission electron microscopy. J. Electron Microsc. [Tokyo] 51:113, 2002, with permission.)

60

Structure of the Peripheral Nervous System

steps in a local interaction between, on the one hand, the axoplasm of a CON segment and its immediate juxtaparanodal extensions, including the ASN, and, on the other hand, various transported elements. There are observations that indicate that the axoplasm of the CON segment may interact actively both with material transported retrogradely from the terminal region, such as intramuscularly injected horseradish peroxidase (HRP) (see Fig. 3–15E),26,29,149 and with materials transported anterogradely from the soma, such as newly synthesized proteins.7,208,323,478 Our studies of alpha motor neuron PNP regions that partake in retrograde transport of intramuscularly injected HRP have shown organelle patterns similar to those of unexposed animals but with a specific temporal order of appearance after the injection.29,313 This segregation and ordered trapping of HRP-positive organelles more or less ceases as the alpha axon enters the CNS (see Fabricius et al.110). Evidently CON segments of peripheral myelinated nerve fibers are sites where elements of the retrograde endocytotic stream are forced into close, probably interactive, contact with elements that belong to the anterograde exocytotic (lysosomal) stream.198,203,440 Extreme segregation patterns appear in the PNP regions of the distal stump in the first days after a crush injury; the axoplasm distal to nodes of Ranvier accumulates large numbers of mitochondrion-like organelles and DLBs.241,453 The MTs of highly segregated CON segments form coherent bundles of up to 20 members. 21,282 The bundles usually project a few micrometers proximally and distally into the juxtaparanodal segments before they become separated into smaller bundles or individual MTs. The various organelles are organized outside and along the MT bundles, to which they seem to cling in a grapevine fashion (see Fig. 3–15A–D). In more relatively empty-looking and less segregated CON segments, the bundling of MTs is less conspicuous and the MTs are more dispersed. No doubt the distribution of MTs in the axonal cross section will, in view of their fundamental role as the rails and promotors of axoplasmic transport, influence the transport capacity of the CON segment with regard to both the amount and the size of transported bodies.30,328 A hypothetical mechanism that was able to influence the distribution of MTs and operated in the PNP region would give the CON segment the combined properties of a sieve and a throttle valve. The claims that the Schwann cells may influence the axon, and in particular the NF and the MT of the PNP regions, are indeed noteworthy in connection with the axoplasmic transport through CON segments.82–84,288 The possibility exists that a highly segregated CON segment could be such a hindrance to the progression of anterogradely and retrogradely transported organelles that turnaround phenomena391 are triggered in the orifices of the CON segment. As a result, the organelles would begin to shuttle back and forth in the respective internode.29,30

The Node–Paranodal Apparatus Landon and Williams, who gave the first comprehensive ultrastructural description of the PNP region in mammalian nerve fibers more than just a few micrometers thick,231,463 used the term paranodal apparatus, referring to the Schwann cell mitochondrion bags, the nodal gap microvilli, and the nodal gap substance. This emphasized the significance of these parts of the Schwann cell in nodal function, which at that time assumed that a fast K⫹ current repolarized the nodal axolemma during the action potential but did not account for the interference of the CON segment with axoplasmic transport. In view of current knowledge and ideas, it seems reasonable to extend the concept of a paranodal apparatus to include the TCS of the Schwann cell, the whole CON segment, and the immediately adjoining parts of the proximal and distal juxtaparanodal segments and their associated ASNs. With this conceptual broadening, the term node–paranodal apparatus could be more informative. Many interesting speculations have been presented with regard to the functional significance of the node– paranodal apparatus and the ways its various components may cooperate102,279,426,441,459 (see also Landon228,231). Hard facts have, however, been remarkably meager, which is surprising in view of the many fundamental functions expressed by the PNP region and its central role in a host of peripheral neuropathies. This is, however, no longer so. During the last two decades an almost overwhelming amount of new molecular and genetic data focusing on the PNP region have been presented.322,414

THE SCHWANN CELL Schwann cells are the satellite cells of peripheral nerves. The majority originate in the neural crest,181 although others appear later as they migrate along the neural tube. The suggestion that Schwann cells in motor nerves arise in the ventral portion of the neural tube226 was supported by experiments using xenoplastic grafts334 and also by autoradiographic studies.457 The contributions of the neural tube and neural crest have been discussed by several authors; other possibilities seem less likely.9,71,458 In the past they have been thought to be merely supporting cells for the production of the myelin sheath where appropriate, but recently their effects on the axon have become a matter of great interest. During embryonic development of early Schwann cells, two growth factors are particularly important: endothelin and ␤-neuroregulin-1. They seem, however, to be required for survival and some aspects of migration rather than establishment of the peripheral nervous system (PNS) glial cell lineage, and it is still unclear what initial signal is required. The transcription factor Sox10 may be important in early development. These

Microscopic Anatomy of the Peripheral Nervous System

factors have been reviewed by Jessen and Mirsky.210 The development of transgenic mouse models to investigate the role of individual proteins in development of Schwann cells and formation of the myelin sheath, axon, cytoplasm, and basal lamina have emphasized the vital role the Schwann cell plays in nerve function and that, without a healthy, fully functional Schwann cell, the axon will degenerate. As already mentioned, Schwann cells multiply dramatically during development. Once they have established a one-to-one relationship with an axon and myelination has started, no further mitosis occurs unless a pathologic process occurs. If contact with a myelin sheath is lost, the Schwann cell is free to divide again. Although there superficially appear to be two types of Schwann cells, those associated with myelinated and with unmyelinated axons, numerous studies have shown that this is due to axonal signals and that they can change from one to another on the receipt of different signals. Recent studies have defined the proteins expressed on the surface of Schwann cells. All Schwann cells from embryonic day 16 in rat nerve express S100 and O4. Myelinating Schwann cells express Krox20, periaxin, the peripheral myelin proteins P zero (P0), peripheral myelin protein-22 (PMP22), P1, myelin basic protein, P2, MAG, connexin 32, proteolipid protein, and galactosylceramide.

FIGURE 3–16 Diagrammatic representation of an unrolled Schwann cell. The white areas indicate the presence of cytoplasm. The lightly stippled areas at the outer and inner portions of the cell represent semicompacted myelin and the heavily stippled areas, compact myelin. Cytoplasmic channels form a continuous network in both compact and semicompact myelin domains. Cytoplasm is present along the whole margin of the cell, the outer and inner “belts” constituting the marginal rim of the abaxonal and adaxonal Schwann cell cytoplasm, respectively. The lateral belt is the terminal cytoplasmic rim at the node of Ranvier, the outer portion bearing the nodal microvilli. One complete and one incomplete Schmidt-Lanterman incisure have been included together with two longitudinal incisures. (From Mugnaini, E., Osen, K. K., Schnapp, B., et al.: Distribution of Schwann cell cytoplasm and plasmalemmal vesicles [caveolae] in peripheral myelin sheaths: an electron microscope study with thin sections and freeze-fracturing. J. Neurocytol. 6:647, 1977, with permission.)

61

Schwann cells associated with unmyelinated axons express a different set of proteins: p75, neural CAM, L1, glial fibrillary acidic protein, growth-associated protein-43, A5E3, and Ran-2. (For a detailed discussion of the signals that determine Schwann cell identity, see Jessen and Mirsky.210) Abnormalities in the genes for some of these proteins have been related to some human genetic diseases; for example, abnormalities in PMP22 and P0 are related to various subtypes of Charcot-Marie-Tooth (CMT) disease type 1 and mutations in connexin 32 are related to X-linked CMT disease (see Chapter 76). Myelin is formed by the compaction of layers of the Schwann cell surface membrane. A diagram of the myelinating cell unrolled shows a trapezoid sheet (Fig. 3–16). The adaxonal Schwann cells form the innermost portion of cytoplasm adjacent to the axon while the abaxonal region is that outside the myelin sheath. There are cytoplasmic channels (incisures of Schmidt-Lanterman) connecting the abaxonal and adaxonal regions, thus permitting the movement of metabolites from one to the other. The nucleus is midway between nodes on the outer surface of the myelin sheath. Both surfaces of the cell are connected to the myelin sheath by extensions of the cell membrane, the inner and outer mesaxons (Fig. 3–17). These are occluded by tight junctions. There is very little cytoplasm and few organelles except in the nuclear and nodal regions.

62

Structure of the Peripheral Nervous System

responsiveness to antigen, cytokines, and T cells suggest that this may be significant in the pathologic development of inflammatory demyelinating neuropathies245 and also possibly in hereditary demyelinating neuropathy.162

THE MYELIN SHEATH Compact myelin is formed by the spiral wrapping of Schwann cell plasma membranes. This forms an alternating structure of light and dark lines. The light band is also divided by a faint line known as the less-dense line. The dark line (the major dense line) is formed by the apposition of the intracytoplasmic surfaces of the Schwann cell and the less-dense line from the juxtaposition of the extracellular surfaces (Fig. 3–19). X-ray crystallography showed that the periodicity of myelin in fresh, unfixed nerve is 18 nm but fixed material shows a considerable reduction to 12 to 15 nm depending on the processing involved. Compaction may also be affected by osmotic effects and x-irradiation and particularly by abnormal serum proteins. Changes in compaction are most commonly seen in the neuropathy related to immunoglobulin M paraproteinemia. The result is an increase in periodicity to 2.5 times normal because of the failure of the extracellular spaces to connect tightly together. In this situation the less-dense line is seen as two components each the correct distance from the major dense line but widely separated from each other (Fig. 3–20). For a discussion of the detailed structure of the myelin sheath, the reader is referred to Chapter 19. FIGURE 3–17 A, An electron micrograph of a transverse section through a small myelinated nerve fiber from a normal adult human sural nerve. Ax ⫽ axon; My ⫽ myelin; R ⫽ Reich granule; Sn ⫽ Schwann cell nucleus. Bar: 0.5 ␮m. B, Detail from A showing inner and outer mesaxons (arrows). Bar: 0.2 ␮m.

Unmyelinated Nerve Fibers R. H. M. King HISTORICAL BACKGROUND

Whereas there may be lipofuscin granules in the Schwann cells of Remak fibers, these are not found where the Schwann cells are associated with myelinated fibers. However, the age-related acid phosphatase–positive granules of myelinating Schwann cells are the ␲ granules of Reich. These are lamellar structures about 1 ␮m in length, with metachromatic properties resulting from the presence of sulfatide or phosphatide (Fig. 3–18), and may have a lysosomal function. They are not found in normal Remak fibers but may persist in the cytoplasm after myelinated fiber degeneration even when the Schwann cell has become associated with unmyelinated axons. Schwann cells can express major histocompatibility complex class II molecules and hence can act as antigenprocessing cells. In vitro observations on Schwann cell

Very fine fibrous structures were described by Remak in 1838 when he teased autonomic nerve bundles apart in water. He distinguished myelinated nerve fibers (tubuli primitivi) from other delicate fibers (fibrae organicae). Although he identified axons within the tubuli primitivi, there is nothing to suggest that he realized that the fibrae organicae also contained axons.338 It is a convenient shorthand to refer to the bundles of Schwann cells containing unmyelinated axons as Remak fibers. In 1839, Schwann identified the nucleated corpuscles on the fibers as cells.369 It was not for another 60 years, however, that Tuckett, using an osmic acid technique, identified both sheath cells and cores (i.e., Schwann cells and axons). He also described the changes produced in these fibers by degeneration.436

Microscopic Anatomy of the Peripheral Nervous System

63

FIGURE 3–18 Electron micrograph of Reich granules from control human sural nerve. Bar: 0.5 ␮m. Inset illustrates the lamellar structure at higher magnification.

FIGURE 3–19 Electron micrograph showing detail of the myelin sheath structure in a control human sural nerve. Arrows indicate the major dense lines and arrowheads the intermediate lines. Bar: 0.5 ␮m.

FIGURE 3–20 Electron micrograph showing detail of an abnormal myelin sheath. Part of the sheath has normal compact myelin (CM) where major dense lines (white arrows) are separated by a less-dense intermediate line that appears double in places. In the adjacent region of the myelin, the intermediate line has failed to compact and the two components are separated by a wide gap (double-headed arrow). Bar: 0.05 ␮m.

64

Structure of the Peripheral Nervous System

Unmyelinated axons vary in diameter from 0.2 to 3 ␮m, and so are difficult to see by light microscopy, especially in paraffin sections whose thickness is greater than the axon diameter. It was only with the advent of electron microscopy that it was confirmed that there was no myelin at all on these small axons in Remak fibers. Phase-contrast microscopy had previously been interpreted to show a very thin myelin sheath, but electron microscopy revealed that some axons did not possess any myelin at all.185 Electron microscopy later confirmed that the axons were enfolded along their length by a series of separate Schwann cells; earlier workers such as Nageotte had thought that this was a syncytium as opposed to separate cells.281 He described the fibers as forming interlacing trabeculae that divided and reunited, with each containing several axons (see Relationship of Unmyelinated Axons with Schwann Cells later). There is very little overlap in size between myelinated and unmyelinated fibers.297 Unmyelinated axons range in size from 0.1 to 3 ␮m. It has been shown that the presence of a myelin sheath does not increase conduction speed for very small fibers.308 In addition, myelin is complex and easily damaged, so an unmyelinated fiber should be more robust and less sensitive to metabolic insults.

DEVELOPMENT Schwann cells develop from the neural crest ectoderm and already show a basal lamina by 9 weeks’ gestation in humans. There are no reports of naked axons being identified before Schwann cells can be seen, as may be the case in regeneration, although before Schwann cells have a basal lamina, they are difficult to distinguish from axons or fibroblast processes. The axons lie in large bundles each enfolded by a single Schwann cell process (Fig. 3–21). In a study of a 14-week fetal nerve, these bundles were seen to contain up to 90 small axons most of which are in contact only with each other, with only the outer ones contacting the surrounding Schwann cell process.146 The apparently greater density of Schwann cell nuclei in transverse sections of fetal compared with adult nerve is probably partly due to the cells being shorter (rounder) than in adult nerves. The Schwann cells divide in response to axonal signals and send cytoplasmic processes out to encircle individual axons on the edge of the bundle to form promyelin fibers with a 1:1 ratio of Schwann cell to axon. This is a necessary precursor to myelination. Myelination commences at different times in different parts of the PNS but seems to start at about 18 weeks’ gestation in the sural nerve in humans.296 By 21 weeks the myelinated fiber density is approximately 5000/mm2, with on average nine lamellae per fiber.376 It has been estimated that Schwann cells only need to divide three or four times during development to reduce the axon:Schwann cell ratio from

FIGURE 3–21 Electron micrograph of nerve from 24-week-old fetus showing a bundle of numerous small unmyelinated axons (asterisks) embedded in one Schwann cell. Schwann cell processes (arrows) only contact the outermost axons in the bundle. Two single, larger unmyelinated axons (A) have formed a one-to-one relationship with a single Schwann cell process (promyelin fiber). Bar: 1 ␮m.

1:100 to 1:12 to the 1:6 seen at birth.146 Further development after birth reduces this still further until eventually only one or two fibers remain small and unmyelinated. In general this progression from small fetal axons to larger mature ones depends on the axon diameter; as they become larger, they become separated and then myelinated. Further axonal growth occurs after myelination.465 The axonal signal leading to growth of some axons and not others is still unclear. Growth factors are clearly involved; high doses of glial cell line–derived neurotrophic factor (GDNF) will change the phenotype of an axon from unmyelinated to myelinated.196 GDNF is also associated with Schwann cell proliferation and increase in size of unmyelinated axons together with a reduction in the number of axons per Remak fiber. The mechanism could be a reaction of the Schwann cells to GDNF or a reaction of a subpopulation of GNDF-dependent, c-Ret–expressing unmyelinated axons to GDNF. In vitro studies show that both nerve growth factor and GDNF will induce myelination.196,278 The axons that are destined to remain unmyelinated become rearranged to lie in closer contact with a Schwann cell process but separated from each other,

Microscopic Anatomy of the Peripheral Nervous System

forming a Remak fiber. The density of unmyelinated axons falls from 235,000/mm2 at birth to 60,000 to 80,000/mm2 at 3 to 10 years,306 and further to 22,000 to 34,000/mm2 by 15 to 20 years.297 Slightly different counts have been obtained by other workers; Jacobs and Love found a density of 30,000 to 40,000/mm2 by 10 years of age.207 The differences may be due to the difficulty of distinguishing small unmyelinated axons from small, circular Schwann cell processes. There is also considerable variation in density in different parts of the fascicle, rendering sampling procedures liable to produce errors. Unmyelinated axons differ considerably from myelinated ones in reaching their final maximum diameter as early as 23 weeks’ gestation.297,376 The number of axons in one Remak fiber varies between species, being up to 10 in rodents and more in cats but usually only 1 or 2 in adult human nerves (Fig. 3–22).214 In neonates there are numerous axons per Schwann cell unit. By about 6 months of age the numbers have reduced to two to six in each Schwann cell, with each axon having its own mesaxon; further maturation occurs later.207 In vitro studies have shown that the presence of fibroblasts is necessary for the maturation of Remak fibers, and the

FIGURE 3–22 Electron micrograph of normal human sural nerve showing numerous unmyelinated axons (A) embedded in denser Schwann cell processes. Two Schwann cell nuclei are sectioned (SN). There are occasional collagen pockets formed by Schwann cells surrounding bundles of collagen fibrils as if they were axons (arrows). Only rare axons share a Schwann cell process with another axon. Bar: 1 ␮m.

65

addition of individual basal laminal components separately causes extensive multiplication of Schwann cell processes.292 Human nerves are not fully mature until after the end of the first decade of life. The first detailed ultrastructural study of unmyelinated fibers in the human sural nerve was by Ochoa and Mair in 1969.297 They found no evidence of the syncytium postulated by Nageotte281 and still supported by Gasser in 1952, despite his electron microscope studies.147

GENERAL ORGANIZATION The components of the unmyelinated axon are the same as in axons with a myelin sheath. The proportions of NFs and MTs, however, differ somewhat. There tend to be fewer NFs and the axons are less densely populated. Conversely, the surrounding Schwann cell processes may have a greater density of intermediate filaments than the density of NFs in the axon. Although there is often a mesaxon as in myelinated fibers, this is not necessarily the case, and unmyelinated axons may merely lie in grooves in the Schwann cell (see Fig. 3–22). They are always covered by Schwann cell basal lamina even when not embedded in cytoplasm. Where adjacent Schwann cell processes touch, there are often dense patches forming structures suggesting junctional complexes.100 The axolemma in well-fixed material is often distinguishable from Schwann cell plasma membrane by its greater electron density and slight irregularity. The proportion of myelinated to unmyelinated fibers varies in different nerves. In the sural nerve in humans, the most commonly studied peripheral nerve, it is approximately 1:4.297 The spatial arrangement within the fascicle is not random but rather patchy. Remak fibers are also often associated with small myelinated axons. Although it has been suggested that this is developmental, it seems possible that these may be sympathetic bundles. The arrangement of unmyelinated axons in a normal Remak fiber is easily distinguishable from a bundle of unmyelinated sprouts originating from a damaged myelinated fiber during regeneration after axonal damage. In the latter situation the original basal laminal ensheathment of the degenerated parent myelinated fiber may be still present, but even if it is not, the group of fibers forms a round collection of axons and Schwann cell processes that still conforms to the original circular profile. The cytoplasm of regenerating Schwann cell processes is also usually less electron dense than that of stable Schwann cells. These collections of Schwann cell processes within a parent basal lamina are often referred to as bands of Büngner after the original description.63 Axonal sprouting is also suggested by the presence of more than three unmyelinated axons in one Schwann cell process.

66

Structure of the Peripheral Nervous System

Relationship of Unmyelinated Axons with Schwann Cells Because unmyelinated axons are not insulated from each other, there is the possibility that an impulse in one could affect a neighboring axon. Laborious studies showing that axons constantly switch between Schwann cells mean that unmyelinated axons are only in close apposition for a very short distance before they move off separately into different Schwann cells. This constant switching between cells means that “cross talk” between axons is very unlikely. Reconstruction of rat nerves from serial sections shows that the Schwann cells are between 200 and 500 ␮m in length and consist of several thin processes (Fig. 3–23).

Spatial Relationships between Axons, Schwann Cells, and Collagen Bundles of collagen fibrils may be enveloped by Schwann cells as if they were axons.297 These are known as “collagen pockets.” It has been shown that they are quite short in longitudinal extent, so they probably do not represent replacement of axons by collagen. However, they are definitely more common in pathologic situations. They might be “hooks” that attach Schwann cells to collagen bundles to provide additional mechanical stability. Although they were specifically looked for, they were not identified in a study of human fetal nerve,146 suggesting that they are a later development. When unmyelinated axons degenerate, the surrounding Schwann cell processes collapse to form a stack of flattened Schwann cell processes (Fig. 3–24). These are readily distinguishable from the rounded profiles that constitute the bands of Büngner formed by myelinated fiber degeneration.423 Sometimes, however, when there is extensive Schwann cell reduplication (e.g., in the hereditary motor and sensory neuropathies), the excess Schwann cell processes unassociated with axons may form similar flat sheets usually near or in classic onion bulbs. The presence of Reich granules in Schwann cells associated with unmyelinated fibers identifies the cells as having originally been associated with a myelinated axon. These axons may therefore be small unmyelinating sprouts arising from a regenerating myelinated fiber, but it is also possible that Schwann cells left without an axon may become associated with an unmyelinated axon rather than the myelinated one that was present originally.217

MORPHOMETRY In well-fixed material, unmyelinated axons may be measured and counted by electron microscopy. At least 10% of the fascicular area should be photographed to give

FIGURE 3–23 Three-dimensional reconstruction of interrelating Schwann cell units from a normal rat cervical sympathetic trunk based upon electron micrographs of serial transverse sections (D to U) cut at 5-␮m intervals. Four of the electron micrographs are reproduced. The arrangement of axons within Schwann cell units varies from level to level because of separation and rejoining of adjacent units. Schwann cell units are present at levels D, O, and U. (Courtesy of A. J. Aguayo and G. M. Bray.)

a good sample, but because of the uneven distribution within the fascicle, this should ideally be chosen by scattering measuring fields across the fascicle. However, in practice grid bars will obscure part of the fascicle, so a compromise is to montage the whole of several grid squares and measure all the axons within them. The average density

Microscopic Anatomy of the Peripheral Nervous System

67

subtle differences in cytoplasmic contents, with axons having more MTs than NFs and a clearer cytoplasmic background. Intermediate filaments in Schwann cell cytoplasm are often grouped, and this rarely happens in axons. With ideal fixation, the axolemma is denser than the Schwann cell plasma membrane. However, when fixation is less than ideal and in pathologic conditions, it may be impossible to distinguish axons from Schwann cell process with sufficient confidence to make morphometry worthwhile.

REGENERATION

FIGURE 3–24 Electron micrograph of flat sheets of Schwann cell processes (SC) formed after unmyelinated axon loss. The basal laminal ensheathment is indicated by an arrow. Bar: 0.2 ␮m.

found by Ochoa and Mair was about 29,000/mm2,298 but intrafascicular variability was estimated as being between 17,000 and 49,000/mm2.297 The size distribution is unimodal until the fifth decade of life,207 and in normal material the smallest fibers are about 0.5 ␮m. Axons with a smaller diameter than this become more common later in life and probably represent unmyelinated fiber regeneration. The highest density in this study was similar to that found by Ochoa and Mair298 in the younger subjects, being 36,000/mm2 at the age of 44 years, and the lowest was 17,300/mm2 in a 67-year-old man. Fibers without myelin and larger than 3 ␮m are almost certainly demyelinated fibers rather than unmyelinated ones. These are easily distinguished by the differences in the relationship of unmyelinated and demyelinated axons with their Schwann cells. As already mentioned, it may also be difficult to distinguish axons from circular Schwann cell profiles. The encircling of Schwann cell processes by other Schwann cell process was described by Ochoa and Vial.299 Axons usually have a lower electron density, but this difference may be obscured in very thin and/or poorly stained sections. If axons are completely encircled by Schwann cell cytoplasm, they will possess a mesaxon, but not all unmyelinated axons are completely indented into the Schwann cell (see Fig. 3–22). In addition, there are very

Schwann cells originally associated with unmyelinated axons but left without an axon after its degeneration still have the ability of adopting regenerative sprouts from myelinated fibers and can construct a myelin sheath on receiving the appropriate axonal signals. The converse can also apply; in regeneration, Schwann cells containing Reich granules, which are only normally found in association with myelinated fibers, may become associated with unmyelinated axons. Schwann cells that have changed from one type of fiber to another will then also change the repertoire of proteins that they express so that they take on the phenotype of the new type of Schwann cell.

The CNS-PNS Transitional Zone J. P. Fraher GENERAL FORM Almost all cranial and spinal nerves are formed from roots. Most roots are in turn formed from a number of rootlets that are attached to an area of the CNS surface (Fig. 3–25). The area through which motor rootlets emerge is the exit zone. The corresponding area for sensory rootlets is the attachment or entry zone. That for the spinal sensory rootlets is the dorsal root entry zone. Each rootlet possesses an individual CNS-PNS interface, at which the two tissue territories meet. It marks the junction of the myelination territories of oligodendrocytes and Schwann cells. The interface generally lies near the plane of the CNS surface and comprises the external surface of the glia limitans, the limiting membrane of the CNS, with its associated basal lamina. The glia limitans is specialized at the transitional zone (TZ), where it is thicker than elsewhere. Its astrocyte processes typically form a barrier across the rootlet that is pierced only by axons. The barrier is irregular, so that CNS and PNS tissues interdigitate over a length of a rootlet, the TZ (see Fig. 3–25).

68

Structure of the Peripheral Nervous System

FIGURE 3–25 A, Diagrammatic transverse section of spinal cord with ventral and dorsal roots, showing the attachment zones (AZ) of their rootlets. B and C, Enlargements of areas outlined in A showing the extents of both dorsal and ventral rootlet transitional zones (TZ) and a central tissue projection (CTP) in the dorsal rootlet TZ. D, Schematic diagram of ventral rootlet TZ indicating an individual myelinated fiber (arrowhead) and a bundle of unmyelinated axons (asterisk) traversing the thickened glia limitans (arrow) in individual tunnels.

The extent of overlap of CNS and PNS tissues at the TZ is often considerable, because the CNS tissue extends outward into the majority of rootlets as a central tissue projection (CTP) (Fig. 3–26). This most frequently takes the form of a tapering cone, embedded in a surrounding PNS tissue compartment (Fig. 3–27D). Its surface is a thick weave of astrocyte processes, while its core resembles white matter, though astrocyte and oligodendrocyte perikarya are absent from smaller examples.127 TZ morphology varies widely among individual nerves (see Fig. 3–27) and also among the constituent rootlets of a given nerve. In some cases the CTP is absent. The interface is then flush with the CNS surface (Fig. 3–27A) or even slightly below it, and the fiber bundle is generally surrounded by a tapering astrocytic collar at and below the plane of the CNS surface (see Fig. 3–25). The thickening of the astrocytic tissue of the glia limitans at

the TZ is often marked, sometimes to the extent that rootlets emerge through an extensive astrocytic pad raised above the plane of the surrounding CNS surface (Fig. 3–27B). Where the root runs close over the CNS surface, the glia limitans may also be thickened, forming a pad underlying the nerve.137 Not uncommonly, the segment of the rootlet nearest the cord consists entirely of CNS tissue (Fig. 3–27E). This arrangement reaches its most extreme expression in the cochlear nerve, which possesses a particularly long CTP. Consequently, in the rat the entire nerve consists of CNS tissue, and PNS tissue is present only in its branches (Fig. 3–27F). Many longer CTPs, such that of the facial nerve, branch distally, with individual branches extending into each division of the nerve (Fig. 3–27G). Unusually, in some rat dorsolateral vagal rootlets an invagination of PNS tissue extends centrally below the brainstem surface349 (Fig. 3–27H). These peripheral tissue insertions have irregular surfaces. As a result, some myelinated axons alternate between the two tissue compartments as they run through the TZ. These variations in form are not systematically related to the fiber composition or rootlet function.136,137 While all TZs tend to be of the same form in any spinal nerve root, the rootlets of many cranial nerves contain a variety of types of TZ, the pattern of which is constant.124,127,136,137,349 CTP size may be associated with fiber bundle size.52,135,388 All larger roots seem to contain one, and the bigger the root, the longer the CTP. Only the smallest rootlets (usually motor) consistently lack one. With regard to spinal roots, the CTP becomes longer in a caudal direction. As a rule, this is more extensive in sensory than in motor rootlets at corresponding levels.135,285,388,415,416 The CTP surface is generally irregular (see Fig. 3–27). Added to this, astrocytic processes project distally from the CTP into the rootlet endoneurium as a glial fringe, which greatly increases the surface area of the CNS-PNS interface.22,24,70 The fringe is much more prominent in the cat than in the rat. The cat possesses dorsal roots that are considerably larger than those of the rat. Its TZ is correspondingly more complex.22,70 On cross section it shows four concentric zones: outer (PNS), glial fringe, mantle, and core (Fig. 3–28). The rat spinal rootlet CTPs lack the well-defined mantle zone found on the cat CTP surface, and that possesses a regular layer of astrocyte cell bodies.

FIBER TRANSITION Myelinated axons traverse the CNS-PNS interface individually, being fully segregated from one another by the abundant astrocytic processes of the TZ as they do so (Fig. 3–29). At the level of the interface, each possesses a transitional node (Fig. 3–30A, B, D, and E).114,141,377 The myelin sheath

Microscopic Anatomy of the Peripheral Nervous System

69

FIGURE 3–26 Light micrograph of longitudinally sectioned rootlets; toluidine blue–stained cryostat sections (⫻40). A ⫽ dorsal rootlet; B ⫽ ventral rootlet; SC ⫽ spinal cord.

FIGURE 3–27 Diagrammatic sections longitudinal to rootlets and at right angles to the CNS surface, showing common types of TZ (A, B, and D–H) and glial islands (C) (see text).

distal to the node is formed by the transitional Schwann cell, and that central to it by the transitional oligodendrocyte (see Fig. 3–30D). The territories affected by demyelinating and other diseases involving these two cell types are correspondingly limited at the TZ. The central end of the Schwann cell commonly lies in a groove or invagination, walled by astrocytic tissue and lined by basal lamina22,70,122,128,131,135–137,260,276,285,347,349,401 (see Fig. 3–30A, D, and F). Invaginations are generally closely packed, so that the TZ astrocyte tissue resembles a honeycomb54,247 (see Fig. 3–30C). The transitional node is essentially a hybrid between central and peripheral types, but with additional unique features.129 For example, an astrocytic flange projects into the node gap, where it separates the oligodendrocyte and Schwann cell paranodes.129,131 The same is true of the Schwann cell and glial basal laminae; these also project into the node gap, where they become continuous with one another and form a further CNS-PNS barrier (see Fig. 3–30E). Because each myelinated fiber possesses a transitional node22–24,53,122,129,130,141,377 over the short TZ segment, node density is several times higher there than in the root or in the cord.118 The possibility of any cross talk taking place between nodes is decreased by their being offset. This is true both for those of ventral motoneurons and those of

70

Structure of the Peripheral Nervous System

FIGURE 3–28 Schematic representation of the transitional region (TR) at the cat nerve root–spinal cord junction. Different central-peripheral borderline arrangements. A, Concave borderline (dashed line) and inverted TR. B, Flat borderline situated at the level of the rootlet (r)–spinal cord junction. C and D, Convex dome-shaped borderline; the CNS expansion into the rootlet is moderate in C and extensive in D. Stippling denotes CNS tissue. The glial fringe is not shown. E, Pointed borderline. The extent of the transitional region (TR) is indicated. The cross-sectional appearance at four different TR levels (A, B, O, and C) and the distribution of the different TR zones are shown in the lower part of the illustration. Light area, endoneurial zone; shaded or cross-hatched areas, glial fringe; light stippling, mantle zone; heavy stippling, core zone. F, Root–spinal cord junction. The root (R) splits into rootlets (r), each with its own TR and attaching separately to the spinal cord (SC). G, Arrangement noted in several cranial nerve roots. The peripheral nervous system component of the root separates into a bundle of closely packed minirootlets, each equipped with a TR. The minirootlets reunite centrally. BS ⫽ brainstem. (From Berthold, C. H., Carlstedt T., and Corneliuson O.: Anatomy of the nerve root at the central-peripheral transition region. In Dyck, P. J., Thomas, P. K., Lambert, E. H., and Bunge, R. [eds.]: Peripheral Neuropathy, 2nd ed. Philadelphia, W. B. Saunders, p. 156, 1984, with permission.)

sensory dorsal rootlet fibers on the CTP surface. In addition, the latter are displaced distally into the rootlet, away from the CNS fibers of the dorsolateral cord. Also, the abundant concentric astrocytic processes that wall the invagination and surround them may insulate them from one another (see Figs. 3–29 and 3–30). Central to the TZ, many of the myelin sheaths become closely apposed to one another. Here also, in rat ventral motoneurons, internodes of one and the same fiber have a smaller axon caliber and thinner myelin sheaths in the CNS than in the PNS (Figs. 3–31 and 3–32).121,123 (The opposite is true of cat S1 dorsal root axons.25) The basal laminae at the TZ form a single compos-

ite continuous surface. That which overlies the glia limitans is projected distally as an array of tubes, each of which surrounds either a single myelinated axon or an unmyelinated axon bundle. Basal lamina thus forms an extensive barrier between the endoneurium and the axonal and glial elements at the TZ. Unmyelinated mammalian axons also traverse the TZ in tunnels. They generally do so as bundles rather than singly (see Fig. 3–25). However, some of the largest examples traverse the TZ glia limitans in individual tunnels, like myelinated axons.121,124 Central to the TZ, unmyelinated axons tend to be less segregated than is

Microscopic Anatomy of the Peripheral Nervous System

71

FIGURE 3–29 Transverse sections through the same fiber bundle at the ventral rootlet (A), TZ (B), and intraspinal (C) levels. B, The TZ is in the upper half of the illustration, transitional fibers being surrounded by pale astrocytic tissue. C, The fibers are sectioned mostly transversely, while those of the surrounding cord are sectioned longitudinally. Bars: 10 ␮m. (Based on Bristol, D. C., et al.: Spacing of central-peripheral transitional nodes of rat motoneurones. Acta Anat. [Basel] 148:206, 1993, with permission of S. Karger AG.)

the case peripherally.349 This is because, over their CNS courses, they lack the segregating processes provided by the Schwann cells peripherally. The mode of ensheathment of some small axons differs in the two milieux. Many that are myelinated peripherally are unmyelinated in the CNS. A few nerves lack a clearly defined TZ glial barrier. These are the olfactory93–95 and vomeronasal (J. P. Fraher, unpublished observations, 1981). In both, the tiny axons traverse the TZ as very large bundles. These are segregated only to a very limited degree, with just a few processes penetrating among them from the sleeve of olfactory ensheathing cells around the periphery of the bundle. These ensheathing cells accompany them through the TZ and into the olfactory bulb.86,126,329

Not uncommonly, isolated clumps of glial tissue are found lying freely in the rootlets, separated from the CNS interface (see Fig. 3–27C). These comprise glial islands.124,415,416 They may consist only of astrocytic tissue, but some also contain oligodendrocytes. They are generally traversed by axons. As these axons enter and leave them, their transitions between the endoneurium and the islands resemble those at the TZ proper.

TISSUE COMPOSITION In addition to axons, myelinating cells, and astrocytes, the TZ contains associated tissues in which they are embedded. These consist peripherally of endoneurium

72

Structure of the Peripheral Nervous System

Microscopic Anatomy of the Peripheral Nervous System

FIGURE 3–30 Myelinated fibers at the TZ. A, Diagram of longitudinally sectioned myelinated axon traversing the TZ. The central end of the Schwann cell is inserted into an invagination bounded by astrocyte processes (A). The CNS-PNS transitional node lies at the bottom of this. Here the basal laminae (BL) of Schwann cell and astrocyte are continuous with one another where they project into the node gap. (From Fraher, J. P.: The transitional zone and CNS regeneration. J. Anat. 196:137, 2000, with permission of Blackwells.) B, Longitudinally sectioned abducent nerve fiber showing transitional node (arrowhead) deep to the plane of the CNS surface (arrow). Thickened glia limitans is denoted by asterisk. (From Fraher, J. P., Smiddy, P. F., and O’Sullivan, V. R.: The central-peripheral transitional regions of cranial nerves: trochlear and abducent nerves. J. Anat. 161:115, 1988, with permission of Blackwells.) C, Scanning electron micrograph of experimentally avulsed cervical ventral rootlet showing the honeycomb-like arrangement of astrocytic tissue at the TZ. The spaces of the honeycomb are the empty invaginations left after avulsion of the Schwann cells. (From Bristol, D. C., and Fraher, J. P.: Experimental traction injuries of ventral spinal nerve roots: a scanning electron microscopic study. Neuropathol. Appl. Neurobiol. 15:549, 1989, with permission of Blackewells.) D, Longitudinally sectioned transitional myelinated fiber showing Schwann cell (top) and oligodendrocytic (bottom) paranodal regions. The invagination surrounding the former is bounded externally by astrocytic processes (asterisk) and is lined by basal lamina (arrows). (From Fraher, J. P., Smiddy, P. F., and O’Sullivan, V. R.: The central-peripheral transitional regions of cranial nerves: trochlear and abducent nerves. J. Anat. 161:115, 1988, with permission of Blackwells.) E, Longitudinal section through a transitional node, showing an astrocytic process (asterisk) extending into the node gap from an adjacent perikaryon. (From Fraher, J.: Node distribution and packing density in the rat CNS-PNS transitional zone. Microsc. Res. Tech. 34:507, 1996, with permission of Wiley-Liss.) F, Electron micrograph of a transversely sectioned transitional Schwann cell just distal to the paranode lying in an invagination bounded by TZ astrocyte processes and lined by basal lamina (arrows). G, Electron micrograph showing the thick, multilayered glia limitans at the TZ. (F and G from Fraher, J. P.: The CNS-PNS transitional zone of the rat: morphometric studies at cranial and spinal levels. Prog. Neurobiol. 38:261, 1992, with permission.) Bars: B and C, 10 ␮m; D and E, 1 ␮m; F and G, 0.5 ␮m.

FIGURE 3–31 Graphs showing myelin sheath thickness plotted against distance along entire peripheral (left) and central (right) internodes belonging to the same typical rat ventral motoneuron axons at 6 days (A) and at maturity (B). The midlevel of the transitional node is arrowed. Note the variation of myelin sheath thickness along the immature internodes. (Based on Fraher, J. P.: Quantitative studies on the maturation of central and peripheral parts of individual ventral motoneuron axons. II. Internodal length. J. Anat. 127:1, 1978.)

73

74

Structure of the Peripheral Nervous System

Microscopic Anatomy of the Peripheral Nervous System

and rootlet sheath,212,413 and centrally of astrocyte processes, which are continuous around the perimeter of the TZ with the glia limitans.129,130,259,260,349 TZ astrocytes have distinctive morphologic features. At least some of them produce a variety of types of processes serving a number of different roles. These include contributing to the TZ glia limitans and to the concentric sheaths walling the invagination around the central ends of the transitional Schwann cells, as well as extending into the TZ node gap (see earlier), where they sometimes contact the nodal axon. Those processes forming the bulk of the CTP and its surface are mainly fine, resembling those of the glia limitans generally41,43,182,260,316,386,409,437,445,464 (see Fig. 3–30G). It is clear, therefore, that individual astrocytes at the TZ give rise to a variety of processes appropriate to their location, and are not specialized as subclasses each of which produces only a specific type of process.

HISTORICAL BACKGROUND Studies on TZ structure (reviewed by Berthold et al.25) date back at least to the mid-19th century. The fundamental light microscopic picture of the TZ was described by Skinner388 and Tarlov.415,416 CTPs were noted from early on,290,443 and Skinner demonstrated the greater size of these in sensory than in motor roots.388 He also identified both astrocytes and oligodendrocytes within them. Tarlov415,416 identified glial islands, and showed that CTPs increase in length rostrocaudally along the neuraxis. Foncin,114 in the first ultrastructural TZ study, showed that the apparent myelin deficiency in the TZ, the Aufhellungszone,202,240,290,443 was a fixation artifact caused

75

by pronounced lamellar separation in the oligodendrocytic paranodes.

DEVELOPMENT The features of TZ development have been studied extensively in rat cervical ventral rootlets.121,124,125,293,294 Axons of developing ventral motoneurons converge on specific exit points in the presumptive glia limitans, possibly influenced by chemoattraction.173,251 They emerge from the neural tube as bundles (the future rootlets) and grow onward across the prospective leptomeninges, where they become associated with cells derived from the neural crest, some of which develop into Schwann cells. At the cord surface these axon bundles are eventually segregated by astrocytic processes, which grow into them from their margins and form a barrier across each (see Fig. 3–32). The segregating processes arise from perikarya located in the glia limitans around the perimeter of the bundle (see Fig. 3–32G). Some of these cells may have migrated outward from the ventricular zone along motoneuron axon bundles.124,125,293,294 Even before barrier formation, clear differences in cellular density (low centrally, high peripherally) commonly mark the presumptive interface. 127,130,135 Segregation begins around embryonic day 13, but is not completed until shortly after birth. Over this period the rootlet axons are extensively segregated in the PNS, a short distance beyond the plane of the cord surface. This is brought about by processes arising from cell clusters that appear on the rootlet surface shortly after axon emergence from the cord (see Fig. 3–32A and C to E).122,124,130,133,134 These form a fine interlocking matrix that segregates the axons and forms a barrier across the rootlet distal

FIGURE 3–32 A and B, Diagrams summarizing early rat ventral rootlet TZ development. A, An axon bundle emerges from the cord (bottom) to form a rootlet that is associated with a cell cluster. Arrows show plane of section of cluster in C. B, The cell cluster disperses in late fetal life, as the bundle begins to be segregated by processes growing in from perikarya surrounding it. Arrows show the plane of section of G. C–E, Electron micrographs of cell clusters. C and D, Transverse sections of a prominent cluster surrounding an axon bundle (C), and the matrix of processes that arises from it and subdivides the bundle (D). E, Longitudinal section of the bundle, showing ensheathing processes extending centrally from cluster cells below the plane of the cord surface, where they are apposed to astrocyte processes of the glia limitans. F–H, Electron micrographs of transversely sectioned axon bundles traversing the TZ, showing near-absence of glial segregating processes at embryonic day 18 (F), axon segregation by astrocyte processes in the immediate postnatal period (G), and subsequent myelination of a fully segregated axon lying in a developing invagination (H). Bars: C, 5 ␮m; D and E, 1 ␮m; F and H, 0.5 ␮m; G, 2 ␮m. (C, D, G, and H from Fraher, J. P.: Axon-glial relationships in early CNS-PNS transitional zone development: an ultrastructural study. J. Neurocytol. 26:41, 1997 with permission of Chapman & Hall.)

76

Structure of the Peripheral Nervous System

to the cord surface (see Fig. 3–32A, C, and D). This precedes formation of the TZ glial barrier and regresses as the former becomes more complete.125 The clusters themselves disperse just before birth. Their component cells serve as a source of Schwann cells.133,134 However, some of these fail to make axonal contact and lie dormant in the rootlet, at least for some time following birth.211 Subsequent TZ development varies between nerves, reflecting the variety of forms of the mature TZ.124 During development, the CNS-PNS interface oscillates and continually changes its form and position as the two tissue classes establish their mutually exclusive territories. (The formation of glial islands is likely to result from glial tissue becoming separated from the CNS during this process and thereafter coming to lie in the endoneurium.124,415,416) Because the astrocytic barrier is at first flush with the adjacent glia limitans, the interface is initially formed in the plane of the surrounding CNS surface. It retains this position, however, in only a minority of locations.124 It most often becomes displaced distally, as CNS tissue grows into the rootlet to form a CTP. Growth of the CTP continues even after transitional node formation. It therefore entails relative elongation of central fiber segments, because the transitional node remains at the interface. Concurrently, the node comes to lie progressively deeper in the invagination as a result of relative distal overgrowth of astrocytic processes around the central end of the transitional Schwann cell. Axons destined for myelination are the first to be fully segregated (see Fig. 3–32B and F to H). Each of these becomes associated with a premyelin Schwann cell and an oligodendrocyte, at some distance peripheral and central, respectively, to the glia limitans. They are at first separated by a bare axon segment. They subsequently extend along the axon toward one another until they reach the interface, where they become closely associated and where the transitional node develops between them.122,129 This entails formation of the Schwann cell and oligodendrocytic paranodes, which come about much as at typical PNS and CNS nodes, respectively. Astrocytic processes and associated basal laminae grow into the nodal gap to provide the basis of a barrier between the two territories. The myelination process is broadly typical for each cell type, though it is specialized in some details.121,143,358 For instance, there is a striking delay in myelination close to the root-cord junction, as compared with the intraspinal and more peripheral parts of the axons (see Fig. 3–31). Thus, during the first postnatal week, there are stretches lacking myelin along axons that are myelinated elsewhere.124 Also, the length of rootlet lying immediately distal to the TZ, the proximal rootlet segment, lags behind the rest of the root in its

development and even in the adult retains some immature features, such as the presence of unusually short internodes.124,130,133,134 Many of these transitional internodes are particularly short, and their myelin shows unusual features.70,134 In the cat, the Schwann cell internuclear distance along unmyelinated axons averages about 12 ␮m in the PNS compartment of the TZ, which should be compared with the 100- to 300-␮m distance found in peripheral nerves generally.68,69 Presumptively unmyelinated axons undergo variable degrees of segregation within the TZ. Some larger ones become segregated completely. Bundles of smaller axons remain much less so. The underlying factors influencing TZ development have been increasingly the subject of recent investigations, many in nonmammalian vertebrates.55,139,163–165,286,395 These studies have shown that the TZ loci of avian cranial nerve rootlet attachments may be determined in the presumptive glia limitans of the early neural tube.286 Cells derived from the neural crest form clusters termed boundary caps (BCs) associated with them. Similar BCs occur at rat presumptive dorsal rootlet TZs.164 They influence sensory axon ingrowth: Axon bundles are arrested near the plane of the developing cord. They may also prevent abnormal cell migration across the CNS-PNS interface, namely, outgrowth of ventral motoneurons into the rootlets.55,139,163–165,395 The cell clusters described earlier on rat ventral rootlets resemble BCs122,124,130,136,137 but are located some distance superficial to the cord surface. Also, they appear well after axon bundle outgrowth and so, unlike BCs, seem not to influence this. The TZ barrier retains its integrity from the outset. In principle, Schwann cells could invade the CNS without penetrating basal lamina. This is because their sheaths of basal lamina are continuous with that covering the glia limitans (see Fig. 3–30). However, this occurs rarely; they are only occasionally found in the undamaged CNS.2,326 Potential sources of these include the isolated cells derived from the cell clusters211 and the Schwann cells of autonomic nerves related to CNS blood vessels. Experimental studies suggest that the TZ astrocytic processes constitute the main barrier that prevents Schwann cell invasion of the CNS. This occurs following irradiation damage to the glia limitans.138,159–161,382–385 The barrier function is also exemplified in the Jimpy mouse, in which, despite severe central myelin deficiency, Schwann cells do not invade the CNS to remedy the defect.276 Neither do Schwann cell precursors invade the CNS before formation of the astrocytic TZ barrier. The cluster cell barrier may play a part in this.124,126 Thus between them the two barriers could prevent Schwann cell invasion during development and subsequently. Resistance could be aided by the expression of the nerve cell adhesion molecule HNK-1 by central axon

Microscopic Anatomy of the Peripheral Nervous System

segments.293 Their consequent close adherence could prevent axons from being pried apart by centrally directed Schwann cell processes.

COMPARATIVE ANATOMY The TZs that have been studied most extensively are those of the cat S1 dorsal root22–24,67,68 and of rat cranial and spinal nerves, in which TZ development has also been investigated in considerable detail.119,124,126 These and other investigations show that, in its general form, TZ structure is similar in a variety of mammals, including humans,285,387,388 rats, mice, cats, monkeys, and cattle.260,276,283,284,349,364,394,401 Although the range of nonmammalian vertebrates studied (including birds, amphibians, cartilaginous fishes, and cyclostomes) is necessarily incomplete, it seems likely that they all typically possess a glial barrier at the TZ. Data from some lower vertebrate species (Squalus acanthias, Tinca vulgaris, and Rana esculenta) indicate that, in these, motor and sensory rootlets do not differ in this respect.467 It is not yet known if, like the mammals, they lack a clearly defined glial barrier at the olfactory and vomeronasal nerve TZs. In the cyclostome Petromyzon (the sea lamprey), the large ventral motoneuron axons traverse the distinctive glia limitans in individual tunnels.120 In this respect they follow the general vertebrate pattern, though they are unmyelinated. They differ from this, however, in being associated at the point of emergence from the cord with an extensive labyrinth of paraxonal spaces. This is bounded by glial processes, Schwann cell processes, and basal lamina, and bears some comparison with the mammalian node gap in that it involves exposing a large surface area to the extracellular spaces, a feature that could play a part in electrolyte control (Fig. 3–33A). In those animals studied that are classified below the cyclostomes, a TZ barrier seems to be absent. For example, in the cephalochordate Amphioxus, neurites traverse the dorsal nerve attachments to the cord.119 While there is no glial barrier here, glial density is increased near the junction (Fig. 3–33B). Amphioxus has no ventral root. Instead, the motoneuron axons remain in the cord and make specialized contacts resembling motor end plates with extensions of the segmental and notochordal musculature, which are directly apposed to the cord surface. In these locations the glia limitans is absent113,144 (Fig. 3–33C and D). Invertebrate nervous systems vary considerably in form. Nevertheless, at least in those insects, copepods, and snails studied at the ultrastructural level, clearly defined glial interfaces seem not to occur where nerve fibers pass through junctions between nerve cords, interconnecting plexuses, and ganglia (J. P. Fraher, unpublished observations).

77

BLOOD SUPPLY The rat ventral rootlet TZ has a very rich blood supply, perhaps to meet the metabolic requirements of the high concentration of nodes there.213 The vessels are mainly capillaries or postcapillary venules.66,180,252,304,317,408,419 They form a densely knit network in the labyrinth of spaces within the pia mater and superficial to the glia limitans. Here they are closely packed around the rootlets, rather than within them. The latter are very small and their endoneurial spaces rarely if ever contain capillaries. Many vessels are continued centrally into the cord along the ventral motoneuron axon bundles, which they commonly follow across the ventral white column as far as the ventral horn gray matter.122,130,212 This arrangement differs from that at the cat S1 dorsal rootlet TZ. Here, the blood vessels do not accompany the axons between PNS and CNS. Instead they deviate from the endoneurial space and join vessels on the cord surface.22,70 Dorsal roots may as a consequence be more susceptible to ischemia than ventral roots.

MECHANICAL FEATURES In the subarachnoid space, each ventral rootlet is covered by a cellular and connective tissue rootlet sheath.212 The outer layers of this are continuous with the pia mater. This arrangement strengthens the rootlet and the junctional region generally. Central to this, the rootlets lie in the labyrinthine intrapial space. Here the rootlet-cord attachments are strengthened by the intricate interdigitation and adhesion that take place between sheath cells and TZ astrocyte processes.132 Longitudinally running rootlet endoneurial collagen fiber bundles may also strengthen the TZ by their becoming attached to the basal lamina. This also helps to anchor the rootlet to the CNS. The high node packing density at the TZ is likely to detract from its mechanical strength, because the node, being bare, is likely to be a weak point on the fiber.54 However, the transitional Schwann cells help to compensate for this by resisting traction through the collagenous connections between their basal laminae and that of the astrocytes bounding the invagination surrounding the central end of the Schwann cell. The invagination could also act as a suction cup, resisting distraction of the paranode. These mechanical features may protect the TZs against stress-related damage during relative movements between the neuraxis and its bony casing, once any slack has been taken up. The effectiveness of these mechanisms is shown by the fact that experimental rootlet traction tends to result in rupture at the level of the rootlet, rather than at the TZ.

FIGURE 3–33 A, Electron micrograph of a longitudinal section of part of a lamprey TZ, showing Schwann cell processes, astrocytic processes, and basal lamina (arrowheads) bounding a complex labyrinth of spaces (asterisks) in relation to the axon (top right) and the glia limitans (bottom). Schwann cell and astrocyte processes are not separated by basal lamina in the vicinity of the axon (bottom right). (From Fraher, J., and Cheong, E.: Glial-Schwann cell specialisations at the central-peripheral nervous system transition of a cyclostome: an ultrastructural study. Acta Anat. (Basel) 154:300, 1995, with permission of S Karger AG, Basel.) B and C, Transverse sections of Amphioxus spinal cord. B, The dorsal nerve extends to the left. At its attachment there is no glial barrier. Glial perikarya are most densely packed in the region of the tapering transitional segment (asterisk). Their density decreases progressively in a central-peripheral direction. (From Fraher, J. P.: The transitional zone and CNS regeneration. J. Anat. 196:137, 2000, with permission of Blackwells, Oxford, UK.) C, In the ventrolateral part of the cord, its surface is closely related to muscle elements (arrows): laterally to processes of myotomal muscles and ventrally to notochordal muscle-like processes. The glia limitans is deficient at both locations and the notochordal sheath is deficient at the second. D, Electron micrograph of the apposed cord surface and a muscle element, showing contacts resembling complex, wide neuromuscular junctions between the two. Here, terminals of propriospinal axons within the cord (arrows) are separated by a synaptic cleft from a muscle element (asterisk). Bars: A and D, 0.5 ␮m; B and C, 25 ␮m.

Microscopic Anatomy of the Peripheral Nervous System

REFERENCES 1. Abbott, N. J., Lieberman, E. M., Pichon, Y., et al.: Periaxonal K⫹ regulation in the small squid Alloteuthis: studies on isolated and in situ axons. Biophys. J. 53:275, 1988. 2. Adelman, L. S., and Aronson, S. M.: Intramedullary nerve fiber and Schwann cell proliferation within the spinal cord (schwannosis). Neurology 22:726, 1972. 3. Alvarez, J.: The autonomous axon: a model based on local synthesis of proteins. Biol. Res. 34:103, 2001. 4. Alvarez-Leefmans, F. J., Leon-Olea, M., Mendoza-Sotelo, J., et al.: Immunolocalization of the Na(⫹)-K(⫹)-2Cl(⫺) cotransporter in peripheral nervous tissue of vertebrates. Neuroscience 104:569, 2001. 5. Ariyasu, R. G., and Ellisman, M. H.: The distribution of (Na⫹ ⫹ K⫹)ATPase is continuous along the axolemma of unensheathed axons from spinal roots of ‘dystrophic’ mice. J. Neurocytol. 16:239, 1987. 6. Ariyasu, R. G., Nichol, J. A., and Ellisman, M. H.: Localization of sodium/potassium adenosine triphosphatase in multiple cell types of the murine nervous system with antibodies raised against the enzyme from kidney. J. Neurosci. 5:2581, 1985. 7. Armstrong, R., Toews, A. D., and Morell, P.: Axonal transport through nodes of Ranvier. Brain Res. 412:196, 1987. 8. Arroyo, E. J., Xu, Y. T., Zhou, L., et al.: Myelinating Schwann cells determine the internodal localization of Kv1.1, Kv1.2, Kvbeta2, and Caspr. J. Neurocytol. 28:333, 1999. 9. Asbury A. K.: The biology of Schwann cells. In Dyck, P. J., Thomas, P. K., and Lambert, E. H. (eds.): Peripheral Neuropathy. Philadelphia, W. B. Saunders, p. 201, 1975. 10. Baas, P. W.: Microtubule transport in the axon. Int. Rev. Cytol. 212:41, 2002. 11. Baas, P. W., and Black, M. M.: Individual microtubules in the axon consist of domains that differ in both composition and stability. J. Cell Biol. 111:495, 1990. 12. Baas, P. W., Slaughter, T., Brown, A., and Black, M. M.: Microtubule dynamics in axons and dendrites. J Neurosci. Res. 30:134, 1991. 13. Bacsich, P.: On the presence of Timofeew’s sensory corpuscles in the autonomic plexuses of the human prostate and seminal vesicles. J. Anat. 104:182, 1969. 14. Baker, M., Bostock, H., Grafe, P., and Martius, P.: Function and distribution of three types of rectifying channel in rat spinal root myelinated axons. J. Physiol. (Lond.) 383:45, 1987. 15. Beamish, N. G., Stolinski, C., Thomas, P. K., and King, R. H. M.: Freeze-fracture observations on normal and abnormal human perineurial tight junctions: alterations in diabetic polyneuropathy. Acta Neuropathol. (Berl.) 81:269, 1991. 16. Beggs, J., Johnson, P. C., Olafsen, A., et al.: Transperineurial arterioles in human sural nerve. J. Neuropathol. Exp. Neurol. 50:704, 1991. 17. Behse, F.: Morphometric studies on the human sural nerve. Acta Neurol. Scand. Suppl. 132:1, 1990. 18. Bennett, V., Lambert, S., Davis, J. Q., and Zhang, X.: Molecular architecture of the specialized axonal membrane at the node of Ranvier. Soc. Gen. Physiol. Ser. 52:107, 1997.

79

19. Berghs, S., Aggujaro, D., Dirkx, R. Jr., et al.: BetaIV spectrin, a new spectrin localized at axon initial segments and nodes of Ranvier in the central and peripheral nervous system. J. Cell Biol. 151:985, 2000. 20. Berthold, C. H.: Morphology of normal peripheral axons. In Waxman, S. G. (ed.): Physiology and Pathobiology of Axons. New York, Raven Press, p. 3, 1978. 21. Berthold, C. H.: Some aspects of the ultrastructural organization of peripheral myelinated axons in the cat. In Weiss, D. G. (ed.): Axoplasmic Transport. Berlin, Springer-Verlag, p. 40, 1982. 22. Berthold, C. H., and Carlstedt, T.: Observations on the morphology at the transition between the peripheral and the central nervous system in the cat. II. General organization of the transitional region in S1 dorsal rootlets. Acta Physiol. Scand. Suppl. 446:23, 1977. 23. Berthold, C. H., and Carlstedt, T.: Observations on the morphology at the transition between the peripheral and the central nervous system in the cat. III. Myelinated fibres in S1 dorsal rootlets. Acta Physiol. Scand. Suppl. 446:43, 1977. 24. Berthold, C. H., and Carlstedt, T.: Observations on the morphology at the transition between the peripheral and the central nervous system in the cat. V. A light microscopical and histochemical study of S1 dorsal rootlets in developing kittens. Acta Physiol. Scand. Suppl. 446:73, 1977. 25. Berthold, C. H., Carlstedt, T., and Corneliuson, O.: Anatomy of the nerve root at the central-peripheral transition region. In Dyck, P. J., Thomas, P. K., Lambert, E. H., and Bunge, R. (eds.): Peripheral Neuropathy, 2nd ed. Philadelphia, W. B. Saunders, p. 156, 1984. 26. Berthold, C. H., Corneliuson, O., and Mellstrom, A.: Peroxidase activity at nodes of Ranvier in lumbosacral ventral spinal roots and in the PNS-CNS transitional region after intramuscular administration of horseradish peroxidase. J. Neurocytol. 15:253, 1986. 27. Berthold, C. H., Corneliuson, O., and Rydmark, M.: Changes in shape and size of cat spinal root myelinated nerve fibers during fixation and Vestopal-w embedding for electron microscopy. J. Ultrastruct. Res. 80:23, 1982. 28. Berthold, C. H., Fabricius, C., Rydmark, M., and Andersen, B.: Axoplasmic organelles at nodes of Ranvier. I. Occurrence and distribution in large myelinated spinal root axons of the adult cat. J. Neurocytol. 22:925, 1993. 29. Berthold, C. H., and Mellstrom, A.: Peroxidase activity at consecutive nodes of Ranvier in the nerve to the medial gastrocnemius muscle after intramuscular administration of horseradish peroxidase. Neuroscience 19:1349, 1986. 30. Berthold, C. H., Nilsson, I., and Rydmark, M.: Axon diameter and myelin sheath thickness in nerve fibres of the ventral spinal root of the seventh lumbar nerve of the adult and developing cat. J. Anat. 136:483, 1983. 31. Berthold, C. H., Nordborg, C., Hildebrand, C., et al.: Sural nerve biopsies from workers with a history of chronic exposure to organic solvents and from normal control cases: morphometric and ultrastructural studies. Acta Neuropathol. (Berl.) 62:73, 1983. 32. Berthold, C. H., and Rydmark, M.: Anatomy of the paranode-node-paranode region in the cat. Experientia 39:964, 1983.

80

Structure of the Peripheral Nervous System

33. Berthold, C. H., and Rydmark, M.: Electron microscopic serial section analysis of nodes of Ranvier in lumbosacral spinal roots of the cat: ultrastructural organization of nodal compartments in fibres of different sizes. J. Neurocytol. 12:475, 1983. 34. Berthold, C. H., and Rydmark, M.: Morphology of normal peripheral axons. In Waxman, S. G., Kocsis, J. D., and Stys, P. K. (eds.): The Axon: Structure, Function and Pathophysiology. Oxford, UK, Oxford University Press, p. 13, 1995. 35. Bischoff, A., and Thomas, P. K.: Microscopic anatomy of myelinated nerve fibers. In Dyck, P. J., Thomas, P. K., and Lambert, E. H. (eds.): Peripheral Neuropathy. Philadelphia, W. B. Saunders, p. 104, 1975. 36. Black, J. A., Friedman, B., Waxman, S. G., et al.: Immunoultrastructural localization of sodium channels at nodes of Ranvier and perinodal astrocytes in rat optic nerve. Proc. R. Soc. Lond. B Biol. Sci. 238:39, 1989. 37. Black, J. A., Kocsis, J. D., and Waxman, S. G.: Ion channel organization of the myelinated fiber. Trends Neurosci. 13:48, 1990. 38. Black, M. M., Chestnut, M. H., Pleasure, I. T., and Keen, J. H.: Stable clathrin: uncoating protein (hsc70) complexes in intact neurons and their axonal transport. J. Neurosci. 11:1163, 1991. 39. Black, M. M., and Lasek, R. J.: Slow components of axonal transport: two cytoskeletal networks. J. Cell Biol. 86:616, 1980. 40. Blanchard, C. E., Mackenzie, M. L., Sikri, K., and Allt, G.: Filipin-sterol complexes at nodes of Ranvier. J. Neurocytol. 14:1053, 1985. 41. Bondareff, W., and McLone, D. G.: The external glial limiting membrane in Macaca: ultrastructure of a laminated glioepithelium. Am. J. Anat. 136:277, 1973. 42. Boyle, M. E., Berglund, E. O., Murai, K. K., et al.: Contactin orchestrates assembly of the septate-like junctions at the paranode in myelinated peripheral nerve. Neuron 30:385, 2001. 43. Braak, E.: On the fine structure of the external glial layer in the isocortex of man. Cell Tissue Res. 157:367, 1975. 44. Brady, S. T., Lasek, R. J., and Allen, R. D.: Video microscopy of fast axonal transport in extruded axoplasm: a new model for study of molecular mechanisms. Cell Motil. 5:81, 1985. 45. Bray, D., and Bunge, M. B.: Serial analysis of microtubules in cultured rat sensory axons. J. Neurocytol. 10:589, 1981. 46. Bray, G. M., Rasminsky, M., and Aguayo, A. J.: Interactions between axons and their sheath cells. Annu. Rev. Neurosci. 4:127, 1981. 47. Bray, J. J., Fernyhough, P., Bamburg, J. R., and Bray, D.: Actin depolymerizing factor is a component of slow axonal transport. J. Neurochem. 58:2081, 1992. 48. Brismar, T.: Potential clamp experiments on myelinated nerve fibres from alloxan diabetic rats. Acta Physiol. Scand. 105:384, 1979. 49. Brismar, T.: Potential clamp analysis of membrane currents in rat myelinated nerve fibres. J. Physiol. (Lond.) 298:171, 1980. 50. Brismar, T., Hildebrand, C., and Berglund, S.: Nodes of Ranvier above a neuroma in the rat sciatic nerve: voltage

51.

52.

53.

54.

55.

56. 57.

58.

59. 60.

61. 62.

63.

64.

65. 66.

67.

68.

clamp analysis and electron microscopy. Brain Res. 378:347, 1986. Brismar, T., Hildebrand, C., and Tegner, R.: Nodes of Ranvier in acrylamide neuropathy: voltage clamp and electron microscopic analysis of rat sciatic nerve fibres at proximal levels. Brain Res. 423:135, 1987. Bristol, D. C.: Aspects of the ventral rootlet transitional zone in the rat: a morphometric and morphological analysis of the L, ventral rootlet transitional zone and the distribution of transitional nodes of Ranvier within it. Ph.D. Thesis, Dublin, National University of Ireland, 1989. Bristol, D. C., and Fraher, J. P.: A morphometric study of the CNS-PNS transitional zone in rat lumbar ventral roots. J. Anat. 152:236, 1987. Bristol, D. C., and Fraher, J. P.: Experimental traction injuries of ventral spinal nerve roots: a scanning electron microscopic study. Neuropathol. Appl. Neurobiol. 15:549, 1989. Britsch, S., Goerich, D. E., Riethmacher, D., et al.: The transcription factor Sox10 is a key regulator of peripheral glial development. Genes Dev. 15:66, 2001. Brophy, P. J.: Axoglial junctions: separate the channels or scramble the message. Curr. Biol. 11:R555, 2001. Brown, A. A., Xu, T., Arroyo, E. J., et al.: Molecular organization of the nodal region is not altered in spontaneously diabetic BB-Wistar rats. J. Neurosci. Res. 65:139, 2001. Brown, E. R., and Abbott, N. J.: Ultrastructure and permeability of the Schwann cell layer surrounding the giant axon of the squid. J. Neurocytol. 22:283, 1993. Bruns, R. R.: Beaded filaments and long-spacing fibrils: relation to type VI collagen. J. Ultrastruct. Res. 89:136, 1984. Bruns, R. R., Press, W., Engvall, E., et al.: Type VI collagen in extracellular, 100-nm periodic filaments and fibrils: identification by immunoelectron microscopy. J. Cell Biol. 103:393, 1986. Bunge, M. B.: The axonal cytoskeleton: its role in generating and maintaining cell forms. Trends Neurosci. 9:477, 1986. Bunge, M. B., Wood, P. M., Tynan, L. B., et al.: Perineurium originates from fibroblasts: demonstration in vitro with a retroviral marker. Science 243:229, 1989. Bungner, O. v.: Versuche zum Studium der histologischen Degeneration und Regeneration verletzer Nerven, ihre Anordnung und ihre Ergebnisse. Zieglers Beitr. 10:321, 1891. Burgoyne, R. D.: Cytoskeleton is a major neuronal organelle. In Burgoyne, R. D. (ed.): The Neuronal Cytoskeleton. New York, Wiley-Liss, p. 1, 1991. Burton, P. R.: Microtubules of frog olfactory axons: their length and number/axon. Brain Res. 409:71, 1987. Caley, D. W., and Maxwell, D. S.: Development of the blood vessels and extracellular spaces during postnatal maturation of rat cerebral cortex. J. Comp. Neurol. 138:31, 1970. Carlstedt, T.: Observations on the morphology at the transition between the peripheral and the central nervous system in the cat. I. A preparative procedure useful for electron microscopy of the lumbosacral dorsal. Acta Physiol. Scand. Suppl. 446:5, 1977. Carlstedt, T.: Observations on the morphology at the transition between the peripheral and the central nervous system in the cat. IV. Unmyelinated fibres in S1 dorsal rootlets. Acta Physiol. Scand. Suppl. 446:61, 1977.

Microscopic Anatomy of the Peripheral Nervous System 69. Carlstedt, T.: Internodal length of nerve fibres in dorsal roots of cat spinal cord. Neurosci. Lett. 19:251, 1980. 70. Carlstedt, T.: An electron-microscopical study of the developing transitional region in feline S1 dorsal rootlets. J. Neurol. Sci. 50:357, 1981. 71. Causey, G.: The Cell of Schwann. Baltimore, Williams & Wilkins, 1960. 72. Chiu, S. Y.: Changes in excitable membrane properties in Schwann cells of adult rabbit sciatic nerves following nerve transection. J. Physiol. (Lond.) 396:173, 1988. 73. Chiu, S. Y., and Ritchie, J. M.: Evidence for the presence of potassium channels in the paranodal region of acutely demyelinated mammalian single nerve fibres. J. Physiol. (Lond.) 313:415, 1981. 74. Chiu, S. Y., Schrager, P., and Ritchie, J. M.: Neuronal-type Na⫹ and K⫹ channels in rabbit cultured Schwann cells. Nature 311:156, 1984. 75. Chiu, S. Y., and Schwarz, W.: Sodium and potassium currents in acutely demyelinated internodes of rabbit sciatic nerves. J. Physiol. (Lond.) 391:631, 1987. 76. Clarke, E., and Bearn, J. G.: The spiral nerve bands of Fontana. Brain 95:1, 1972. 77. Clermont, Y.: Contractile elements in the limiting membranes of the seminiferous tubules. Exp. Cell Res. 15:438, 1958. 78. Cui, S. Y., Christensen, E. I., and Nielsen, S.: Membrane traffic after inhibition of endocytosis in renal proximal tubules. J. Struct. Biol. 107:201, 1991. 79. Dahlstrom, A. B., Czernik, A. J., and Li, J. Y.: Organelles in fast axonal transport: what molecules do they carry in anterograde vs retrograde directions, as observed in mammalian systems? Mol. Neurobiol. 6:157, 1992. 80. Davis, J. Q., Lambert, S., and Bennett, V.: Molecular composition of the node of Ranvier: identification of ankyrin-binding cell adhesion molecules neurofascin (mucin ⫹ /third FNIII domain–) and NrCAM at nodal axon segments. J. Cell Biol. 135:1355, 1996. 81. de la Motte, D. J., Hall, S. M., and Allt, G.: A study of the perineurium in peripheral nerve pathology. Acta Neuropathol. 33:257, 1975. 82. De Waegh, S., and Brady, S. T.: Altered slow axonal transport and regeneration in a myelin-deficient mutant mouse: the trembler as an in vivo model for Schwann cell-axon interactions. J. Neurosci. 10:1855, 1990. 83. De Waegh, S. M., and Brady, S. T.: Local control of axonal properties by Schwann cells: neurofilaments and axonal transport in homologous and heterologous nerve grafts. J. Neurosci. Res. 30:201, 1991. 84. De Waegh, S. M., Lee, V. M., and Brady, S. T.: Local modulation of neurofilament phosphorylation, axonal caliber, and slow axonal transport by myelinating Schwann cells. Cell 68:451, 1992. 85. DeCoster, M. A., and DeVries, G. H.: Evidence that the axolemmal mitogen for cultured Schwann cells is a positively charged, heparan sulfate proteoglycan-bound, heparin-displaceable molecule. J. Neurosci. Res. 22:283, 1989. 86. Devon, R., and Doucette, R.: Olfactory ensheathing cells do not require l-ascorbic acid in vitro to assemble a basal lamina or to myelinate dorsal root ganglion neurites. Brain Res. 688:223, 1995.

81

87. Dingemans, K. P., and Teeling, P.: Long-spacing collagen and proteoglycans in pathologic tissues. Ultrastruct. Pathol. 18:539, 1994. 88. Dolapchieva, S., Ichev, K., and Ovtscharoff, W.: Bismuth iodide impregnation of the peripheral nerve fibres. Z. Mikrosk. Anat. Forsch. 99:804, 1985. 89. Dolapchieva, S., Ichev, K., and Ovtscharoff, W.: Lectin binding-sites in axon-myelin-Schwann cell complex. Acta Histochem. Cytochem. 19:253, 1986. 90. Dolapchieva, S., Ichev, K., and Ovtscharoff, W.: Ultrastructural cytochemical localization of 5⬘-nucleotidase activity in axonmyelin-Schwann cell complex. Acta Histochem. 83:125, 1988. 91. Dolapchieva, S., Ichev, K., and Ovtscharoff, W.: Cytochemical localization of ATPase in axon-myelin-Schwann cell complex type. Z. Mikrosk. Anat. Forsch. 103:151, 1989. 92. Dolapchieva, S., Ovtscharoff, W., and Ichev, K.: Localizations of ruthenium red positive material in rabbit peripheral nerves. Acta Histochem. 78:19, 1986. 93. Doucette, R.: Glial influences on axonal growth in the primary olfactory system. Glia 3:433, 1990. 94. Doucette, R.: PNS-CNS transitional zone of the first cranial nerve. J. Comp. Neurol. 312:451, 1991. 95. Doucette, R.: Glial cells in the nerve fiber layer of the main olfactory bulb of embryonic and adult mammals. Microsc. Res. Tech. 24:113, 1993. 96. Dubois, J. M.: Capsaicin blocks one class of K⫹ channels in the frog node of Ranvier. Brain Res. 245:372, 1982. 97. Duce, I. R., and Keen, P.: Can neuronal smooth endoplasmic reticulum function as a calcium reservoir? Neuroscience 3:837, 1978. 98. Dupree, J. L., Girault, J. A., and Popko, B.: Axo-glial interactions regulate the localization of axonal paranodal proteins. J. Cell Biol. 147:1145, 1999. 99. Dyck, P. J., Giannini, C., and Lais, A.: Pathologic alterations in nerves. In Dyck, P. J., Thomas, P. K., Griffin, J. W., et al. (eds.): Peripheral Neuropathy, 3rd ed. Philadelphia, W. B. Saunders, p. 514, 1993. 100. Eames, R. A., and Gamble, H. J.: Schwann cell relationships in normal human cutaneous nerves. J. Anat. 106:417, 1970. 101. Ellisman, M. H.: Beyond neurofilaments and microtubules. Neurosci. Res. Program Bull. 19:43, 1981. 102. Ellisman, M. H., Friedman, P. L., and Hamilton, W. J.: The localization of sodium and calcium to Schwann cell paranodal loops at nodes of Ranvier and of calcium to compact myelin. J. Neurocytol. 9:185, 1980. 103. Ellisman, M. H., and Levinson, S. R.: Immunocytochemical localization of sodium channel distributions in the excitable membranes of Electrophorus electricus. Proc. Natl. Acad. Sci. U. S. A. 79:6707, 1982. 104. Ellisman, M. H., and Lindsey, J. D.: The axoplasmic reticulum within myelinated axons is not transported rapidly. J. Neurocytol. 12:393, 1983. 105. Ellisman, M. H., and Porter, K. R.: Microtrabecular structure of the axoplasmic matrix: visualization of cross-linking structures and their distribution. J. Cell Biol. 87:464, 1980. 106. Ellisman, M. H., Wiley, C. A., Lindsey, J. D., et al.: Structure and function of the cytoskeleton and endomembrane systems at the node of Ranvier. In Zagoren, J. C., and Fedoroff, S. (eds.): The Node of Ranvier. New York, Academic Press, p. 153, 1984.

82

Structure of the Peripheral Nervous System

107. Espejo, F., and Alvarez, J.: Microtubules and calibers in normal and regenerating axons of the sural nerve of the rat. J. Comp. Neurol. 250:65, 1986. 108. Evans, P. D., Reale, V., Merzon, R. M., and Villegas, J.: The effect of a glutamate uptake inhibitor on axon-Schwann cell signalling in the squid giant nerve fibre. J. Exp. Biol. 173:251, 1992. 109. Evans, P. D., Reale, V., Merzon, R. M., and Villegas, J.: N-methyl-D-aspartate (NMDA) and non-NMDA (metabotropic) type glutamate receptors modulate the membrane potential of the Schwann cell of the squid giant nerve fibre. J. Exp. Biol. 173:229, 1992. 110. Fabricius, C., Berthold, C. H., and Rydmark, M.: Axoplasmic organelles at nodes of Ranvier. II. Occurrence and distribution in large myelinated spinal cord axons of the adult cat. J. Neurocytol. 22:941, 1993. 111. Fannon, A. M., Sherman, D. L., Ilyina-Gragerova, G., et al.: Novel E-cadherin-mediated adhesion in peripheral nerve: Schwann cell architecture is stabilized by autotypic adherens junctions. J. Cell Biol. 129:189, 1995. 112. Fath, K. R., and Lasek, R. J.: Two classes of actin microfilaments are associated with the inner cytoskeleton of axons. J. Cell Biol. 107:613, 1988. 113. Flood, P. R.: A peculiar mode of muscular innervation in Amphioxus: light and electron microscopic studies of the so-called ventral roots. J. Comp. Neurol. 126:181, 1966. 114. Foncin, J. F.: Structure fine de la zone de passage radiculomedullaire. Rev. Neurol. (Paris) 105:509, 1961. 115. Fontana, F.: Traité sur le Vénin de la Vipere sur les Poisons Americains. Florence, 1781. 116. Forman, D. S.: Axonal transport of mitochondria. In Smith, R. S., and Bisby, M. A. (eds.): Axonal Transport. New York, Alan R Liss, p. 155, 1987. 117. Forman, D. S., Lynch, K. J., and Smith, R. S.: Organelle dynamics in lobster axons: anterograde, retrograde and stationary mitochondria. Brain Res. 412:96, 1987. 118. Fraher, J.: Node distribution and packing density in the rat CNS-PNS transitional zone. Microsc. Res. Tech. 34:507, 1996. 119. Fraher, J.: Axons and glial interfaces: ultrastructural studies. J. Anat. 200:415, 2002. 120. Fraher, J., and Cheong, E.: Glial-Schwann cell specialisations at the central-peripheral nervous system transition of a cyclostome: an ultrastructural study. Acta Anat. (Basel) 154:300, 1995. 121. Fraher, J. P.: The growth and myelination of central and peripheral segments of ventral motoneurone axons: a quantitative ultrastructural study. Brain Res. 105:193, 1976. 122. Fraher, J. P.: The maturation of the ventral root-spinal cord transitional zone: an ultrastructural study. J. Neurol. Sci. 36:427, 1978. 123. Fraher, J. P.: Quantitative studies on the maturation of central and peripheral parts of individual ventral motoneuron axons. I. Myelin sheath and axon calibre. J. Anat. 126:509, 1978. 124. Fraher, J. P.: The CNS-PNS transitional zone of the rat: morphometric studies at cranial and spinal levels. Prog. Neurobiol. 38:261, 1992. 125. Fraher, J. P.: Axon-glial relationships in early CNS-PNS transitional zone development: an ultrastructural study. J. Neurocytol. 26:41, 1997.

126. Fraher, J. P.: The transitional zone and CNS regeneration. J. Anat. 196:137, 2000. 127. Fraher, J. P., and Delanty, F. J.: The development of the central-peripheral transitional zone of the rat cochlear nerve: a light microscopic study. J. Anat. 155:109, 1987. 128. Fraher, J. P., and Kaar, G. F.: The maturation of the ventral root-spinal cord transitional zone. Part 2. A quantitative ultrastructural study of the dynamics of its early development. J. Neurol. Sci. 53:63, 1982. 129. Fraher, J. P., and Kaar, G. F.: The transitional node of Ranvier at the junction of the central and peripheral nervous systems: an ultrastructural study of its development and mature form. J. Anat. 139:215, 1984. 130. Fraher, J. P., and Kaar, G. F.: The lumbar ventral root-spinal cord transitional zone in the rat: a morphological study during development and at maturity. J. Anat. 145:109, 1986. 131. Fraher, J. P., Kaar, G. F., Bristol, D. C., and Rossiter, J. P.: Development of ventral spinal motoneurone fibres: a correlative study of the growth and maturation of central and peripheral segments of large and small fibre classes. Prog. Neurobiol. 31:199, 1988. 132. Fraher, J. P., and O’Leary, D.: Morphological specialisations of rat cranial nerve transitional zones. J. Anat. 184:119, 1994. 133. Fraher, J. P., and Rossiter, J. P.: Cell clusters on fetal rat ventral roots: prenatal development. J. Anat. 136:111, 1983. 134. Fraher, J. P., and Rossiter, J. P.: Cell clusters on rat ventral roots: postnatal development. J. Anat. 137:555, 1983. 135. Fraher, J. P., and Sheehan, M. M.: The CNS-PNS transitional zone of rat cervical dorsal roots during development and at maturity: a morphological and morphometric study. J. Anat. 152:189, 1987. 136. Fraher, J. P., Smiddy, P. F., and O’Sullivan, V. R.: The centralperipheral transitional regions of cranial nerves: oculomotor nerve. J. Anat. 161:103, 1988. 137. Fraher, J. P., Smiddy, P. F., and O’Sullivan, V. R.: The centralperipheral transitional regions of cranial nerves: trochlear and abducent nerves. J. Anat. 161:115, 1988. 138. Franklin, R. J., and Blakemore, W. F.: Requirements for Schwann cell migration within CNS environments: a viewpoint. Int. J. Dev. Neurosci. 11:641, 1993. 139. Franz, T.: Defective ensheathment of motoric nerves in the Splotch mutant mouse. Acta Anat. (Basel) 138:246, 1990. 140. Freedman, S. D., and Lentz, T. L.: Binding of horseradish peroxidase-alpha-bungarotoxin to axonal membranes at the node of Ranvier. J. Comp. Neurol. 193:179, 1980. 141. Friede, R. L.: A Histochemical Atlas of Tissue Oxidation in the Brain Stem of the Cat. Basel, S. Karger, p. 18, 1961. 142. Friede, R. L., and Bischhausen, R.: The organization of endoneural collagen in peripheral nerves as revealed with the scanning electron microscope. J. Neurol. Sci. 38:83, 1978. 143. Friede, R. L., and Samorajski, T.: Myelin formation in the sciatic nerve of the rat—a quantitative electron microscopic, histochemical and radioautographic study. J. Neuropathol. Exp. Neurol. 27:546, 1968. 144. Fritzsch, B., and Northcutt, R. G.: Cranial and spinal nerve organization in Amphioxus and lampreys: evidence for an ancestral craniate pattern. Acta Anat. (Basel) 148:96, 1993.

Microscopic Anatomy of the Peripheral Nervous System 145. Gabriel, G., Thomas, P. K., King, R. H. M., et al.: Freezefracture observations on human peripheral nerve. J. Anat. 146:153, 1986. 146. Gamble, H. J., and Breathnach, A. S.: An electron-microscope study of human foetal peripheral nerve. J. Anat. 99:573, 1965. 147. Gasser, H. S.: Discussion of the hypothesis of saltatory conduction. Cold Spring Harb. Symp. Quant. Biol. 17:32, 1952. 148. Gatzinsky, K. P.: Node-paranode regions as local degradative centres in alpha-motor axons. Microsc. Res. Tech. 34:492, 1996. 149. Gatzinsky, K. P., and Berthold, C. H.: Lysosomal activity at nodes of Ranvier during retrograde axonal transport of horseradish peroxidase in alpha-motor neurons of the cat. J. Neurocytol. 19:989, 1990. 150. Gatzinsky, K. P., Berthold, C. H., and Corneliuson, O.: Acid phosphatase activity at nodes of Ranvier in alpha-motor and dorsal root ganglion neurons of the cat. J. Neurocytol. 17:531, 1988. 151. Gatzinsky, K. P., Berthold, C. H., and Rydmark, M.: Axon-Schwann cell networks are regular components of nodal regions in normal large nerve fibres of cat spinal nerves. Neurosci. Lett. 124:264, 1991. 152. Gatzinsky, K. P., Holtmann, B. B., Daraie, B., et al.: Early onset of degenerative changes at nodes of Ranvier in alpha-motor axons of CNTF null (-/-) mutant mice. Glia 42:340, 2003. 153. Gatzinsky, K. P., Persson, G. H., and Berthold, C. H.: Removal of retrogradely transported material from rat lumbosacral alpha-motor axons by paranodal axon-Schwann cell networks. Glia 20:115, 1997. 154. Gerhardt, E.: Are Ranvier’s nodes artifacts by preparation? Acta Anat. 141:132, 1991. 155. Ghabriel, M. N., and Allt, G.: Incisures of SchmidtLanterman. Prog. Neurobiol. 17:25, 1981. 156. Ghabriel, M. N., and Allt, G.: The node of Ranvier. Prog. Anat. 2:137, 1982. 157. Ghadially, F. N.: Ultrastructural Pathology of the Cell and Matrix. London, Butterworths, p. 1252, 1988. 158. Giannini, C., and Dyck, P. J.: Axoglial dysjunction: a critical appraisal of definition, techniques, and previous results. Microsc. Res. Tech. 34:436, 1996. 159. Gilmore, S. A., and Duncan, D.: On the presence of peripheral-like nervous and connective tissue within irradiated spinal cord. Anat. Rec. 160:675, 1968. 160. Gilmore, S. A., and Sims, T. J.: The role of Schwann cells in the repair of glial cell deficits in the spinal cord. In Das, G. D., and Wallace, R. B. (eds.): Neural Transplantation and Regeneration. New York, Springer-Verlag, p. 245, 1986. 161. Gilmore, S. A., and Sims, T. J.: Glial-glial and glial-neuronal interfaces in radiation-induced, glia-depleted spinal cord. J. Anat. 190:5, 1997. 162. Ginsberg, L., Malik, O., Kenton, A. R., et al.: Co-existence of hereditary and inflammatory neuropathy. Brain 127:1, 2003. 163. Golding, J., Shewan, D., and Cohen, J.: Maturation of the mammalian dorsal root entry zone—from entry to no entry. Trends Neurosci. 20:303, 1997. 164. Golding, J. P., and Cohen, J.: Border controls at the mammalian spinal cord: late-surviving neural crest boundary

165.

166.

167.

168. 169. 170.

171.

172.

173.

174.

175.

176.

177.

178.

179.

180.

181.

182.

83

cap cells at dorsal root entry sites may regulate sensory afferent ingrowth and entry zone morphogenesis. Mol. Cell. Neurosci. 9:381, 1997. Golding, J. P., Shewan, D., Berry, M., and Cohen, J.: An in vitro model of the rat dorsal root entry zone reveals developmental changes in the extent of sensory axon growth into the spinal cord. Mol. Cell. Neurosci. 7:191, 1996. Gorio, A.: Sprouting and regeneration of peripheral nerve. In Zagoren, J. C., and Fedoroff, S. (eds.): The Node of Ranvier. New York, Academic Press, p. 353, 1984. Grafstein, B.: Axoplasmic transport: function and mechanisms. In Waxman, S. G., Kocsis, J. D., and Stys, P. K. (eds.): The Axon: Structure, Function and Pathophysiology. Oxford, UK, Oxford University Press, p. 185, 1995. Grafstein, B., and Forman, D. S.: Intracellular transport in neurons. Physiol. Rev. 60:1167, 1980. Gray, P. T. A., and Ritchie, J. M.: Ion channels in Schwann and glial-cells. Trends Neurosci. 8:411, 1985. Griffin, J. W., George, R., and Ho, T.: Macrophage systems in peripheral nerves: a review. J. Neuropathol. Exp. Neurol. 52:553, 1993. Gupta, S. K., Poduslo, J. F., Dunn, R., et al.: Myelin-associated glycoprotein gene expression in the presence and absence of Schwann cell-axonal contact. Dev. Neurosci. 12:22, 1990. Gupta, S. K., Poduslo, J. F., and Mezei, C.: Temporal changes in PO and MBP gene expression after crush-injury of the adult peripheral nerve. Brain Res. 464:133, 1988. Guthrie, S., and Lumsden, A.: Motor neuron pathfinding following rhombomere reversals in the chick embryo hindbrain. Development 114:663, 1992. Haak, R. A., Kleinhans, F. W., and Ochs, S.: The viscosity of mammalian nerve axoplasm measured by electron spin resonance. J. Physiol. (Lond.) 263:115, 1976. Hall, S., Hughes, R., and Atkinson, P.: Lamellated sensory corpuscles within the endoneurium. J. Neurol. Neurosurg. Psychiatry 54:744, 1991. Hall, S. M., and Williams, P. L.: The distribution of electrondense tracers in peripheral nerve fibres. J. Cell Sci. 8:541, 1971. Hallin, R. G., Ekedahl, R., and Frank, O.: Segregation by modality of myelinated and unmyelinated fibers in human sensory nerve fascicles. Muscle Nerve 14:157, 1991. Halter, J. A., and Clark, J. W. Jr.: The influence of nodal constriction on conduction velocity in myelinated nerve fibers. Neuroreport 4:89, 1993. Hammerschlag, R., and Stone, G. C.: Further studies on the initiation of fast axonal transport. In Smith, R. S., and Bisby, M. A. (eds.): Axonal Transport. New York, Alan R Liss, p. 37, 1987. Hannah, R. S., and Nathaniel, E. J.: The postnatal development of blood vessels in the substantia gelatinosa of rat cervical cord—an ultrastructural study. Anat. Rec. 178:691, 1974. Harrison, R. G.: Neuroblast versus sheath cell in the development of peripheral nerves. J. Comp. Neurol. 37:123, 1924. Haug, H.: The membrane limitans gliae superficialis of cat’s visual cortex [in German]. Z. Zellforsch. Mikrosk. Anat. 115:79, 1971.

84

Structure of the Peripheral Nervous System

183. Herbst, F.: Untersuchungen ueber Metallsalzreaktionen an den Ranvierschen Schnuerringen. Acta Histochem. 22:223, 1965. 184. Heriot, K., Gambetti, P., and Lasek, R. J.: Proteins transported in slow components a and b of axonal transport are distributed differently in the transverse plane of the axon. J. Cell Biol. 100:1167, 1985. 185. Hess, A., and Lansing, A. I.: The fine structure of peripheral nerve fibres. Anat. Rec. 117:175, 1953. 186. Hess, A., and Young, J. Z.: The nodes of Ranvier. Proc. R. Soc. Lond. B 140:301, 1952. 187. Hirano, A.: Nodes of Ranvier in pathological conditions. In Zagoren, J. C., and Fedoroff, S. (eds.): The Node of Ranvier. New York, Academic Press, p. 213, 1984. 188. Hirano, A., and Dembitzer, H. M.: Further studies on the transverse bands. J. Neurocytol. 11:861, 1982. 189. Hirokawa, N.: Cross-linker system between neurofilaments, microtubules, and membranous organelles in frog axons revealed by the quick-freeze, deep-etching method. J. Cell Biol. 94:129, 1982. 190. Hirokawa, N.: 270K microtubule-associated protein crossreacting with anti-MAP2 IgG in the crayfish peripheral nerve axon. J. Cell Biol. 103:33, 1986. 191. Hirokawa, N.: Molecular architecture and dynamics of the neuronal cytoskeleton. In Burgoyne, R. D. (ed.): The Neuronal Cytoskeleton. New York, Wiley-Liss, p. 5, 1991. 192. Hirokawa, N., Bloom, G. S., and Vallee, R. B.: Cytoskeletal architecture and immunocytochemical localization of microtubule-associated proteins in regions of axons associated with rapid axonal transport: the beta,beta⬘iminodipropionitrile-intoxicated axon as a model system. J. Cell Biol. 101:227, 1985. 193. Hoffman, P. N., Cleveland, D. W., Griffin, J. W., et al.: Neurofilament gene expression: a major determinant of axonal caliber. Proc. Natl. Acad. Sci. U. S. A. 84:3472, 1987. 194. Hoffman, P. N., Griffin, J. W., and Price, D. L.: Control of axonal caliber by neurofilament transport. J. Cell Biol. 99:705, 1984. 195. Hoffman, P. N., Thompson, G. W., Griffin, J. W., and Price, D. L.: Changes in neurofilament transport coincide temporally with alterations in the caliber of axons in regenerating motor fibers. J. Cell Biol. 101:1332, 1985. 196. Hoke, A., Ho, T., Crawford, T. O., et al.: Glial cell linederived neurotrophic factor alters axon Schwann cell units and promotes myelination in unmyelinated nerve fibers. J. Neurosci. 23:561, 2003. 197. Hollenbeck, P. J.: The transport and assembly of the axonal cytoskeleton. J. Cell Biol. 108:223, 1989. 198. Holtzman, E.: Membrane trafficking in neurons. Curr. Opin. Neurobiol. 2:607, 1992. 199. Hopkins, W. G., and Brown, M. C.: The distribution of nodal sprouts in a paralysed or partly denervated mouse muscle. Neuroscience 7:37, 1982. 200. Hopkins, W. G., Brown, M. C., and Keynes, R. J.: Nerve growth from nodes of Ranvier in inactive muscle. Brain Res. 222:125, 1981. 201. Hromada, J.: On the nerve supply of the connective tissue of some peripheral nervous system components. Acta Anat. 55:343, 1963.

202. Hulles, E.: Beiträge zur Kenntnis der sensiblen Wurzeln der Medulla oblongata beim Menschen. Arb. Neurol. Inst. Wien Univ. 7:392, 1900. 203. Huttner, W. B., and Dotti, C. G.: Exocytotic and endocytotic membrane traffic in neurons. Curr. Opin. Neurobiol. 1:388, 1991. 204. Ichimura, T., and Ellisman, M. H.: Three-dimensional fine structure of cytoskeletal-membrane interactions at nodes of Ranvier. J. Neurocytol. 20:667, 1991. 205. Jacobs, J. M.: Toxic effects on the node of Ranvier. In Zagoren, J. C., and Fedoroff, S. (eds.): The Node of Ranvier. New York, Academic Press, p. 245, 1984. 206. Jacobs, J. M.: Morphological changes at paranodes in IgM paraproteinaemic neuropathy. Microsc. Res. Tech. 34:544, 1996. 207. Jacobs, J. M., and Love, S.: Qualitative and quantitative morphology of human sural nerve at different ages. Brain 108:897, 1985. 208. Janetzko, A., Zimmermann, H., and Volknandt, W.: Intraneuronal distribution of a synaptic vesicle membrane protein: antibody binding sites at axonal membrane compartments and trans-Golgi network and accumulation at nodes of Ranvier. Neuroscience 32:65, 1989. 209. Jessen, K. R., and Mirsky, R.: Schwann cells and their precursors emerge as major regulators of nerve development. Trends Neurosci. 22:402, 1999. 210. Jessen, K. R., and Mirsky, R.: Signals that determine Schwann cell identity. J. Anat. 200:367, 2002. 211. Kaar, G. F. Development of the transitional region between central and peripheral nervous systems: a morphological and quantitative study of the L5 ventral root. Ph.D. Thesis, Dublin, National University of Ireland, 1984. 212. Kaar, G. F., and Fraher, J. P.: The sheaths surrounding the attachments of rat lumbar ventral roots to the spinal cord: a light and electron microscopical study. J. Anat. 148:137, 1986. 213. Kaar, G. F., and Fraher, J. P.: The vascularisation of the central-peripheral transitional zone of rat lumbar ventral rootlets: a morphological and morphometric study. J. Anat. 150:145, 1987. 214. Kanda, T., Tsukagoshi, H., Oda, M., et al.: Morphological changes in unmyelinated nerve fibres in the sural nerve with age. Brain 114:585, 1991. 215. Kenny, A. J., and Booth, A. G.: Microvilli: their ultrastructure, enzymology and molecular organization. Essays Biochem. 14:1, 1978. 216. Key, A., and Retzius, G.: Studien in der Anatomie des Nervensystems und des Bindegewebes. Stockholm, Samson & Wallin, 1876. 217. King, R. H. M., and Thomas, P. K.: Electron microscope observations on aberrant regeneration of unmyelinated axons in the vagus nerve of the rabbit. Acta Neuropathol. (Berl.) 18:150, 1971. 218. Kocsis, J. D.: Functional organization of potassium channels in normal and pathological mammalian axons. In Zagoren, J. C., and Fedoroff, S. (eds.): The Node of Ranvier. New York, Academic Press, p. 183, 1984. 219. Kocsis, J. D., and Waxman, S. G.: Ionic channel organization of normal and regenerating mammalian axons. Prog. Brain Res. 71:89, 1987.

Microscopic Anatomy of the Peripheral Nervous System 220. Kordeli, E., Davis, J., Trapp, B., Bennett, V.: An isoform of ankyrin is localized at nodes of Ranvier in myelinated axons of central and peripheral nerves. J. Cell Biol. 110:1341, 1990. 221. Kosik, K. S.: The molecular and cellular biology of tau. Brain Pathol. 3:39, 1993. 222. Krammer, E. B., and Lischka, M. F.: Schwermetallaffine Strukturen des peripheren Nerven. I. Potentieller Stürfaktor beim cytochemischen AChENachweis. Histochemie 36:269, 1973. 223. Kristensson, K., and Olsson, Y.: The perineurium as a diffusion barrier to protein tracers: differences between mature and immature animals. Acta Neuropathol. (Berl.) 17:127, 1971. 224. Kristol, C., Sandri, C., and Akert, K.: Intramembranous particles at the nodes of Ranvier of the cat spinal cord: a morphometric study. Brain Res. 142:391, 1978. 225. Krnjevic, K.: The distribution of Na and K in cat nerves. J. Physiol. (Lond.) 128:473, 1955. 226. Kuntz, A.: Experimental studies on the histogenesis of the sympathetic nervous system. J. Comp. Neurol. 34:1, 1922. 227. Laidlaw, G. F.: Silver staining of the endoneurial fibers of the cerebrospinal nerves. Am. J. Pathol. 6:435, 1930. 228. Landon, D. N.: Structure of normal peripheral myelinated nerve fibres. In Waxman, S. G., and Ritchie, J. M. (eds.): Demyelinating Disease: Basic and Clinical Electrophysiology. New York, Raven Press, p. 25, 1981. 229. Landon, D. N., and Hall, S.: The myelinated nerve fibre. In Landon, D. N. (ed.): The Peripheral Nerve. London, Chapman and Hall, p. 1, 1976. 230. Landon, D. N., and Langley, O. K.: The local chemical environment of nodes of Ranvier: a study of cation binding. J. Anat. 108:419, 1971. 231. Landon, D. N., and Williams, P. L.: Ultrastructure of the node of Ranvier. Nature 199:575, 1963. 232. Langford, G. M., Allen, R. D., and Weiss, D. G.: Substructure of sidearms on squid axoplasmic vesicles and microtubules visualized by negative contrast electron microscopy. Cell Motil. Cytoskeleton 7:20, 1987. 233. Langley, O. K.: A comparison of the binding of Alcian blue and inorganic cations to polyanions in peripheral nerve. Histochem. J. 3:251, 1971. 234. Langley, O. K., and Landon, D. N.: Copper binding at nodes of Ranvier: a new electron histochemical technique for the demonstration of polyanions. J. Histochem. Cytochem. 17:66, 1969. 235. Lasek, R. J.: Axonal transport: a dynamic view of neuronal structures. Trends Neurosci. 3:87, 1980. 236. Lasek, R. J.: Translocation of the neuronal cytoskeleton and axonal locomotion. Philos. Trans. R. Soc. Lond. B Biol. Sci. 299:313, 1982. 237. Lasek, R. J., Garner, J. A., and Brady, S. T.: Axonal transport of the cytoplasmic matrix. J. Cell Biol. 99:212s, 1984. 238. Le Beau, J. M., Powell, H. C., and Ellisman, M. H.: Node of Ranvier formation along fibres regenerating through silicone tube implants: a freeze-fracture and thin-section electron microscopic study. J. Neurocytol. 16:347, 1987. 239. Lemke, G., and Chao, M.: Axons regulate Schwann cell expression of the major myelin and NGF receptor genes. Development 102:499, 1988.

85

240. Levi, E.: Studien zur normalen und pathologischen Anatomie der hinteren Rückenmarkswurzeln. Arb. Neurol. Inst. Wien Univ. 13:62, 1906. 241. Levine, J., and Willard, M.: Fodrin: axonally transported polypeptides associated with the internal periphery of many cells. J. Cell Biol. 90:631, 1981. 242. Lev-Ram, V., and Ellisman, M. H.: Axonal activation-induced calcium transients in myelinating Schwann cells, sources, and mechanisms. J. Neurosci. 15:2628, 1995. 243. Lewis, S. A., Ivanov, I. E., Lee, G. H., Cowan, N. J: Organization of microtubules in dendrites and axons is determined by a short hydrophobic zipper in microtubuleassociated proteins MAP2 and tau. Nature 342:498, 1989. 244. Lieberman, E. M., Abbott, N. J., and Hassan, S.: Evidence that glutamate mediates axon-to-Schwann cell signaling in the squid. Glia 2:94, 1989. 245. Lilje, O.: The processing and presentation of endogenous and exogenous antigen by Schwann cells in vitro. Cell. Mol. Life Sci. 59:2191, 2002. 246. Lindsey, J. D., and Ellisman, M. H.: The neuronal endomembrane system. III. The origins of the axoplasmic reticulum and discrete axonal cisternae at the axon hillock. J. Neurosci. 5:3135, 1985. 247. Livesey, F. J., and Fraher, J. P.: Experimental traction injuries of cervical spinal nerve roots: a scanning EM study of rupture patterns in fresh tissue. Neuropathol. Appl. Neurobiol. 18:376, 1992. 248. Livingston, R. B., Pfenniger, K., Moor, H., Akert, K.: Specialized paranodal and interparanodal glial-axonal junctions in the peripheral and central nervous system: a freeze-etching study. Brain Res. 58:1, 1973. 249. LoPachin, R. M., Castiglia, C. M., Lehning, E., and Saubermann, A. J.: Effects of acrylamide on subcellular distribution of elements in rat sciatic nerve myelinated axons and Schwann cells. Brain Res. 608:238, 1993. 250. Lubinska, L., and Lukaszewska, I.: Shape of myelinated nerve fibres and proximodistal flow of axoplasm. Acta Biol. Exp. 17:115, 1956. 251. Lumsden, A., and Keynes, R.: Segmental patterns of neuronal development in the chick hindbrain. Nature 337:424, 1989. 252. Lundborg, G., and Brånemark, P.-I.: Microvascular structure and function of peripheral nerves. Adv. Microcirc. 1:66, 1968. 253. Luse, S. A.: Electron microscopic study of brain tumours. Neurology 10:881, 1960. 254. Lustig, M., Zanazzi, G., Sakurai, T., et al.: Nr-CAM and neurofascin interactions regulate ankyrin G and sodium channel clustering at the node of Ranvier. Curr. Biol. 11:1864, 2001. 255. Martin, S., Levine, A. K., Chen, Z. J., et al.: Deposition of the NG2 proteoglycan at nodes of Ranvier in the peripheral nervous system. J. Neurosci. 21:8119, 2001. 256. Martz, D., Lasek, R. J., Brady, S. T., Allen R. D.: Mitochondrial motility in axons: membranous organelles may interact with the force generating system through multiple surface binding sites. Cell. Motil. 4:89, 1984. 257. Mata, M., Kupina, N., and Fink, D. J.: Phosphorylationdependent neurofilament epitopes are reduced at the node of Ranvier. J. Neurocytol. 21:199, 1992. 258. Mata, M., Staple, J., and Fink, D. J.: Cytochemical localization of Ca2⫹-ATPase activity in peripheral nerve. Brain Res. 445:47, 1988.

86

Structure of the Peripheral Nervous System

259. Maxwell, D. S.: Fine structure of the normal trigeminal ganglion in the cat and monkey. J. Neurosurg. 26:127, 1967. 260. Maxwell, D. S., Kruger, L., and Pineda, A.: The trigeminal nerve root with special reference to the central-peripheral transition zone: an electron microscopic study in the macaque. Anat. Rec. 164:113, 1969. 261. McDonald, K.: Osmium ferricyanide fixation improves microfilament preservation and membrane visualization in a variety of animal cell types. J. Ultrastruct. Res. 86:107, 1984. 262. McIntyre, C. C., Richardson, A. G., and Grill, W. M.: Modeling the excitability of mammalian nerve fibers: influence of afterpotentials on the recovery cycle. J. Neurophysiol. 87:995, 2002. 263. McQuarrie, I. G.: Effect of conditioning lesion on axonal sprout formation at nodes of Ranvier. J. Comp. Neurol. 231:239, 1985. 264. Mecham, R. P., and Heuser, J. E.: The elastic fiber. In Day, E. D. (ed.): Cell Biology of the Extracellular Matrix. New York, Plenum Press, p. 79, 1991. 265. Melendez-Vasquez, C. V., Rios, J. C., Zanazzi, G., et al.: Nodes of Ranvier form in association with ezrin-radixin-moesin (ERM)-positive Schwann cell processes. Proc. Natl. Acad. Sci. U. S. A. 98:1235, 2001. 266. Meller, K.: Early structural changes in the axoplasmic cytoskeleton after axotomy studied by cryofixation. Cell Tissue Res. 250:663, 1987. 267. Menegoz, M., Gaspar, P., Le Bert, M., et al.: Paranodin, a glycoprotein of neuronal paranodal membranes. Neuron 19:319, 1997. 268. Metuzals, J., and Tasaki, I.: Subaxolemmal filamentous network in the giant nerve fiber of the squid (Loligo pealei L.) and its possible role in excitability. J. Cell Biol. 78:597, 1978. 269. Mi, H., Deerinck, T. J., Ellisman, M. H., Schwarz, T. L.: Differential distribution of closely related potassium channels in rat Schwann cells. J. Neurosci. 15:3761, 1995. 270. Midroni, G., Bilbao, J. M., and Cohen, S. M.: Normal anatomy of peripheral (sural) nerve. In Biopsy Diagnosis of Peripheral Neuropathy. Boston, Butterworth-Heinemann, p. 13, 1995. 271. Miller, C. C., Ackerley, S., Brownlees, J., et al.: Axonal transport of neurofilaments in normal and disease states. Cell. Mol. Life Sci. 59:323, 2002. 272. Miller, R. G., and da Silva, P. P.: Particle rosettes in the periaxonal Schwann cell membrane and particle clusters in the axolemma of rat sciatic nerve. Brain Res. 130:135, 1977. 273. Mohammed, U. H. M., and Landon, D. N.: Axoplasmic asymmetry at the node of Ranvier. J. Anat. 137:820, 1983. 274. Mohammed, U. H. M., Landon, D. N., and Love, S.: Ultrastructural morphometry of frog nodes of Ranvier. J. Anat. 137:822, 1983. 275. Molander, M., Berthold, C. H., Persson, H., et al.: Monosialoganglioside (GM1) immunofluorescence in rat spinal roots studied with a monoclonal antibody. J. Neurocytol. 26:101, 1997. 276. Moll, C., and Meier, C.: The central-peripheral transition zone of cervical spinal nerve roots in Jimpy mutant and normal mice: light- and electron-microscopic study. Acta Neuropathol. (Berl.) 60:241, 1983.

277. Moore, G. R., Boegman, R. J., Robertson, D. M., and Raine, C. S.: Acute stages of batrachotoxin-induced neuropathy: a morphologic study of a sodium-channel toxin. J. Neurocytol. 15:573, 1986. 278. Moya, F., Bunge, M. B., and Bunge, R. P.: Schwann cells proliferate but fail to differentiate in defined medium. Proc. Natl. Acad. Sci. U. S. A. 77:6902, 1980. 279. Muller-Mohnssen, H., Tippe, A., Hillenkamp, F., and Unsold, E.: Is the rise of the action potential of the Ranvier node controlled by a paranodal organ? Naturwissenschaften 61:369, 1974. 280. Myers, R. R., Heckman, H. M., and Powell, H. C.: Endoneurial fluid is hypertonic: results of microanalysis and its significance in neuropathy. J. Neuropathol. Exp. Neurol. 42:217, 1983. 281. Nageotte, J.: L’Organisation de la Matière. Paris, Felix Alcan, 1922. 282. Nakazawa, E., and Ishikawa, H.: Occurrence of fasciculated microtubules at nodes of Ranvier in rat spinal roots. J. Neurocytol. 24:399, 1995. 283. Nathaniel, E. J., and Nathaniel, D. R.: Electron microscopic observations on the dorsal root-spinal cord junction. J. Cell Biol. 19:52a, 1963. 284. Nathaniel, E. J. H., and Pease, D. C.: Collagen and basement membrane formation by Schwann cells during nerve regeneration. J. Ultrastruct. Res. 9:550, 1963. 285. Nemecek, S., Parizek, J., Spacek, J., and Nemeckova, J.: Histological, histochemical and ultrastructural appearance of the transitional zone of the cranial and spinal nerve roots. Folia Morphol. (Warsz.) 17:171, 1969. 286. Niederlander, C., and Lumsden, A.: Late emigrating neural crest cells migrate specifically to the exit points of cranial branchiomotor nerves. Development 122:2367, 1996. 287. Nilsson, I., and Berthold, C. H.: Axon classes and internodal growth in the ventral spinal root L7 of adult and developing cats. J. Anat. 156:71, 1988. 288. Nixon, R. A.: The regulation of neurofilament protein dynamics by phosphorylation: clues to neurofibrillary pathobiology. Brain Pathol. 3:29, 1993. 289. Nixon, R. A., and Sihag, R. K.: Neurofilament phosphorylation: a new look at regulation and function. Trends Neurosci. 14:501, 1991. 290. Obersteiner, H., and Redlich, E.: Uber Wesen und Pathogenese der Tabischen Hinterstrangsdegeneration. Arb. Neurol. Inst. Wien Univ. 1:158, 1894. 291. Oblinger, M. M., Brady, S. T., McQuarrie, I. G., and Lasek, R. J.: Cytotypic differences in the protein composition of the axonally transported cytoskeleton in mammalian neurons. J. Neurosci. 7:453, 1987. 292. Obremski, V. J., Wood, P. M., and Bunge, M. B.: Fibroblasts promote Schwann cell basal lamina deposition and elongation in the absence of neurons in culture. Dev. Biol. 160:119, 1993. 293. O’Brien, D., Dockery, P., McDermott, K., and Fraher, J. P.: The ventral motoneurone axon bundle in the CNS—a cordone system? J. Neurocytol. 27:247, 1998. 294. O’Brien, D., Dockery, P., McDermott, K., and Fraher, J. P.: Early development of rat ventral root transitional zone: an immunohistochemical and morphometric study. J. Neurocytol. 30:11, 2001.

Microscopic Anatomy of the Peripheral Nervous System 295. Obst, T.: Über das Endgebeit des Perineuriums an den Zahnnerven der Ratte. Z. Zellforsch. Mikrosk. Anat. 114:515, 1971. 296. Ochoa, J.: The sural nerve of the human foetus: electron microscope observations and counts of axons. J. Anat. 108:231, 1971. 297. Ochoa, J., and Mair, W. G.: The normal sural nerve in man. I. Ultrastructure and numbers of fibres and cells. Acta Neuropathol. (Berl.) 13:197, 1969. 298. Ochoa, J., and Mair, W. G.: The normal sural nerve in man. II. Changes in the axons and Schwann cells due to ageing. Acta Neuropathol. (Berl.) 13:217, 1969. 299. Ochoa, J., and Vial, J. D.: Behaviour of peripheral nerve structures in chronic neuropathies, with special reference to the Schwann cell. J. Anat. 102:95, 1967. 300. Ochs, S., Jersild, R. A. Jr., and Li, J. M.: Slow transport of freely movable cytoskeletal components shown by beading partition of nerve fibers in the cat. Neuroscience 33:421, 1989. 301. Okabe, S., and Hirokawa, N.: Microtubule dynamics in nerve cells: analysis using microinjection of biotinylated tubulin into PC12 cells. J. Cell Biol. 107:651, 1988. 302. Okabe, S., and Hirokawa, N.: Turnover of fluorescently labelled tubulin and actin in the axon. Nature 343:479, 1990. 303. Okabe, S., and Hirokawa, N.: Do photobleached fluorescent microtubules move? Re-evaluation of fluorescence laser photobleaching both in vitro and in growing Xenopus axon. J. Cell Biol. 120:1177, 1993. 304. Olsson, Y.: Vascular permeability in the peripheral nervous system. In Dyck, P. J., Thomas, P. K., and Lambert, E. H. (eds.): Peripheral Neuropathy. Philadelphia, W. B. Saunders, p. 190–200, 1975. 305. Olsson, Y., and Reese, T. S.: Permeability of vasa nervorum and perineurium in mouse sciatic nerve studied by fluorescence and electron microscopy. J. Neuropathol. Exp. Neurol. 30:105, 1971. 306. Origuchi, Y.: Quantitative histological study in the sural nerves of children. Brain. Dev. 3:395, 1981. 307. Orkand, R. K., Nicholls, J. G., and Kuffler, S. W.: Effect of nerve impulses on the membrane potential of glial cells in the central nervous system of Amphibia. J. Neurophysiol. 29:788, 1966. 308. Paintal, A. S.: A comparison of the nerve impulses of mammalian non-medullated nerve fibres with those of the smallest diameter medullated fibres. J. Physiol. (Lond.) 193:523, 1967. 309. Pannese, E., Procacci, P., Ledda, M., et al.: A quantitative study of microtubules in motor and sensory axons. Acta Anat. (Basel) 118:193, 1984. 310. Pannese, E., Procacci, P., Ledda, M., et al.: Internodal microvilli of Schwann cells of myelinated fibres in lizard spinal roots project onto unmyelinated axons. J. Neurocytol. 18:295, 1989. 311. Parhad, I. M., Clark, A. W., and Griffin, J. W.: The effect of impairment of slow transport on axonal caliber. In Smith, R. S., and Bisby, M. A. (eds.): Axonal Transport. New York, Alan R Liss, p. 263, 1987. 312. Parmantier, E., Lynn, B., Lawson, D., et al.: Schwann cell-derived desert hedgehog controls the development of peripheral nerve sheaths. Neuron 23:713, 1999.

87

313. Pascher, R., Berthold, C. H., and Rydmark, M.: Computerassisted simulation of high-voltage electron microscopy using serial images recorded by conventional transmission electron microscopy. J. Electron Microsc. (Tokyo) 51:113, 2002. 314. Pedraza, L., Huang, J. K., and Colman, D. R.: Organizing principles of the axoglial apparatus. Neuron 30:335, 2001. 315. Peltonen, J., Jaakkola, S., Virtanen, I., and Pelliniemi, L.: Perineurial cells in culture: an immunocytochemical and electron microscopic study. Lab. Invest. 57:480, 1987. 316. Peters, A., Palay, S. L., Webster, H. deF.: The Fine Structure of the Nervous System: The Neurons and Supporting Cells. Philadelphia, W. B. Saunders, p. 233, 1976. 317. Phelps, C. H.: The development of glio-vascular relationships in the rat spinal cord: an electron microscopic study. Z. Zellforsch. Mikrosk. Anat. 128:555, 1972. 318. Phillips, D. D., Hibbs, R. G., Ellison, J. P., and Shapiro, H.: An electron microscopic study of central and peripheral nodes of Ranvier. J. Anat. 111:229, 1972. 319. Plenk, H.: Über argyrophile Fasern (Gitterfasern) und ihre Bildungszellen. Ergeb. Anat. Entwicklungsgesch. 27:302, 1927. 320. Plenk, H.: Der Schwannsche Scheide der markhaltigen Nervenfasern. Z. Mikrosk. Anat. Forsch. 36:191, 1934. 321. Poliak, S., Gollan, L., Salomon, D., et al.: Localization of Caspr2 in myelinated nerves depends on axon-glia interactions and the generation of barriers along the axon. J. Neurosci. 21:7568, 2001. 322. Poliak, S., and Peles, E.: The local differentiation of myelinated axons at nodes of Ranvier. Nat. Rev. Neurosci. 4:968, 2003. 323. Price, D. L., and Griffin, J. W.: Neurons and ensheathing cells as targets of disease processes. In Spencer, P. S., and Schaumburg, H. H. (eds.): Experimental and Clinical Neurotoxicity. Baltimore, Williams & Wilkins, p. 1, 1980. 324. Quarles, R. H., Hammer, J. H., and Trapp, B. D.: The immunoglobulin gene superfamily and myelination. In Dynamic Interactions of Myelin Proteins. New York, Alan R Liss, p. 49, 1990. 325. Quatacker, J.: The axonal reticulum in the neurons of the superior cervical ganglion of the rat as a direct extension of the Golgi apparatus. Histochem. J. 13:109, 1981. 326. Raine, C. S.: On the occurrence of Schwann cells within the normal central nervous system. J. Neurocytol. 5:371, 1976. 327. Raine, C. S.: Differences between the nodes of Ranvier of large and small diameter fibres in the P.N.S. J. Neurocytol. 11:935, 1982. 328. Raine, C. S., Finch, H., and Masone, A.: Axoplasmic asymmetry at the node of Ranvier. J. Neurocytol. 12:533, 1983. 329. Raisman, G.: Specialized neuroglial arrangement may explain the capacity of vomeronasal axons to reinnervate central neurons. Neuroscience 14:237, 1985. 330. Rambourg, A., Clermont, Y., and Beaudet, A.: Ultrastructural features of six types of neurons in rat dorsal root ganglia. J. Neurocytol. 12:47, 1983. 331. Rambourg, A., and Droz, B.: Smooth endoplasmic reticulum and axonal transport. J. Neurochem. 35:16, 1980. 332. Ranvier, M. L.: Leçons sur l’Histologie du Système Nerveux. Paris, Librairie F. Savy, 1878.

88

Structure of the Peripheral Nervous System

333. Rasband, M. N., Trimmer, J. S., Schwarz, T. L., et al.: Potassium channel distribution, clustering, and function in remyelinating rat axons. J. Neurosci. 18:36, 1998. 334. Raven, C. P.: Experiments on the origin of sheath cells and sympathetic neuroblasts in Amphibia. J. Comp. Neurol. 67:221, 1937. 335. Rayner, D. A., and Baines, A. J.: A novel component of the axonal cortical cytoskeleton, A60, defined by a monoclonal antibody. J. Cell Sci. 94(Pt. 3):489, 1989. 336. Reale, E., Luciano, L., and Spitznas, M.: Freeze-fracture faces of the perineurial sheath of the rabbit sciatic nerve. J. Neurocytol. 4:261, 1975. 337. Reale, E., Luciano, L., and Spitznas, M.: Freeze-fracture aspects of the perineurium of spinal ganglia. J. Neurocytol. 5:385, 1976. 338. Remak, R.: Observationes Anatomicae et Microscopicae de Systematis Nervosi Structura. Berlin, Sumptibus et Formis Reimerianis, 1838. 339. Reynolds, R. J., and Heath, J. W.: Patterns of morphological variation within myelin internodes of normal peripheral nerve: quantitative analysis by confocal microscopy. J. Anat. 187:369, 1995. 340. Rieger, F., Daniloff, J. K., Pincon-Raymond, M., et al.: Neuronal cell adhesion molecules and cytotactin are colocalized at the node of Ranvier. J. Cell Biol. 103:379, 1986. 341. Ritchie, J. M., and Chiu, S. Y.: Distribution of sodium and potassium channels in mammalian myelinated nerve. In Waxman, S. G., and Ritchie, J. M. (eds.): Demyelinating Disease: Basic and Clinical Electrophysiology. New York, Raven Press, p. 329, 1981. 342. Rogart, R.: Structural and functional relationships of ion conduction in the myelinated and demyelinated axon. In Zagoren, J. C., and Fedoroff, S. (eds.): The Node of Ranvier. New York, Academic Press, p. 109, 1984. 343. Roper, J., and Schwarz, J. R.: Heterogeneous distribution of fast and slow potassium channels in myelinated rat nerve fibres. J. Physiol. (Lond.) 416:93, 1989. 344. Rosenbluth, J.: Intramembranous particle distribution at the node of Ranvier and adjacent axolemma in myelinated axons of the frog brain. J. Neurocytol. 5:731, 1976. 345. Rosenbluth, J.: Membrane specializations at the nodes of Ranvier and paranodal and juxtaparanodal regions of myelinated central and peripheral nerve fibers. In Zagoren, J. C., and Fedoroff, S. (eds.): The Node of Ranvier. New York, Academic Press, p. 31, 1984. 346. Rosenbluth, J.: Role of glial cells in the differentiation and function of myelinated axons. Int. J. Dev. Neurosci. 6:3, 1988. 347. Ross, M. D., and Burkel, W.: Electron microscopic observations of the nucleus, glial dome, and meninges of the rat acoustic nerve. Am. J. Anat. 130:73, 1971. 348. Ross, M. H., and Reith, E. J.: Perineurium: evidence for contractile elements. Science 165:604, 1965. 349. Rossiter, J. P., and Fraher, J. P.: Intermingling of central and peripheral nervous tissues in rat dorsolateral vagal rootlet transitional zones. J. Neurocytol. 19:385, 1990. 350. Rydmark, M.: Nodal axon diameter correlates linearly with internodal axon diameter in spinal roots of the cat. Neurosci. Lett. 24:247, 1981. 351. Rydmark, M., and Berthold, C. H.: Electron microscopic serial section analysis of nodes of Ranvier in lumbar spinal

352.

353.

354.

355.

356.

357. 358.

359.

360.

361. 362.

363.

364.

365.

366. 367.

368.

369.

roots of the cat: a morphometric study of nodal compartments in fibres of different sizes. J. Neurocytol. 12:537, 1983. Rydmark, M., Berthold, C. H., and Gatzinsky, K. P.: Paranodal Schwann cell mitochondria in spinal roots of the cat: an ultrastructural morphometric analysis. J. Neurocytol. 27:99, 1998. Saito, A., and Zachs, S. I.: Ultrastructure of Schwann and perineurial sheaths at the mouse neuromuscular junction. Anat. Rec. 164:379, 1969. Saitua, F., and Alvarez, J.: Microtubular packing varies along the course of motor and sensory axons: possible regulation of microtubules by environmental cues. Neurosci. Lett. 104:249, 1989. Salzer J. L.: Mechanisms of adhesion between axons and glial cells. In Waxman, S. G., Kocsis, J. D., and Stys, P. K. (eds.): The Axon: Structure, Function and Pathophysiology. Oxford, UK, Oxford University Press, p. 164, 1995. Salzer, J. L.: Clustering sodium channels at the node of Ranvier: close encounters of the axon-glia kind. Neuron 18:843, 1997. Salzer, J. L.: Nodes of Ranvier come of age. Trends Neurosci. 25:2, 2002. Samorajski, T., and Friede, R. L.: A quantitative electron microscopic study of myelination in the pyramidal tract of rat. J. Comp. Neurol. 134:323, 1968. Sawada, H., Furthmayr, H., Konomi, H., and Nagai, Y.: Immunoelectronmicroscopic localization of extracellular matrix components produced by bovine corneal endothelial cells in vitro. Exp. Cell Res. 171:94, 1987. Scherer, S. S.: Molecular specializations at nodes and paranodes in peripheral nerve. Microsc. Res. Tech. 34:452, 1996. Scherer, S. S.: Nodes, paranodes, and incisures: from form to function. Ann. N. Y. Acad. Sci. 883:131, 1999. Scherer, S. S., and Arroyo, E. J.: Recent progress on the molecular organization of myelinated axons. J. Peripher. Nerv. Syst. 7:1, 2002. Scherer, S. S., Xu, T., Crino, P., et al.: Ezrin, radixin, and moesin are components of Schwann cell microvilli. J. Neurosci. Res. 65:150, 2001. Schlaepfer, W. W., Freeman, L. A., and Eng, L. F.: Studies of human and bovine spinal nerve roots and the outgrowth of CNS tissues into the nerve root entry zone. Brain Res. 177:219, 1979. Schnapp, B., and Mugnaini, E.: Membrane architecture of myelinated fibers as seen by freeze-fracture. In Waxman, S. G. (eds.): Physiology and Pathobiology of Axons. New York, Raven Press, p. 83, 1978. Schnapp, B. J., and Reese, T. S.: Cytoplasmic structure in rapid-frozen axons. J. Cell Biol. 94:667, 1982. Schroder, J. M.: Developmental and pathological changes at the node and paranode in human sural nerves. Microsc. Res. Tech. 34:422, 1996. Schroder, J. M., Bohl, J., and Brodda, K.: Changes of the ratio between myelin thickness and axon diameter in the human developing sural nerve. Acta Neuropathol. (Berl.) 43:169, 1978. Schwann, T.: Mikroskopische Untersuchungen über die Übereinstimmung in der Struktur und dem Wachstum der Tiere und Pflanzen. Berlin, Sander, 1839.

Microscopic Anatomy of the Peripheral Nervous System 370. Schwartz, M., Ernst, S. A., Siegel, G. J., Agranoff, B. W.: Immunocytochemical localization of (Na⫹, K⫹)-ATPase in the goldfish optic nerve. J. Neurochem. 36:107, 1981. 371. Shantha, T. R., and Bourne, G. H.: The perineurial epithelium—a new concept. In The Structure and Function of Nervous Tissue. New York, Academic Press, p. 379, 1968. 372. Shanthaveerappa, T. R., and Bourne, G. H.: ‘The perineurial epithelium,’ a metabolically active, continuous, protoplasmic cell barrier surrounding peripheral nerve fasciculi. J. Anat. 96:527, 1962. 373. Shanthaveerappa, T. R., and Bourne, G. H.: Perineurial epithelium: a new concept of its role in the integrity of the peripheral nervous system. Science 153:1464, 1966. 374. Sheetz, M. P., and Martenson, C. H.: Axonal transport: beyond kinesin and cytoplasmic dynein. Curr. Opin. Neurobiol. 1:393, 1991. 375. Shellswell, G. B., Restall, D. J., Duance, V. C., and Bailey, A. J.: Identification and differential distribution of collagen types in the central and peripheral nervous systems. FEBS Lett. 106:305, 1979. 376. Shield, L. K., King, R. H. M., and Thomas, P. K.: A morphometric study of human fetal sural nerve. Acta Neuropathol. (Berl.) 70:60, 1986. 377. Shimazono, J.: Über das Verhalten der Zentralen und der peripheren Nervensubstanz bei verschiedenen Vergiftungen und Ernährungsstörungen. Arch. Psychiatr. Nervenkr. 53:872, 1914. 378. Shinogami, H.: Cytochemical-localization of K⫹-Nppase in the heminode of Ranvier in the rat optic-nerve. Acta Histochem. Cytochem. 22:207, 1989. 379. Sievers, J.: Basic two-dye stains for epoxy-embedded 0.3-1␮ sections. Stain Technol. 46:195, 1971. 380. Sima, A., and Sourander, P.: The effect of perinatal undernutrition on perineurial diffusion barrier to exogenous protein: an experimental study on rat sciatic nerve. Acta Neuropathol. (Berl.) 24:263, 1973. 381. Sima, A. A., and Brismar, T.: Reversible diabetic nerve dysfunction: structural correlates to electrophysiological abnormalities. Ann. Neurol. 18:21, 1985. 382. Sims, T. J., and Gilmore, S. A.: Interactions between intraspinal Schwann cells and the cellular constituents normally occurring in the spinal cord: an ultrastructural study in the irradiated rat. Brain Res. 276:17, 1983. 383. Sims, T. J., and Gilmore, S. A.: Interactions between Schwann cells and CNS axons following a delay in the normal formation of central myelin. Exp. Brain Res. 75:513, 1989. 384. Sims, T. J., and Gilmore, S. A.: Regeneration of dorsal root axons into experimentally altered glial environments in the rat spinal cord. Exp. Brain Res. 99:25, 1994. 385. Sims, T. J., and Gilmore, S. A.: Regrowth of dorsal root axons into a radiation-induced glial-deficient environment in the spinal cord. Brain Res. 634:113, 1994. 386. Sims, T. J., Gilmore, S. A., Waxman, S. G., and Klinge, E.: Dorsal-ventral differences in the glia limitans of the spinal cord: an ultrastructural study in developing normal and irradiated rats. J. Neuropathol. Exp. Neurol. 44:415, 1985.

89

387. Sindou, M., Quoex, C., and Baleydier, C.: Fiber organization at the posterior spinal cord-rootlet junction in man. J. Comp. Neurol. 153:15, 1974. 388. Skinner, H. A.: Some histologic features of cranial nerves. Arch. Neurol. Psychiatry 25:356, 1931. 389. Small, J. R., Scadding, J. W., and Landon, D. N.: Ultrastructural localization of sympathetic axons in experimental rat sciatic nerve neuromas. J. Neurocytol. 25:573, 1996. 390. Smith, R. S.: The short term accumulation of axonally transported organelles in the region of localized lesions of single myelinated axons. J. Neurocytol. 9:39, 1980. 391. Smith, R. S.: Control of the direction of rapid axonal transport in the vertebrates. In Smith, R. S., and Bisby, M. A. (eds.): Axonal Transport. New York, Alan R Liss Inc, p. 139, 1987. 392. Smith, R. S.: Real-time imaging of axonally transported subresolution organelles in vertebrate myelinated axons. J. Neurosci. Methods 26:203, 1989. 393. Smith, R. S., and Forman, D. S.: Organelle dynamics in lobster axons: anterograde and retrograde particulate organelles. Brain Res. 446:26, 1988. 394. Snyder, R.: The organization of the dorsal root entry zone in cats and monkeys. J. Comp. Neurol. 174:47, 1977. 395. Sonnenberg-Riethmacher, E., Miehe, M., Stolt, C. C., et al.: Development and degeneration of dorsal root ganglia in the absence of the HMG-domain transcription factor Sox10. Mech. Dev. 109:253, 2001. 396. Spacek, J., and Lieberman, A. R.: Relationships between mitochondrial outer membranes and agranular reticulum in nervous tissue: ultrastructural observations and a new interpretation. J. Cell Sci. 46:129, 1980. 397. Spencer, P. S.: Reappraisal of the model for “bulk axoplasmic flow.” Nat. New Biol. 240:283, 1972. 398. Spencer, P. S., and Thomas, P. K.: Ultrastructural studies of the dying-back process. II. The sequestration and removal by Schwann cells and oligodendrocytes of organelles from normal and diseased axons. J. Neurocytol. 3:763, 1974. 399. Spencer, P. S., Weinberg, H. J., Raine, C. S., and Prineas, J. W.: The perineurial window—a new model of focal demyelination and remyelination. Brain Res. 96:323, 1975. 400. Spray, D. C., and Dermietzel, R.: X-linked dominant CharcotMarie-Tooth disease and other potential gap-junction diseases of the nervous system. J. Comp. Neurol. 18:256, 1995. 401. Steer, J. M.: Some observations on the fine structure of rat dorsal spinal nerve roots. J. Anat. 109:467, 1971. 402. Stelzner, D. J.: The relationship between synaptic vesicles, Golgi apparatus, and smooth endoplasmic reticulum: a developmental study using the zinc iodide-osmium technique. Z. Zellforsch. Mikrosk. Anat. 120:332, 1971. 403. Stolinski, C.: Structure and composition of the outer connective tissue sheaths of peripheral nerve. J. Anat. 186:123, 1995. 404. Stolinski, C., and Breathnach, A. S.: Freeze-fracture replication of mammalian peripheral nerve—a review. J. Neurol. Sci. 57:1, 1982. 405. Stolinski, C., Breathnach, A. S., Martin, B., et al.: Associated particle aggregates in juxtaparanodal axolemma and adaxonal Schwann cell membrane of rat peripheral nerve. J. Neurocytol. 10:679, 1981.

90

Structure of the Peripheral Nervous System

406. Stone, G. C., and Hammerschlag, R.: Molecular mechanisms involved in sorting of fast-transported proteins. In Smith, R. S., and Bisby, M. A. (eds.): Axonal Transport. New York, Alan R Liss, p. 15, 1987. 407. Strupp, M., and Grafe, P.: A chloride channel in rat and human axons. Neurosci. Lett. 133:237, 1991. 408. Sturrock, R. R.: A quantitative and morphological study of vascularisation of the developing mouse spinal cord. J. Anat. 132:203, 1981. 409. Suarez, I., and Fernandez, B.: Structure and ultrastructure of the external glial layer in the hypothalamus of the hamster. J. Hirnforsch. 24:99, 1983. 410. Suarez, I., and Raff, M. C.: Subpial and perivascular astrocytes associated with nodes of Ranvier in the rat optic-nerve. J. Neurocytol. 18:577, 1989. 411. Sugimoto, Y., Takayama, S., Horiuchi, Y., Toyama, Y.: An experimental study on the perineurial window. J. Peripher. Nerv. Syst. 7:104, 2002. 412. Sunderland, S.: The connective tissues of peripheral nerves. Brain 88:841, 1965. 413. Sunderland, S.: Nerve and Nerve Injuries. Edinburgh, Churchill Livingstone, 1978. 414. Suter, U., and Scherer, S.: Disease mechanisms in inherited neuropathies. Nat. Rev. Neurosci. 4:714, 2003. 415. Tarlov, I. M.: Structure of the nerve root. I. Nature of the junction between the central and the peripheral nervous system. Arch. Neurol. Psychiatry 37:555, 1937. 416. Tarlov, I. M.: Structure of the nerve root. II. Differentiation of sensory from motor roots: observations on identification of function in roots of mixed cranial nerves. Arch. Neurol. Psychiatry 37:1338, 1937. 417. Tashiro, T., and Komiya, Y.: Maturation and aging of the axonal cytoskeleton: biochemical analysis of transported tubulin. J. Neurosci. Res. 30:192, 1991. 418. Thomas, F. P.: Antibodies to GM1 and Gal(beta 1–3)GalNAc at the nodes of Ranvier in human and experimental autoimmune neuropathy. Microsc. Res. Tech. 34:536, 1996. 419. Thomas, P. K.: The connective tissue of peripheral nerve: an electron microscope study. J. Anat. 97:35, 1963. 420. Thomas, P. K., Beamish, N. G., Small, J. R., et al.: Paranodal structure in diabetic sensory polyneuropathy. Acta Neuropathol. (Berl.) 92:614, 1996. 421. Thomas, P. K., and Bhagat, S.: The effect of extraction of the intrafascicular contents of peripheral nerve trunks on perineurial structure. Acta Neuropathol. (Berl.) 43:135, 1978. 422. Thomas, P. K., and Jones, D. G.: The cellular response to nerve injury. II. Regeneration of the perineurium after nerve section. J. Anat. 101:45, 1967. 423. Thomas, P. K., and King, R. H. M.: The degeneration of unmyelinated axons following nerve section: an ultrastructural study. J. Neurocytol. 3:497, 1974. 424. Thomas, P. K., King, R. H. M., and Sharma, A. K.: Changes with age in the peripheral nerves of the rat. Acta Neuropathol. (Berl.) 52:1, 1980. 425. Thomas, P. K., Ochoa, J. L., Berthold, C. H., et al.: Microscopic anatomy of the peripheral nervous system. In Dyck, P. J., Thomas, P. K., Griffin, J. W., et al. (eds.): Peripheral Neuropathy, 3rd ed. Philadelphia, W. B. Saunders, p. 28, 1993. 426. Tippe, A., and Muller-Mohnssen, H.: Further experimental evidence for the synapse hypothesis of Na⫹-current activation

427.

428.

429.

430.

431.

432.

433.

434.

435.

436.

437.

438.

439.

440. 441.

442.

443.

and inactivation at the Ranvier node. Naturwissenschaften 62:490, 1975. Todd, B. A., Inman, C., Sedgwick, E. M., and Abbott, N. J.: Ionic permeability of the frog sciatic nerve perineurium: parallel studies of potassium and lanthanum penetration using electrophysiological and electron microscopic techniques. J. Neurocytol. 29:551, 2000. Toews, A. D., Armstrong, R., Holshek, J., et al.: Unloading and transfer of axonally transported lipids. In Smith, R. S., and Bisby, M. A. (eds.): Axonal Transport. New York, Alan R Liss, p. 327, 1987. Trapp, B. D., Hauer, P., and Lemke, G.: Axonal regulation of myelin protein mRNA levels in actively myelinating Schwann cells. J. Neurosci. 8:3515, 1988. Trapp, B. D., and Kidd, G. J.: Axo-glial septate junctions: the maestro of nodal formation and myelination? J. Cell Biol. 150:F97, 2000. Tsukita, S., and Ishikawa, H.: Three-dimensional distribution of smooth endoplasmic reticulum in myelinated axons. J. Electron Microsc. (Tokyo) 25:141, 1976. Tsukita, S., and Ishikawa, H.: The movement of membranous organelles in axons: electron microscopic identification of anterogradely and retrogradely transported organelles. J. Cell Biol. 84:513, 1980. Tsukita, S., and Ishikawa, H.: The cytoskeleton in myelinated axons—serial section study. Biomed. Res. (Tokyo) 2:424, 1981. Tsukita, S., Tsukita, S., Kobayashi, T., and Matsumoto, G.: Subaxolemmal cytoskeleton in squid giant axon. II. Morphological identification of microtubule- and microfilamentassociated domains of axolemma. J. Cell Biol. 102:1710, 1986. Tsukita, S., Usukura, J., Tsukita, S., and Ishikawa, H.: The cytoskeleton in myelinated axons: a freeze-etch replica study. Neuroscience 7:2135, 1982. Tuckett, I. L.: On the structure and degeneration of nonmyelinated nerve fibres. J. Physiol. (Lond.) 19:267, 1895. Uehara, M., and Ueshima, T.: Scanning electron microscopy of the superficial glial limiting membrane in the cat brain and spinal cord. Nippon Juigaku Zasshi 50:115, 1988. Uhrik, B., and Stampfli, R.: Ultrastructural observations on nodes of Ranvier from isolated single frog peripheral nerve fibres. Brain Res. 215:93, 1981. Ushiki, T., and Ide, C.: Three-dimensional architecture of the endoneurium with special reference to the collagen fibril arrangement in relation to nerve fibers. Arch. Histol. Jpn. 49:553, 1986. van Deurs, B., Petersen, O. W., Olsnes, S., and Sandvig, K.: The ways of endocytosis. Int. Rev. Cytol. 117:131, 1989. Vasilescu, V., and Filip, D. A.: The Ranvier node as a chemo-electric pulsatory unit: a study of its structurefunctions relations. Physiologie 16:83, 1979. Vega, J. A., Del Valle, M. M., Haro, J. J., et al.: The inner-core, outer-core and capsule cells of the human pacinian corpuscles: an immunohistochemical study. Eur. J. Morphol. 32:11, 1994. Virchow, R.: Rückenmark und Gehirn. In Die Cellularpathologie in ihrer Begrundung auf Physiologische und Pathologische Gewebelehre. Berlin, Hirschwald, p. 238, 1858.

Microscopic Anatomy of the Peripheral Nervous System 444. Voyvodic, J. T.: Target size regulates calibre and myelination of sympathetic axons. Nature 342:430, 1989. 445. Wagner, H. J., Barthel, J., and Pilgrim, C.: Permeability of the external glial limiting membrane of rat parietal cortex. Anat. Embryol. (Berl.) 166:427, 1983. 446. Wang, L., and Brown, A.: Rapid movement of microtubules in axons. Curr. Biol. 12:1496, 2002. 447. Watson, D.: Regional variation in the abundance of axonal cytoskeletal proteins. J. Neurosci. Res. 30:226, 1991. 448. Waxman, S. G.: Regional differentiation of the axon: a review with special reference to the concept of the multiplex neuron. Brain Res. 47:269, 1972. 449. Waxman, S. G.: Voltage-gated ion channels in axons: localization, function, and development. In Waxman, S. G., Kocsis, J. D., and Stys, P. K. (eds.): The Axon: Structure, Function and Pathophysiology. Oxford, UK, Oxford University Press, p. 218, 1995. 450. Waxman, S. G., and Black, J. A.: Macromolecular structure of the Schwann cell membrane: perinodal microvilli. J. Neurol. Sci. 77:23, 1987. 451. Waxman, S. G., and Quick, D. C.: Intra-axonal ferric ion-ferrocyanide staining of nodes of Ranvier and initial segments in central myelinated fibers. Brain Res. 144:1, 1978. 452. Waxman, S. G., and Ritchie, J. M.: Molecular dissection of the myelinated axon. Ann. Neurol. 33:121, 1993. 453. Webster, H. deF.: Transient, focal accumulation of axonal mitochondria during the early stages of wallerian degeneration. J. Cell Biol. 12:361, 1962. 454. Webster, H. deF., and Spiro, D.: Phase and electron microscopic studies of experimental demyelination. I. Variations in myelin sheath contour in normal guinea pig sciatic nerve. J. Neuropathol. Exp. Neurol. 19:42, 1960. 455. Weis, J., Alexianu, M. E., Heide, G., and Schröder, J. M.: Renaut bodies contain elastic fibre components. J. Neuropathol. Exp. Neurol. 52:444, 1993. 456. Weiss, D. G., Seitz-Tutter, D., Langford, G. M., et al.: The native microtubule as the engine for bidirectional organelle movements. In Smith, R. S., and Bisby, M. A. (eds.): Axonal Transport. New York, Alan R Liss, p. 91, 1987. 457. Weston, J. A.: A radioautographic analysis of the migration and localisation of trunk neural crest cells in the chick. Dev. Biol. 6:279, 1963. 458. Weston, J. A.: The migration and differentiation of neural crest cells. In Abercrombie, M., Brachet, J., and King, T. (eds.): Advances in Morphogenesis. New York, Academic Press, p. 127, 1970. 459. Wiley, C. A., and Ellisman, M. H.: Rows of dimericparticles within the axolemma and juxtaposed particles within glia, incorporated into a new model for the paranodal glial-axonal junction at the node of Ranvier. J. Cell Biol. 84:261, 1980. 460. Williams, P. L., and Hall, S. M.: In vivo observations on mature myelinated nerve fibres of the mouse. J. Anat. 107:31, 1970. 461. Williams, P. L., and Hall, S. M.: Prolonged in vivo observations of normal peripheral nerve fibres and their acute

462. 463.

464.

465.

466.

467.

468.

469.

470.

471.

472.

473.

474.

475.

476.

477.

478.

91

reactions to crush and deliberate trauma. J. Anat. 108:397, 1971. Williams, P. L., and Kashef, R.: Asymmetry of the node of Ranvier. J. Cell Sci. 3:341, 1968. Williams, P. L., and Landon, D. N.: Paranodal apparatus of peripheral nerve fibres of mammals. Nature 198:670, 1963. Williams, V.: Intercellular relationships in the external glial limiting membrane of the neocortex of the cat and rat. Am. J. Anat. 144:421, 1975. Windebank, A. J., Wood, P., Bunge, R. P., and Dyck, P. J.: Myelination determines the caliber of dorsal root ganglion neurons in culture. J. Neurosci. 5:1563, 1985. Wood, J. G., Jean, D. H., Whitaker, J. N., et al.: Immunocytochemical localization of the sodium, potassium activated ATPase in knifefish brain. J. Neurocytol. 6:571, 1977. Wulfhekel, U.: The passage from central to peripheral structure of cranial nerves in the lower vertebrates [in German]. Acta Anat. (Basel) 74:183, 1969. Wurtz, C. C., and Ellisman, M. H.: Alterations in the ultrastructure of peripheral nodes of Ranvier associated with repetitive action potential propagation. J. Neurosci. 6:3133, 1986. Wurtz, C. C., and Ellisman, M. H.: Activity associated ultrastructural changes in peripheral nodes of Ranvier are independent of fixation. Exp. Neurol. 101:87, 1988. Yamamoto, K., Merry, A. C., and Sima, A. A.: An orderly development of paranodal axoglial junctions and bracelets of Nageotte in the rat sural nerve. Brain Res. Dev. Brain Res. 96:36, 1996. Zagoren, J. C.: Cation binding at the node of Ranvier. In Zagoren, J. C., and Fedoroff, S. (eds.): The Node of Ranvier. New York, Academic Press, p. 69, 1984. Zagoren, J. C., and Arezzo, J. C.: Cation binding at the node of Ranvier: II. Redistribution of binding sites during electrical stimulation. Brain Res. 242:27, 1982. Zagoren, J. C., Raine, C. S., and Suzuki, K.: Cation binding at the node of Ranvier: I. Localization of binding sites during development. Brain Res. 242:19, 1982. Zelena, J.: Arrays of glycogen granules in the axoplasm of peripheral nerves at pre-ovoid stages of wallerian degeneration. Acta Neuropathol. (Berl.) 50:227, 1980. Zenker, W., and Hohberg, E.: A-alpha-nerve-fiber: number of neurotubules in the stem fibre and in the terminal branches. J. Neurocytol. 2:143, 1973. Zenker, W., Mayr, R., and Gruber, H.: Axoplasmic organelles: quantitative differences between ventral and dorsal root fibres of the rat. Experientia 29:77, 1973. Zenker, W., Mayr, R., and Gruber, H.: Neurotubules: different densities in peripheral motor and sensory nerve fibres. Experientia 31:318, 1975. Zimmermann, H., and Vogt, M.: Membrane proteins of synaptic vesicles and cytoskeletal specializations at the node of Ranvier in electric ray and rat. Cell Tissue Res. 258:617, 1989.

4 Channel Function in Mammalian Axons and Support Cells S. Y. CHIU

Na⫹ Channels Na⫹ Channels on Axons Molecular Identity of Nav Channels on Axons Mechanisms for Nav Channel Segregation

Na⫹ Channels on Schwann Cells K⫹ Channels Molecular Identity of Axonal Kv Channels: A Tripartite of Kv1.1/Kv1.2/Kv␤2

Voltage-gated ion channels are porelike, membranelodged proteins that mediate rapid ionic flux (106 ions/s) across cell membranes. Voltage-gated channels are important for axons because they allow action potentials to be generated; the initial rapid influx of sodium ions through sodium channels causes the rapid upstroke of an action potential followed by a delayed efflux of potassium ions through potassium channels to repolarize the membrane. This cycle of depolarization and hyperpolarization also generates axial current along the axoplasm, which excites neighboring membrane patches and propels action potentials along nerve fibers at a high speed. This model of excitation, referred to as the Hodgkin-Huxley model of nerve excitation,52 has, since its inception in 1952, formed the basis for understanding fast electrical signaling in all nerve models, in both health and disease. In the early 1980s, Neher and Sakmann ushered in the phase of molecular study of ion channels with the invention of the patchclamp technique, which allows the electrical events associated with the opening and closing of a single ion channel protein to be observed. Cloning of ion channels then followed and dominated the field. As of the writing of this chapter, 50 potassium channel genes and 10 sodium channel genes have been identified in mammals; many of them will undoubtedly be given proper addresses in the human genome once the Human Genome Project is completed. Molecular biology also allows ion channel crystals to be grown and their three-dimensional structure solved. In a

Mechanisms for Clustering Kv Channels Function of Kv Channels K⫹ Channels on Schwann Cells

brief span of 3 years (1998 to 2001), MacKinnon and colleagues42,51,84,149 revolutionized the biophysical study of ion channels by solving the pore structure of potassium channels by x-ray crystallography at 2.0- to 3.2-Å resolution, producing a vivid molecular contour of the selectivity filter through which K⫹ ions are seen navigating, captured in a stunning series of frozen molecular snapshots. Basic studies of ion channels in axons and ensheathing cells have gained considerably from genetic cloning, because many ion channel genes are expressed in peripheral nerves, allowing ion channels in these nerves to be studied in molecular terms unthinkable 9 years ago when this chapter was last written. On the clinical front, neuropathy increasingly is linked to mutations of known ion channel genes, offering the potential for designing treatment based on precise molecular targets. This chapter summarizes the current status of ion channel research in axons and ensheathing cells of the peripheral nerves with a primary focus on three major issues: the molecular identity of ion channels, the mechanisms responsible for localization of ion channels, and the normal and pathophysiologic functions of ion channels. When necessary, I have tried to give a brief account of the background to the rapidly evolving molecular research that, for the past 5 years, has shaped progress and defined future questions. Readers interested in a fuller description of the pre–molecular biologic background should consult my earlier chapter written for the third edition of Peripheral Neuropathy.29 95

96

Function of the Peripheral Nervous System

Naⴙ CHANNELS

A

intact

paranodal demyelination

ⴙ

Na Channels on Axons The peripheral nervous system has two types of axon: the large, fast-conducting myelinated axons and the small, slower conducting nonmyelinated axons. Voltage-gated Na⫹ channels (Nav) are sodium-selective pores whose opening and closing is gated by voltage. Nav channels have different distribution patterns on these two types of axon. In the nonmyelinated axons, Nav channels are uniformly distributed; in contrast, Nav channels in the myelinated axon are clustered at the node of Ranvier. Nodal clustering of ion channels was first hypothesized by Rosenbluth,108 who, based on freeze-fracture studies of myelinated axons, suggested that Nav channels are localized to the nodal membrane and trapped there by the paranodal junction. The first experimental evidence for Nav channel segregation came from studies of binding of 3H-saxitoxin to sodium channels of myelinated nerves. Noting that there is no difference in saxitoxin binding to intact and homogenized sciatic nerves, Ritchie and Rogart106 concluded that Nav channels are concentrated at the nodal region and absent in the internodal membrane masked by the myelin. This finding was later confirmed in voltage-clamp studies of single mammalian myelinated fibers in which acute paranodal demyelination was found to add little to the nodal sodium currents (Fig. 4–1A).31 Interestingly, by directly dissolving the myelin of a single internode, voltage clamp studies do show a low density of Nav channel in the internodes of adult myelinated nerves.37 Measurement of gating current (the tiny capacitative current associated with the activation of a single Nav channel) suggests that a single mammalian node of Ranvier has approximately 82,000 channels,24 which, with an assumed nodal area of approximately 60 ␮m2, translates to a very high channel density of approximately 1400/␮m2.

Molecular Identity of Nav Channels on Axons Do axonal Nav channels form a single homogeneous pool? Do different types of axons express different types of Nav channels? Earlier voltage-clamp studies revealed that nodal sodium currents in myelinated fibers inactivate in a double exponential time course, a finding consistent either with a single Nav type with second-order inactivation kinetics23 or two types of Nav channels with different inactivation kinetics.10 Furthermore, Nav channels exhibit different tetrodotoxin sensitivity. Hence, classic studies have pointed to diversity in axonal Nav expression. This issue is now solved in spectacular molecular detail by the cloning of Nav channel genes. At the writing of this chapter, 10 Nav channel genes (Nav1.1 through Nav1.9) have been cloned in mammals, and many of these are expressed

10

10

nA

nA 0

0

-10

-10

0 nA -5

0 nA -5

0

B

C

msec

0

4

msec

4

nodal Nav1.6

node

Juxtaparanodal Kv1.1

FIGURE 4–1 Complementary distribution of Na⫹ and K⫹ channels in mammalian PNS myelinated axons. A, Discovery of complementary distribution of Na⫹ and K⫹ channels in 1980. A single mammalian myelinated fiber was voltage-clamped before (left) and 15 minutes after (right) acute paranodal demyelination. Demyelination reveals a large outward K⫹ current (upward deflection) without affecting the inward Na⫹ current (downward deflection), suggesting that Na⫹ channels are localized to the node and K⫹ channels are under the myelin sheath. (From Chiu, S. Y., and Ritchie, J. M.: Potassium channels in nodal and internodal axonal membrane of mammalian myelinated fibres. Nature 284:170, 1980, with permission.) B and C, Recent immunohistochemical confirmation of complementary distribution of Na⫹ and K⫹ channels. B, Staining of Nav1.6 at the node (red). The paranodal adhesion molecule contactin is stained green. (Courtesy of Rock Levinson.) C, Staining of Kv1.1 in the juxtaparanode (green). Note the lack of Kv1.1 staining at the nodal gap. Paranodal contactin is stained red. (Courtesy of Steve Scherer and Arroyo Edgardo.)

in neurons.47,48 These different Nav channel genes encode sodium channel proteins with similar structural motifs but with different kinetic properties as a result of variation in other regions of their primary amino acid sequence.47

Channel Function in Mammalian Axons and Support Cells

Expression of different Nav isoforms in different axon types would greatly enrich electrophysiologic signaling in the peripheral nervous system (PNS). When investigators perform systematic histochemical analysis of Nav channel expression on axons with isoform-specific antibodies, a fascinating picture of Nav channel expression emerges. Small C-type nonmyelinated axons of neurons express Nav1.9 channels.71 This channel generates a persistent sodium current, potentially rendering C fibers sensitive to subthreshold stimulations. Nav1.2 is expressed predominantly in nonmyelinated axons,49,137 whereas Nav1.6 is expressed predominantly at nodes of Ranvier (red in Fig. 4–1B).21 The story took an intriguing turn when it was found that the Nav subtype switched from Nav1.2 to Nav1.6 when axons become myelinated, a switch first identified in central nervous system (CNS) axons14 but that since has been found applicable to PNS axons as well.99 At present, the functional consequence of this developmental switch in Nav channel subtype is unclear. One idea is that this developmental switch reflects a change in the anchoring proteins needed to cluster Nav channels during myelinogenesis; apparently changing the lock changes the key that fits. Mutations of Nav channels are neuropathologic.62,119 For example, mutation in Nav1.2 causes seizures in mice in transgenic studies, and it has been suggested that humans with mutation of this gene may be candidates for seizure disorders.62 A mutation of the Nav1.6 channel in mice (med) results in a truncated, nonfunctional gene product,20,64 causing mice to die 21 to 28 days after birth. Interestingly, even though the med mice lack immunostaining of Nav1.6 channels in the Ranvier nodes in sciatic nerves,21 the conduction velocity is only mildly affected.43,104 Furthermore, [3H]-tetrodotoxin binding is significantly increased in med sciatic nerves,104 suggesting an upregulation of a yet-to-be identified Nav isoform that compensates for the loss of Nav1.6 channels. The interesting point here is that deleting Nav1.6 also causes paranodal dysmyelination and nodal widening. This finding, together with the switch from Nav1.2 to Nav1.6 channels during myelination, suggests that inappropriate expression of nodal Nav channel isoforms is incompatible with normal nodal structure.

Mechanisms for Nav Channel Segregation Clustering of Nav channels at the Ranvier node is critical for generating high-speed nerve conduction in adult myelinated fibers. How does this clustering occur during development, and how are the clusters maintained after they are formed? For obvious reasons, both in terms of normal physiology and in terms of the neuropathologic impact in multiple sclerosis, the molecular basis for Nav channel clustering has become one of the most intensely researched areas in the field. As will be evident in the

97

discussion that follows, at least four different mechanisms contribute to Nav channel clustering at the Ranvier nodes: the physical barrier imposed by the paranodal junction, direct anchoring of Nav channels to extracellular and intracellular molecules, linkage to auxiliary ␤ subunits of Nav channels, and diffusible factors. In an elegant study combining Nav staining and electrophysiology in rat sciatic nerves, Vabnick and co-workers127 detected Nav clustering as early as postnatal day 1. As Nav clusters started to mature during myelinogenesis, there was a dramatic increase in conduction velocity up to 10 m/s. Clearly, insulation of leakage current across the internodal axon by myelin ensheathment contributes to the increase in conduction velocity. Nevertheless, the concomitant clustering of nodal Nav channels at high density is absolutely required on pure electrophysiologic grounds to allow the tiny nodal membranes to generate an intense focal Na influx to propel saltatory conduction. By following Nav staining before and after myelination formation in the rat sciatic nerve, Vabnick and co-workers127 found that the early Nav channel clusters appear to be pushed ahead by the elongating edge of Schwann cells as they stretch and occupy the future internodal domain on the growing axons. This picturesque finding suggests that Nav clustering requires physical contact between axons and Schwann cells. However, diffusible factors might also contribute to channel clustering. The first serious evidence for this theory comes from in vitro studies of CNS nerves. Kaplan and colleagues61 found that mammalian retinal ganglion cells, when cultured in the absence of direct physical contact with oligodendrocytes (the myelinating cells in the CNS), could form Nav clusters on axons if oligodendrocyteconditioned medium is present. These two types of clustering mechanism might cofunction and are not necessarily mutually exclusive; nevertheless, the debate on the relative role of diffusible factors and physical contact in Nav channel clustering will likely go on with renewed vigor as these factors are identified in the future and subjected to genetic manipulations. Intracellular cytoskeletal elements and extracellular adhesion molecules also contribute to Nav channel clustering. Nav channels have extensive cytoplasmic domains containing structural motifs that might interact with intracellular anchoring proteins. One key intracellular anchoring protein has been identified as ankyrin G, which interacts with purified Nav proteins in biochemical assays120 and colocalizes with Nav channels at the Ranvier nodes in immunohistochemical studies. A cerebellar-specific gene knockout of ankyrin G greatly reduces Nav channels at the initial segments of granule cells.9,145 Another cytoskeletal protein, spectrin isoform ␤IV, co-localizes with Nav channel clusters.11 Extracellular cell adhesion molecules (CAMs) found on the nodal surface with potential for Nav channel clustering activity include neurofascin and Nr-CAM.4 Extracellular and intracellular anchoring proteins apparently interact.

Function of the Peripheral Nervous System

Na⫹ Channels on Schwann Cells Although the main function of Schwann cells is myelin formation and axon insulation, these cells surprisingly express voltage-gated Na⫹ and K⫹ channels, the hallmark of

excitable cells. This discovery was first made in patch-clamp recordings from Schwann cells in culture (Fig. 4–2A).36 Subsequent patch-clamp studies on acutely isolated Schwann cells, with axons still attached to them, confirmed that Nav channels are normally expressed in Schwann cells in vivo (Fig. 4–2B).25 Interestingly, Nav channel activity is downregulated at the Schwann cell soma during myelin elaboration,25 suggesting that the expression of Nav channels on Schwann cells is related to myelinogenesis.30 It has been suggested that during myelin formation, Nav channels on the Schwann cells are redistributed from the soma to distal processes.28 During wallerian degeneration, dying

A

Cultured Schwann cell

400

pA

For example, ankyrin interacts with the cytoplasmic domains of neurofascin, Nr-CAM, and Nav channels,4,120 and ankyrin G knockout mice exhibit altered neurofascin expression in the initial segments of granule neurons in the cerebellum.145 An intriguing finding is that there is a shift in the ankyrin isoform expression (from ankyrin B to ankyrin G) that parallels a shift in the Nav isoform expression (Nav1.2 to Nav1.6) during myelinogenesis.69 The emerging molecular picture appears to be one in which each Nav channel subtype is tailor-made to fit only a specific subset of anchoring proteins, and that maintaining a stable Nav channel cluster in a mature nerve requires a dynamic deployment of specific anchoring proteins. Once a mature Nav channel cluster is formed, anchoring to axonal cytoskeletal elements appears to be more important than the Schwann cells in maintaining the Nav channel clusters. This conclusion comes from a recent study in which contactin, a paranodal protein involved in tightening the paranodal seal, was genetically deleted from mice. Interestingly, despite loosening of the paranodal seal, which causes a dispersion of the juxtaparanodal potassium channels (see later in this chapter), the nodal Nav channel clusters remain intact, suggesting that the paranodal seal is not very important in preserving Nav channel clusters after they are formed. Besides physical constraint by the paranodal junction and anchorage to an assortment of extracellular and intracellular proteins, auxiliary ␤ subunits (Nav␤1, ␤2, ␤3) of Nav channels also have been suggested to contribute to Nav channel clustering at the Ranvier nodes. Nav␤ subunits are transmembrane proteins with an extracellular immunoglobulinlike domain and a cytoplasmic domain.54–56 The cytoplasmic domain has structural motifs that have been hypothesized to interact with ankyrin, whereas the extracellular domain has been suggested to interact with like molecules from Schwann cells in a homophilic fashion75 or with extracellular adhesion molecules such as tenascins and neurofascin.101 Neurofascin and Nav␤1 are co-localized to Ranvier nodes in rat sciatic nerves,101 and in vitro studies using co-cultures of Schwann cells and dorsal root ganglion neurons show that perturbing the function of Nr-CAM, a CAM expressed at the axon surface, blocks Nav channel and ankyrin accumulation at the nodes of Ranvier.73 Researchers now envision a Nav channel raft using the membrane-spanning Nav␤ auxiliary subunit as an intermediary to coordinate extracellular and intracellular interactions with other anchoring elements.73,75 Gene knockouts of Nav␤ will undoubtedly be forthcoming to shed light on the role of ␤ auxiliary subunits on Nav channel clustering.

200 0 -200

5 msec

Acutely isolated Schwann cell

B 200

ionic current (pA)

98

0

Na+ current -300 5 msec

FIGURE 4–2 Na⫹ and K⫹ channels in mammalian Schwann cells. A, Wholecell ionic currents recorded from a single cultured Schwann cell showing Na⫹ currents (downward) and K⫹ currents (upward). (From Chiu, S. Y., Schrager, P., and Ritchie, J. M.: Neuronal-type Na⫹ and K⫹ channels in rabbit cultured Schwann cells. Nature 311:156, 1984, with permission.) B, Na⫹ currents recorded from a single, acutely isolated mammalian Schwann cell. The picture shows the glass patch-pipette used to record the currents from the cell soma. (From Chiu, S. Y.: Sodium currents in axonassociated Schwann cells from adult rabbits. J. Physiol. [Lond.] 386:181, 1987, with permission.)

Channel Function in Mammalian Axons and Support Cells

axons apparently provide a signal to Schwann cells to reexpress Nav channels at the soma.26 Intriguingly, messenger RNA (mRNA) for Nav␤1 appears to be present in Schwann cells of normal myelinated fibers.87 It is unclear where on the myelinating Schwann cell the Nav␤1 protein is expressed, an issue that could be of particular interest given the hypothesis that homophilic interactions between the extracellular domain of Nav␤1 from axons and Schwann cells might play a role in Nav channel clustering at the Ranvier nodes.75 What are the functions of Nav channels on Schwann cells? Their low density appears to preclude Schwann cells from generating action potentials.25,36,113 One possible function is that Nav channels in Schwann cells act to recruit the auxiliary Nav␤1 subunits to the Schwann cell surface where, as discussed previously, they could provide homophilic interactions with axonal Nav␤1 to stabilize Nav channel clusters on axons. The viability of this hypothesis clearly awaits definitive immunohistochemical staining of Nav␤1 on mature Schwann cells to see if it co-localizes and apposes to Nav␤1 expressed on the axons. In a different type of functional speculation, Chiu and colleagues proposed that Schwann cells synthesize Nav channels for axons.36 With the identification of Nav channel subtypes through gene cloning, this hypothesis can be examined by looking for a common set of Nav isoforms expressed in Schwann cells and axons. Teleologically, a transfer of Nav channels from Schwann cells to axons would relieve the biosynthetic load of the neurons.105,118 Finally, Sontheimer and colleagues have advanced the theory that the sodium influx through glial cell Nav channels can be used to fuel potassium uptake through the glial Na⫹,K⫹-ATPase, thereby contributing to potassium buffering.118 If Schwann cell Nav channels are activated during axonal activity, Sontheimer et al.’s model allows an activity-dependent augmentation of potassium uptake through the Schwann cell Na⫹,K⫹-ATPase that would be matched to nerve activity.

K⫹ CHANNELS In contrast to Nav channels, whose primary role is to mediate the rapid upstroke of an action potential, voltagegated potassium channels (Kv) play an important role in the modulation of excitability such as frequency coding. In their pioneering work on excitability of the nonmyelinated squid giant axon, Hodgkin and Huxley52 concluded that Kv channels play a major role in terminating the action potential by providing a delayed potassium efflux to repolarize the membrane. The function of Kv channels in myelinated axons was addressed only when a new voltageclamp technique was invented for studying single myelinated fibers.41,86 The first studies with these new voltage-clamp techniques showed that Kv channels are

99

present in amphibian nodes of Ranvier and play a central role in action potential repolarization (Fig. 4–3). Surprisingly, subsequent voltage-clamp studies on single mammalian myelinated axons from peripheral nerves, including human nerves, showed that the nodal membrane of these fibers lacks Kv channels.17,34,53,110 Action potential repolarization in the mammalian PNS, in contrast to the amphibian PNS myelinated nerves, relies only on Nav channel inactivation and passive leakage current (Fig. 4–3).34 Are Kv channels absent from mammalian myelinated nerves, or are they present in extranodal regions? The first clue for the latter possibility came from a study111 showing that Kv channels are present on chronically demyelinated axons. Chiu and Ritchie31 then showed in acute paranodal demyelination studies that Kv channels are present normally under the myelin sheath (see Fig. 4–1A). By early 1980, the conclusion was established that Kv channels are segregated in a fashion opposite to that of Nav channels, in that Kv channels are absent in the node but present under the myelin (see Fig. 4–1A). This was followed by a period of extensive electrophysiologic characterization of mammalian internodal potassium channels. Baker and colleagues6 defined the presence, as well as the probable function, of three types of rectifying channel (one sensitive to 4-aminopyridine [4-AP], one to tetraethylammonium [TEA], and the third to cesium) under intact myelin in rat spinal root myelinated axons. A major breakthrough in electrophysiologic analysis of axonal channels occurred when highly skilled investigators demonstrated that they can acutely peel away the myelin and subject the naked but tiny axon to patch-clamp recordings.58 These studies identified up to eight different types of axonal potassium channels, including Kv channels, leakage K⫹ channels that show weak voltage dependence, and K⫹ channels sensitive to intracellular metabolic chemistry. Readers interested in the electrophysiologic classification of axonal K⫹ channels should consult several excellent reviews on this subject.130,133 Our main focus here is on the molecular study of axonal Kv channels, which has seen a quantum leap in new knowledge fueled by the pace-setting application of molecular biology and transgenic techniques.

Molecular Identity of Axonal Kv Channels: A Tripartite of Kv1.1/Kv1.2/Kv␤2 Cloning of K⫹ channel genes from mammalian brain has identified three large families of K⫹ channels classified structurally according to the number of transmembrane domains (six, four, or two) in each ␣ subunit.57 Of the three families, the one with six transmembrane regions constitutes the largest family and is the primary focus of this chapter. Within this family, at least 16 subfamilies have been identified based on similarity in the primary amino acid sequence. Most, but not all, of the K⫹ channels in these 16 subfamilies

100

Function of the Peripheral Nervous System

FIGURE 4–3 Macroscopic ionic currents (top) and action potentials (bottom) in nodes of Ranvier. Top, Na⫹ currents are displayed as downward deflections and K⫹ currents as upward deflections. Note that K⫹ current is present in the amphibian (right) but absent in the mammal (left). Bottom, Computer reconstruction of the mammalian action potential based only on Na⫹ and leakage currents: empirical data (left) and computer reconstruction (right). Repolarization is accounted for by Na⫹ channel inactivation and passive leakage current. Unlike the amphibian myelinated nodes, voltage-gated K⫹ channels are not required. (From Chiu, S. Y., Ritchie, J. M., Rogart, R. B., and Stagg, D.: A quantitative description of membrane currents in rabbit myelinated nerve. J. Physiol. [Lond.] 292:149, 1979, with permission.)

are voltage gated (Kv), and each subfamily consists of numerous members. For example, the Kv1 subfamily (also called Shaker subfamily) has up to nine members, Kv1.1 through Kv1.9. Each functional K⫹ channel is composed of four ␣ subunits, and only members of the same subfamily can co-assemble.70 This tetramultimeric nature of K⫹ channels provides great functional diversity through mixing and matching different members of the same subfamily to form potassium channels with unique properties. Auxiliary ␤ subunits (Kv␤s) further diversify Kv channel function.96,125 Three Kv␤ subfamilies have been identified (Kv␤1, Kv␤2, and Kv␤3) and various functions of the Kv␤ subunits have been proposed. Some of these cytoplasmic Kv␤ subunits can induce Kv␣ channel inactivation, leading to the hypothesis that Kv␤ subunit association modulates frequency coding in excitable cells.96 Kv␤ subunits have also been

suggested to be chaperones to facilitate surface expression of Kv␣ subunits.112 Intriguingly, Kv␤ belongs to the aldoketoreductase (AKR) family,40,79 which has led to the hypothesis that the association of Kv␤ with Kv␣ subunits allows excitability of cells to be modulated by intracellular oxidation-reduction chemistry. As for Nav channels, axonal Kv channels in the mammalian peripheral nerves do not form a homogeneous pool. Using isoform-specific antibodies in immunohistochemistry, the most conspicuous landmark is a tripartite Kv1.1/Kv1.2/Kv␤2 clustered at the juxtaparanodes of mature myelinated axons (see Fig. 4–1C),98,100,103,132 a finding confirming the conclusion of earlier electrophysiologic studies regarding Kv channel localization under the myelin.31 It should be noted that this immunohistochemical survey might not be exhaustive; given the diverse population

101

Channel Function in Mammalian Axons and Support Cells

of K channels characterized in patch-clamp studies, other molecular species might well be discovered in the future. Nevertheless, the discovery of the Kv1.1/Kv1.2/Kv␤2 juxtaparanodal tripartite in mammalian peripheral axons has directed research along two major lines of questions. First, what maintains the Kv1.1/Kv1.2/Kv␤2 tripartite in the juxtaparanode of a mature myelinated axon? Second, what is the function of the juxtaparanodal Kv tripartite?

Mechanisms for Clustering Kv Channels During development of the sciatic nerves, Kv1.1/Kv1.2/ Kv␤2 clusters first appear at the nodal membrane, then quickly migrate toward the paranode before stabilizing in the juxtaparanode in a mature myelinated axon,128 an observation that accounts for the gradual diminution of sensitivity of sciatic nerves to Kv channel blockers (TEA and 4-AP) during development.63 What maintains the Kv channel clusters at the juxtaparanode in a mature nerve? As in the case of Nav channels, a physical barrier imposed by the paranodal junction and anchorage to an assortment of proteins have been shown to play a role. The role of Schwann cells in organization of axonal Kv channels has been investigated using various PNS dysmyelinating mutants, including the Trembler mice. Axonal Kv1.1 channels are desegregated in the Trembler dysmyelinated nerves,131 suggesting that the Schwann cell–axon junction

at the paranode is important for maintaining channel clustering. Recently, the hypothesis regarding the paranodal junction as a key factor in positioning the Kv channel clusters has been definitively confirmed by molecular biology studies in which the nuts and bolts of the protein assembly at the junction have been identified and selectively deleted in knockout mice.13,16,91 At least three assembly proteins have been identified that pull the paranodal Schwann cell loops toward the axon at the paranode to form a tight junction. They are Caspr1 and Contactin (which form a functional complex) from the axonal membrane,16,44,80 which interlock with neurofascin (NF)-155 protruding from the Schwann cell loops122 (Fig. 4–4C).16 After the contactin gene is genetically inactivated,16 Caspr expression is abolished from the paranodes. The NF-155 from the Schwann cell fails to interlock to the axons, which causes the Schwann cell loops to recede slightly from the axon, resulting in loosening of the paranodal junction and slowing of the nerve conduction (Fig. 4–4A and B). Kv1.1 is found to migrate from the juxtaparanode to the paranode and stops next to the nodal Nav channels, whose cluster is not affected by contactin deletion (see Fig. 4–4C). Anchoring proteins also have been suggested to contribute to Kv1 channel clustering. Recently, Kv1 channels in the juxtaparanode were found to co-localize and coimmunoprecipate with Caspr2, another member of the neurexin superfamily to which Caspr1 belongs.95 The large

C

A

wild type

rec

stim

NF-155 Caspr +-

distance

- + Na+ - channels

B

juxtaparanode

paranode

node

juxtaparanode

paranode

node

Contactin

mutant + /+ CNTN

1 msec

CNTN

-/-

FIGURE 4–4 Neuropathology of genetic ablation of the paranodal adhesion protein contactin. A, Experimental configuration for measurement of conduction of compound action potential in sciatic nerves of mice. B, Slowing of nerve conduction in the contactin knockout (bottom) compared with the wild type (top). C, Model (top) showing the three assembly proteins (NF-155 from Schwann cells and Caspr and Contactin from the axon) in a normal paranode. Deletion of contactin (bottom) results in lifting of the Schwann cell paranodal loops and migration of Kv channels from the juxtaparanode to the paranode. (From Boyle, M. E., Berglund, E. O., Murai, K. K., et al.: Contactin orchestrates assembly of the septate-like junctions at the paranode in myelinated peripheral nerve. Neuron 30:385, 2001, with permission.)

102

Function of the Peripheral Nervous System

extracellular domain of Caspr2 might interact with binding partners from Schwann cells, helping to regulate Kv1 channel localization to not only the juxtaparanodes but along the inner mesaxon as well.5 Binding to various intracellular proteins, as in the case of Nav channels, might also determine Kv1 localization. For example, there is a PDZbinding domain in the cytoplasmic C-terminus of the Kv1 channel that could act as an anchoring site for other cytoplasmic PDZ-containing elements such as PSD-95.95 As discussed in the case of Nav channel clustering, cytoplasmic ankyrin and spectrin ␤IV appear to anchor Kv channels. The role of spectrin ␤IV in channel organization has been clarified by the autosomal recessive mutation quivering, a spontaneous mutation in mice known since 1953 to produce ataxia, hind limb paralysis, and deafness. Recently, the quivering mutation has been traced to a lossof-function mutation in spectrin ␤IV, which interestingly correlates with an alteration of Kv1 channel localization in myelinated axons of quivering mice.90 Does Kv␤2 promote trafficking and appropriate localization of the Kv1.1/Kv1.2/Kv␤2 tripartite in myelinated axons? As for Nav channels, auxiliary subunits have also been proposed to organize axonal localization of Kv channels. Kv␤2 binds to K⫹ channel ␣ subunits and is a member of the AKR superfamily. However, the Kv␤ subunits, unlike the Nav␤ subunits, are strictly cytoplasmic proteins with no transmembrane domains, a feature that precludes extracellular homophilic interactions with apposing Schwann cell Kv␤s as a mechanism for clustering Kv␣ subunits on axons. Proposed functions for Kv␤s include a chaperone-like role in Kv1 ␣-subunit biogenesis and catalytic activity as an AKR oxidoreductase. In cultured mammalian cells, the absence of Kv␤2 results in unglycosylated forms of Kv1.1 and Kv1.2 that are not efficiently transported to the membrane surface, leading to the hypothesis112 that Kv␤2 facilitates the glycosylation of Kv1␣ subunits in the endoplasmic reticulum (ER), thereby promoting the trafficking of Kv␣/␤ complexes from the ER to the surface cell membrane. Furthermore, in cultured hypocampal neurons, Kv1.2 was appropriately targeted to axons only when cotransfected with Kv␤2.22 Whether Kv␤2 is important for Kv1.1/Kv1.2 clustering in vivo has recently been tested by the generation of a Kv␤2 knockout mouse.78 These mutant mice have a shortened lifespan and a cold swim–induced myotonia indicating hyperexcitability. Surprisingly, Kv1.1/Kv1.2 clusters form normally in the juxtaparanodes of mature PNS myelinated axons, suggesting normal biogenesis and trafficking of Kv1␣s in the absence of Kv␤2.78 The differences between the in vivo and in vitro studies might be reconciled if, for example, other Kv␤ subunits (such as Kv␤1) were upregulated in the Kv␤2-null mice to compensate for the loss of Kv␤2. However, PNS myelinated axons normally do not express Kv␤1, and Western blot analysis of whole-brain lysates revealed no significant change in the Kv␤1 protein

expression in the Kv␤2-null mice.78 Alternatively, anchorage to cytoskeletal elements such as spectrin ␤IV might compensate for the loss of Kv␤2. Finally, other yet-to-be identified Kv1␣ subunits in axons could facilitate efficient surface expression of heteromultimeric Kv1 channels at the juxtaparanodes in the absence of Kv␤2.77 It is likely that multiple binding partners coordinate the juxtaparanodal clustering of Kv1 channels with a high degree of redundancy, and simultaneous genetic inactivation of multiple binding partners might be needed to shed light on the role of protein-protein interactions in Kv1 channel localization. Figure 4–5 summarizes the evolution of our knowledge on the organization of ion channels in myelinated axons from the early 1960s (top) to 2002 (bottom).

Function of Kv Channels Physiologic Functions Having reviewed the mechanisms for Kv1 channel clustering, we next turn to the physiologic and pathophysiologic functions of Kv1 channels in PNS myelinated axons. During development, the transient presence of Kv1 channels on the nodal membrane has been suggested to prevent abnormal excitation in developing myelinating nerves.128 However, what might the functions of Kv1 channels be after they have been fully sequestrated in the juxtaparanode, which is thought to be sealed off from the nodal membrane by the paranodal junction108? In the early 1980s, it was proposed that a mature paranodal junction has significant finite leakage across it, allowing electrotonic interactions between the nodal and the internodal axonal membrane.8,33 As a consequence, it has been proposed that nodal excitation can evoke a delayed, secondary depolarization of the juxtaparanodal membrane, which when amplified by a trace amount of internodal Nav channels,37 can lead to reentrant excitation of the node.32 The juxtaparanodal Kv1 channels have been suggested to dampen this putative reentrant excitation, preserving faithful electrical transmission along the axons. This hypothesis has been difficult to test because of the difficulty in experimentally manipulating the juxtaparanodal Kv1 channels, which are normally wrapped by the myelin. However, a rare human neurologic disorder, episodic ataxia type 1 (EA-1), provides clues to the function of Kv1 channels in myelinated axons. Episodic ataxia is characterized by stress-induced hyperexcitability; EA-1 patients appear normal but intermittently display a protracted phase of ataxia and tremors brought on by stress, including startle and exercise.19 The myokymia component in the clinical manifestations suggests PNS hyperexcitability. Recently, EA-1 has been traced to various missense and nonsense mutations of the human Kv1.1 gene, which result in reduced levels of functional channels, abnormal trafficking, and intracellular aggregation of mutated Kv1.1 proteins.1,19,76,143 Genetic ablation of Kv1.1 in mice was therefore used to assess Kv1.1 functions in vivo.114 Kv1.1-null mice display

103

Channel Function in Mammalian Axons and Support Cells

Node of Ranvier

Initial segment Axon hillock

Dendrite

Na+ Myelin

+

+ -

Axon

+ Myelin

+

-

Na+

microvilli

internode

basal lamina

incisure

juxtapara para

Kv1.1 Kv1.2 Caspr Kvβ2 Caspr2

node

Nav1.6 ankyrinG

PNS CNS

Perinodal astrocyte Compact Myelin

Adaxonal Membrane Microvilli or Perinodal Astrocyte

15 5 Co

nt a cti n

NF

sp r Ca

Kvβ2

Tenascin-R β2

Kv1.1 Kv1.2

NF186

NF

18 6

Paranodal Myelin Loop

Tenascin-C

Caspr-2

K+ Channels

Nr-CAM

FIGURE 4–5 Summary of the evolution of our knowledge on ion channel localization in the mammalian myelinated nerves from the 1950s to 2002. Top, Early view (the 1950s) of myelinated nerves. Action potentials are generated in the initial segment of the neuron and propagate in a saltatory fashion from node to node. The myelin sheath at the internode prevents capacitative and resistive loss, allowing highspeed nerve conduction. (From Peles, E., and Salzer, J. L.: Molecular domains of myelinated axons. Curr. Opin. Neurobiol. 10:558, 2000, with permission.) Bottom, Current view (2002) of myelinated nerves. Molecular biology and gene cloning have identified the ion channel species, clarified their localization, and identified many anchoring proteins involved in organizing the ion channels. (Courtesy of Steve Scherer and Arroyo Edgardo.)

Band 4.1B PDZ?

Spectrin βIVS1 AnkyrinG Na+Channel

104

Function of the Peripheral Nervous System

epilepsy and ataxia resulting from deleting Kv1.1 expression in the CNS.114,144 What is the physiologic impact of deleting Kv1.1 on PNS myelinated axons? First, there is a slight increase in the refractory period of nerve transmission in the adult sciatic nerves,114 suggesting that juxtaparanodal Kv1.1 contributes to electrogenesis. Second, the Kv1.1-null mice display cold swim–induced myokymia, which hints at a stress-inducible hyperexcitability of the neuromuscular junction. Immunohistochemical, electrophysiologic, and theoretical analysis of the neuromuscular transmission in the mutants revealed a temperaturedependent repetitive discharge from a short myelinated segment just preceding the nerve terminal.147,148 The hypothesis is that a preshortening of the internodes prior to the nerve terminal, while needed to facilitate safe invasion of the nerve terminal by action potentials,97,102,134 also increases the electrotonic coupling between the nodal and internodal axonal membrane, which is inherently destabilizing. This juxtaparanodal Kv1.1 is theorized to stabilize the preterminal, myelinated segment of the axon from reentrant excitation.147,148 This proposed action of juxtaparanodal Kv1.1 could be of general applicability to branch points. Computer simulations have shown that an inappropriate shortening of internodes prior to a branch point, while providing safe conduction across the branch point, also creates inherent instability in the pre–branch point segments.146 Juxtaparanodal Kv1.1 might act to prevent reentrant excitation in myelinated segments in the immediate vicinity of branch points, preserving fidelity of electrical signaling in axonal trees. It should be pointed out that, because Kv1.2 channels still remain in the juxtaparanodes in the Kv1.1-null mutants, a double knockout of Kv1.1 and Kv1.2 might be needed to assess fully the physiologic role of juxtaparanodal Kv1 channels. Pathophysiologic Functions Kv1 channels in myelinated axons are highly relevant in demyelinating diseases. Once exposed by demyelination, the internodal Kv1 channels act like a brake to retard conduction across the lesion. Although the scarcity of internodal Nav channels coupled with impedance mismatch134 contributes to conduction failures in demyelination, neutralizing the braking action of Kv channels by pharmacologic intervention has been explored as a means to restore conduction in demyelination. Even a slight migration of Kv1 channels from the juxtaparanode to the paranode, as occurs in the contactin knockout mice (see above), profoundly affects nerve conduction in myelinated fibers.16 Normally, Nav and Kv1 channels are physically delineated by a 2- to 4-␮m paranodal seal. Studies of the contactin knockout mice show that inappropriate encroachment of Kv1 channels on Nav channels consequent on a slight loosening of the paranodal pockets may be incompatible with normal nerve function.16 Schauf and Davis109 first pointed out on electrophysiologic grounds that blockage of Kv1

channels with 4-AP should increase the safety factor of conduction in demyelinated axons. This idea was first confirmed in animal models of demyelination,15,123 then followed by clinical trials121 showing that oral ingestion of 4-AP transiently improves neurologic functions in patients with multiple sclerosis over those given placebo. The use of K⫹ channel blockers in clinical trials in patients with multiple sclerosis has been reviewed.12,83,121 Other reviews on the role of Kv channels in electrogenesis of myelinated axons can be found elsewhere.39,135 The identification of the juxtaparanodal tripartite Kv1.1/Kv1.2/Kv␤2 should provide a more specific target for therapeutic design in pharmacologic intervention strategies in multiple sclerosis.

K⫹ Channels on Schwann Cells Kv channels were first described in mammalian Schwann cells in culture conditions using patch-clamp techniques (see Fig. 4–2A).36 Subsequent patch-clamp studies from acutely isolated Schwann cells demonstrate that Kv channel expression is not a tissue culture artifact.25 Electrophysiologic studies have distinguished various types of Kv channels in Schwann cells according to kinetics and pharmacology.2,7,65,129 Furthermore, inward rectifier K⫹ channels (Kir) have also been described in mammalian Schwann cells.66,67,138,139 Inward rectifiers, unlike Kv, pass inward currents more readily than outward currents, as a result of a cytoplasmic blocking particle (Mg2⫹ or polyamines such as spermine and spermidine) that gets swept into the pore when the cell is depolarized, thereby obstructing the efflux of K⫹ ions. With the presence of Kv and Kir channels firmly established in Schwann cells, research has focused on three broad issues: the relationship between Schwann cell K⫹ channel expression and myelinogenesis, the molecular identity of Schwann cell K⫹ channels, and the function of Schwann cell K⫹ channels. Expression of Schwann Cell K⫹ Channels Linked to Myelinogenesis Patch-clamp studies from Schwann cells acutely isolated from sciatic nerves at different stages of myelination have shown that Kv and Kir channel activity is downregulated at the Schwann cell soma as myelin is elaborated (Fig. 4–6, left).139 Interestingly, in an adult myelinated fiber, both Kv and Kir channel activity can be detected at Schwann cell processes near the node of Ranvier,138 leading to the suggestion of a developmental redistribution of both channels to the Schwann cell microvilli during myelin formation. This pattern of Schwann cell channel expression during normal myelination is reversed when axons are transected in mature myelinated nerve fibers26; as the axons and myelin sheath degenerate during wallerian degeneration, Nav and Kv channel activity reappears on the soma of Schwann cells (Fig. 4–6, right).26 It is likely that a combination of trophic signals both from axons and the myelin

Channel Function in Mammalian Axons and Support Cells

105

FIGURE 4–6 Relation between K⫹ channel activity and proliferation in mammalian Schwann cells. Left, Reduction in Schwann cell proliferation during normal myelinogenesis is paralleled by a reduction in K⫹ channel activity at the Schwann cell soma. (From Wilson, G. F., and Chiu, S. Y.: Potassium channel regulation in Schwann cells during early developmental myelinogenesis. J. Neurosci. 10:1615, 1990, with permission.) Right, Increase in Schwann cell proliferation during wallerian degeneration is paralleled by an increase in K⫹ channel activity at the Schwann cell soma. (From Chiu, S. Y.: Changes in excitable membrane properties in Schwann cells of adult rabbit sciatic nerves following nerve transection. J. Physiol. [Lond.] 396:173, 1988, with permission.)

sheath act in combination to program channel expression in the Schwann cells. Molecular Identity and Localization of Schwann Cell Kv and Kir Channels The first clue to the molecular identity of Kv channels in Schwann cells came from mRNA analysis of sciatic nerves. Because axons are commonly thought to lack transcripts, mRNA extracted from sciatic nerves probably has a predominant Schwann cell origin. Kv1.1 and Kv1.2 mRNAs have been detected in mammalian sciatic nerves.35 Sobko and colleagues116 carried out a more exhaustive analysis of Kv channel identity in Schwann cells, at the protein level, in Schwann cells cultured from neonatal mouse sciatic nerves. Under these conditions, Schwann cells express four members of the Kv1 subfamily (Kv1.1, Kv1.2, Kv1.4, and Kv1.5), one member of the Kv2 subfamily (Kv2.1), and two members of the Kv3 subfamily (Kv3.1 and Kv3.2). These K⫹ channels are also detected in Western blots of sciatic nerves; however, an axonal contribution to the signal in the Western blots cannot be ruled out.116 Electron microscopic immunohistochemistry shows that Kv1.1 and Kv1.5 are differentially localized in a mature myelinating Schwann cell; Kv1.5 is localized to the Schwann cell membrane near the nodes of Ranvier and in bands that run along the outer surface of the myelin, whereas Kv1.1 is present predominantly in the perinuclear regions and in intracellular compartments.81 Perinuclear location of Kv1.1 proteins in Schwann cells has also been

observed in remyelinating sciatic nerves following lysolecithin-induced demyelination.98 The identity of Kir channels has also been clarified in Schwann cells with the identification of Kir2.1 and Kir2.3 on the Schwann cell microvilli facing the nodal surface.81 Thus, as in axons, K⫹ channel localization in a fully differentiated Schwann cell is highly polarized. What are the molecular mechanisms responsible for the differential localization of K⫹ channels in myelinating Schwann cells? Of particular interest is Kv1.1. This channel in neurons is transported away from the cell soma (dorsal root ganglion) and inserted on the axon at the juxtaparanode, which can be contrasted with the situation in a mature myelinating Schwann cell, in which Kv1.1 is trapped in intracellular compartments and apparently fails to be transported to the cell surface. It is possible that the normal absence of Kv␤2 in Schwann cells98 fails to provide a chaperone for surface expression of Kv1.1 in Schwann cells. However, Kv1.5 shows efficient surface expression in Schwann cells81 in the absence of Kv␤2. It is known from transfection studies using cultured mammalian cell lines that different Kv␣ subunits have different surface expression efficiency77: the surface expression efficiency is good for Kv1.2 homotetramers but poor for Kv1.1 homotetramers, the latter being trapped in the ER. Interestingly, in neonatal sciatic nerves, most of the Kv1.5 subunits are involved in heteromulteric association with Kv1.2,116 a finding that might explain why Kv1.5 shows efficient surface expression on Schwann cells.81 It appears that Kv1.1

106

Function of the Peripheral Nervous System

subunits might exist in Schwann cells as homotetramers, which are inherently poorly transported and trapped in the intracellular compartments.81 Here, the comparison with Kv1.1 trafficking in axons is interesting. As discussed previously, Kv1.1 trafficking to the juxtaparanodes is not affected by genetic deletion of Kv␤2 proteins.78 Perhaps the efficient surface expression of Kv1.1 in the Kv␤2-null axons is the result of its association with Kv1.2, an association that might be lacking in Schwann cells. Functions of K⫹ Channels in Schwann Cells Three hypotheses concerning K⫹ channel functions in Schwann cells have been proposed. The first is that K⫹ channels are synthesized in Schwann cells and transferred to axons to relieve synthetic burden on the neurons. The second is that K⫹ channels in Schwann cells contribute to potassium buffering. The third is that K⫹ channels in Schwann cells modulate Schwann cell proliferation. Schwann Cell–to-Axon Transfer of Kv Channels. With regard to the first hypothesis of a Schwann cell–toaxon transfer of K⫹ channels, a minimum requirement for the hypothesis has been satisfied by the demonstration that Kv1.1/Kv1.2, two members of the axonal juxtaparanodal tripartite (Kv1.1/Kv1.2/Kv␤2), are expressed in Schwann cells.35,116 However, the present difficulty with this hypothesis, at least for Kv1.1, is that this protein appears to be retained in the intracellular compartments of Schwann cells. In contrast, the vesicular form of Kv1.1 proteins in Schwann cells might reflect proteins that are destined for transfer to axons; this transfer, if it exists, could be below our current detection limit. Kv1.2, however, should be expressed efficiently on the Schwann cell surface, but immunohistochemical localization of Kv1.2 in myelinating Schwann cells has not been examined. In squid giant axons, a bilateral transfer of large proteins between Schwann cells and axons has been demonstrated.50,126 Our laboratory, in collaboration with Albee Messing and Larry Wrabetz, is currently exploring transgenic techniques45 to test the transfer hypothesis by selectively expressing Kv1.1 in Schwann cells of the Kv1.1 global knockout mouse to see if juxtaparanodal expression of Kv1.1 can be restored. Potassium Buffering. The level of endoneurial fluid potassium is known to have a profound effect on the excitability of peripheral nerves; depending on the extracellular potassium level, excitability of nerves can be either enhanced or depressed. Normal nerve activity, as well as damaged tissues, could be sources of extracellular potassium. Abnormal elevation of endoneurial potassium could be responsible for positive symptoms such as paresthesias and pain in peripheral neuropathies caused by spontaneous firing of axons,124 and an elevation of endoneurial potassium concentration has been demonstrated in demyelinated and regenerating nerves.72 Glial

cells have traditionally been suggested to play a role in potassium buffering.68,88 Transport systems such as Na⫹,K⫹-ATPase on glial cells can transport potassium ions into glial cells and lower the extracellular potassium level. In addition, the various Kv and Kir channels described on Schwann cells could make an important contribution to potassium buffering. In the peripheral nerves, activitydependent release of potassium is spatially homogeneous in neonates before myelin formation and becomes highly localized to the nodal region once myelin is formed. The expression of Kv and Kir channels in Schwann cells appears to be functionally matched to the developmental change in potassium release. In neonates, activitydependent release of axonal K⫹ is large because of the participation of the entire axonal surface in electrogenesis; correspondingly, Schwann cell expression of Kir and Kv channels is large in neonates. As myelination proceeds, Schwann cell expression of both K⫹ channels is reduced, consistent with a marked reduction in K⫹ released per unit axonal length as the myelin sheath is deposited on the axons. However, theoretical calculations have shown that the intense and highly focal ionic flux at the nodal regions, coupled with a restricted extracellular volume at the nodalparanodal-juxtaparanodal compartments, could produce significant local accumulation of K⫹ ions.28 The Schwann cell microvilli facing the nodal membrane, with a large surface-to-volume ratio, are an ideal system for absorbing K⫹ ions released during activity (note that, even though Kv channels are absent at the node, the leakage channels that help repolarize an action potential are likely potassium channels that mediate K⫹ efflux during the repolarization phase). Once inside the Schwann cells, the K⫹ ions could flow down the concentration gradient and exit elsewhere, away from the node. The Kir2.1 and Kir2.3 channels on the microvilli have been hypothesized to mediate the nodal absorption, whereas the Kv1.5 channels located on the outer myelin surface have been suggested to divert the absorbed K⫹ ions and release them away from the node.81,138 Testing this model with global genetic ablation of Kir2.1 is hampered by lethality in mutant mice at 5 to 10 hours after birth because of cleft palate.142 Functional Linkage between Ion Channels and Proliferation. Insights into channel function in Schwann cells can be gained by examining channel expression at various stages of the life cycle of a Schwann cell. Myelinogenesis and proliferation, two of the best studied features of Schwann cell biology, undergo characteristic changes during Schwann cell–axon interactions. During the first postnatal week, there is a wave of Schwann cell proliferation as cells contact growing axons.18,136,141 This initial proliferation soon subsides and gives way to myelinogenesis as cells are committed to a 1:1 relation with axons destined to become myelinated. As described in previous sections, patch-clamp analyses have shown that

Channel Function in Mammalian Axons and Support Cells

the Kv and Kir channels cease to be active on the Schwann cell soma as the cells leave the proliferation phase and enter the myelinogenic phase. The parallel reduction in K⫹ channel activity and Schwann cell proliferation suggests a functional linkage between the two (see Fig. 4–6, left). This functional linkage was explored in a study in which quiescent, adult Schwann cells were driven into proliferation by nerve transection.26 Here again, the K⫹ channel activity and proliferation increase in parallel (see Fig. 4–6, right). Blocking Kv channels with quinine, TEA, and 4-AP also, in most cases, reversibly blocks Schwann cell proliferation.38 Finally, application of Schwann cell mitogens (glial growth factor and myelin and axolemmal fragments) to cultured Schwann cells enhances Kv channel activity.140 Taken together, these studies suggest that potassium channel activity is an important early signal that modulates the transition of Schwann cells from a proliferative to a myelinogenic phase. The mechanism underlying this signal transduction remains unclear. One possibility is that Kv channel activity alters the resting potential of the Schwann cell,38 which produces ion fluxes that alter the intracellular concentration of second messengers such as sodium and calcium, which are thought to be important for mitosis.117 Since the discovery that K⫹ channels are linked to Schwann cell proliferation, research has focused on identifying the molecular species of K⫹ channels responsible for modulating Schwann cell proliferation. As discussed in previous sections, molecular biology has identified at least seven Kv channels and two Kir channels in Schwann cells.35,81,82,116 Which of these K⫹ channel subtypes are key modulators of Schwann cell proliferation? One approach is to examine the antiproliferative effects of K⫹ channel blockers specific for certain K⫹ channel subtypes. Besides the broad-spectrum K⫹ channel blockers such as TEA and 4-AP, various venom toxins (e.g., the dendrotoxin family) have been found to block certain K⫹ channel subtypes specifically. Based on the lack of an antiproliferative effect of dendrotoxin-I116 and dendrotoxin-␣,89 Kv1.1 and Kv1.2 have been excluded as possible candidates. Of the remaining Kv channels, Kv2.1 and Kv1.5 are strong contenders as important modulators of Schwann cell proliferation. The evidence comes from both CNS and PNS studies. In the CNS, astrocyte proliferation is also linked functionally to Kv channel activity. Using an antisense oligodeoxynucleotide technique, MacFarlane and Sontheimer74 found that Kv1.5 is the major Kv channel expressed in cultured astrocytes. Furthermore, inhibition of Kv1.5 protein expression by antisense oligodeoxynucleotides significantly inhibits astrocyte proliferation.74 The next interesting question that MacFarlane and Sontheimer74 addressed is whether the reduction in proliferation accompanying normal astrocyte differentiation is due to a reduction in protein expression level for Kv1.5 or a reduction in the activity of an existing pool of Kv1.5

107

channels. The correct picture that is emerging appears to be the latter one. How might the activity of existing Kv1.5 channels be modulated without a change in the protein expression level? MacFarlane and Sontheimer74 identified the key to be channel phosphorylation; phosphorylation of the Kv1.5 channel increases channel activity and increases proliferation, whereas dephosphorylation of the channel decreases its activity and decreases proliferation. What, then, controls phosphorylation of Kv channels in glial cells? Immunoprecipitation shows that Kv1.5 is associated with the Src family of protein tyrosine kinases, and this association remains unaltered during astrocyte proliferation.74 Similarly in Schwann cells, Kv1.5 and Kv2.1 have been shown to be physically associated with a Src family tyrosine kinase, which is probably Fyn.115 Indeed, in patchclamp studies of cultured Schwann cells, pharmacologic inhibition of tyrosine kinases inhibits Kv channel activity, and artificial activation of tyrosine kinases increases Kv channel activity.93,115 Taken together, the emerging picture is that certain Kv channels (such as Kv1.5/Kv2.1) and certain Src tyrosine kinases (such as Fyn) coexist physically as a modulatory complex in Schwann cells, that these Kv channels are constitutively phosphorylated by Src tyrosine kinase, and that an understanding of the endogenous factors that regulate Src tyrosine activity might be the key to understand how Kv channel activity modulates Schwann cell proliferation during development. It is important to emphasize that this model for linking K⫹ channels to Schwann cell proliferation rests mostly on tissue culture and explant studies. The central question that remains unanswered is whether Kv channels modulate Schwann cell proliferation in the more complex setting of a whole animal. Nevertheless, two intriguing in vivo studies using global gene knockouts suggest that the in vitro model might be applicable in vivo. First, Peretz and co-workers92 examined whether a particular tyrosine phosphatase, protein tyrosine phosphatase ␧ (PTP␧), might be an in vivo factor regulating phosphorylation of the Kv1.5/Kv2.1/Src complex in Schwann cells. In general, tyrosine phosphatases counter the action of Src tyrosine protein kinases, causing dephosphorylation of Kv1.5 and reducing its channel activity. Consistent with this, global genetic ablation of PTP␧ in mice leads to hyperphosphorylation of Kv1.5 and Kv2.1 proteins and enhancement of Kv channel activity in Schwann cells.92 More importantly, this increased Kv1.5/Kv2.1 channel activity presumably prolongs the proliferative phase of Schwann cells, thereby delaying their entrance into the myelinogenic phase and causing a transient hypomyelination in the PTP␧-null sciatic nerves shortly after birth.92 Interestingly, the PTP␧null mice exhibit normal myelination as adults, suggesting that other PTP species might be present to downregulate Kv channel activity in the adult, a prerequisite for myelination.92 Also, patch-clamp recordings reveal a large Kv current in Schwann cells of dysmyelinating mutant Trembler

108

Function of the Peripheral Nervous System

mice,27 suggesting that pathologic hyperphosphorylation of Kv channels in adult Schwann cells might be incompatible with myelination. Whether Kv1.5 is the Schwann cell channel species expressed in the Trembler Schwann cell is unclear. Clinically, pathologic upregulation of Kv channel activity in Schwann cells might cause schwannomas, and blockage of Kv channels has been suggested as a possible therapeutic intervention to slow the growth of this type of tumor.60,107 Second, genetic ablation of K⫹ channels in mice has in some cases produced hypomyelination. A striking example can be found in genetic ablation of a Kir channel (Kir4.1), which causes severe hypomyelination in the CNS axons.85 Kir4.1 is expressed in oligodendrocytes, the myelinating glia of CNS axons, and deleting Kir4.1 channels from oligodendrocytes results in a depolarized resting membrane potential and immature morphology in these cells.85 Global genetic ablation in mice of Kir2.1, the Kir species found in Schwann cells, results in lethality, which has prevented any study of possible perturbation on myelination.142 Interestingly, mutation in Kir2.1 in humans causes Andersen’s syndrome (AS), a rare disorder characterized by periodic paralysis and cardiac arrhythmias.3,94 Whether patients with AS have dysmyelination in the PNS remains unexplored. Kir2.1 channels maintain a stable resting potential that is important for cell function,59 and it remains possible that mutation of this channel causes a deviation of the resting potential of Schwann cells to a pathologic level that is incompatible with normal myelination. The limitation of the global knockout is the lack of tissue specificity, and in some cases lethality, a limitation that can be ameliorated by the use of the Cre-Lox recombinase system to limit gene inactivation to Schwann cells. This technique has recently been successfully applied to selectively ablate non–ion channel genes in Schwann cells,46 and selective inactivation of ion channel genes in Schwann cells using this technique will probably be the next step for shedding light on in vivo functions of K⫹ channels in Schwann cells.

3.

4. 5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

ACKNOWLEDGMENTS

16.

Work in the author’s laboratory was supported by grants from the National Institutes of Health, the National Multiple Sclerosis Society, and a Pew Scholar Award in the Biomedical Sciences.

17.

REFERENCES

19.

1. Aelman, J. P., Bond, C. T., Pessia, M., and Maylie, J.: Episodic ataxia results from voltage-dependent potassium channels with altered functions. Neuron 15:1449, 1995. 2. Amedee, T., Ellie, E., Dupouy, B., and Vincent, J. D.: Voltagedependent calcium and potassium channels in Schwann cells

20.

18.

cultured from dorsal root ganglia of the mouse. J. Physiol. (Lond.) 441:35, 1991. Andelfinger, G., Tapper, A. R., Welch, R. C., et al.: KCNJ2 mutation results in Andersen syndrome with sex-specific cardiac and skeletal muscle phenotypes. Am. J. Hum. Genet. 71:663, 2002. Arroyo, E. J., and Scherer, S. S.: On the molecular architecture of myelinated fibers. Histochem. Cell Biol. 113:1, 2000. Arroyo, E. J., Xu, Y. T., Zhou, L., et al.: Myelinating Schwann cells determine the internodal localization of Kv1.1, Kv1.2, Kv␤2, and Caspr. J. Neurocytol. 28:333, 1999. Baker, M., Bostock, H., Grafe, P., and Martius, P.: Function and distribution of three types of rectifying channel in rat spinal root myelinated axons. J. Physiol. (Lond.) 383:45, 1987. Baker, M., Howe, J. R., and Ritchie, J. M.: Two types of 4-aminopyridine-sensitive potassium current in rabbit Schwann cells. J. Physiol. (Lond.) 464:321, 1993. Barrett, E. F., and Barrett, J. N.: Intracellular recording from vertebrate myelinated axons: mechanism of the depolarizing afterpotential. J. Physiol. (Lond.) 323:117, 1982. Bennett, V., and Lambert, S.: Physiological roles of axonal ankyrins in survival of premyelinated axons and localization of voltage-gated sodium channels. J. Neurocytol. 28:303, 1999. Benoit, E., Corbier, A., and Dubois, J. M.: Evidence for two transient sodium currents in the frog node of Ranvier. J. Physiol. (Lond.) 361:339, 1985. Berghs, S., Aggujaro, D., Dirkx, R., et al.: Beta IV spectrin, a new spectrin localized at axon initial segments and nodes of Ranvier in the central and peripheral nervous system. J. Cell Biol. 151:985, 2000. Bever, C. T. Jr., Young, D., Anderson, P. A., et al.: The effects of 4-aminopyridine in multiple sclerosis patients: results of a randomized, placebo-controlled, double-blind, concentration-controlled, crossover trial. Neurology 44:1054, 1994. Bhat, M. A., Rios, J. C., Lu, Y., et al.: Axon-glia interactions and the domain organization of myelinated axons requires Neurexin IV/Caspr/Paranodin. Neuron 30:369, 2001. Boiko, T., Rasband, M. N., Levinson, S. R., et al.: Compact myelin dictates the differential targeting of two sodium channel isoforms in the same axon. Neuron 30:91, 2001. Bostock, H., Sears, T. A., and Sherratt, R. M.: The effects of 4-aminopyridine and tetraethylammonium ions on normal and demyelinated mammalian nerve fibres. J. Physiol. (Lond.) 313:301, 1981. Boyle, M. E., Berglund, E. O., Murai, K. K., et al.: Contactin orchestrates assembly of the septate-like junctions at the paranode in myelinated peripheral nerve. Neuron 30:385, 2001. Brismar, T.: Potential clamp analysis of membrane currents in rat myelinated nerve fibres. J. Physiol. (Lond.) 298:171, 1980. Brown, M. J., and Asbury, A. K.: Schwann cell proliferation in the postnatal mouse: timing and topography. Exp. Neurol. 74:170, 1981. Browne, D. L., Gancher, S. T., Nutt, J. G., et al.: Episodic ataxia/myokymia syndrome is associated with point mutations in the human potassium channel gene, KCNA1 [see comments]. Nat. Genet. 8:136, 1994. Burgess, D. L., Kohrman, D. C., Galt, J., et al.: Mutation of a new sodium channel gene, Scn8a, in the mouse mutant “motor endplate disease.” Nat. Genet. 10:461, 1995.

Channel Function in Mammalian Axons and Support Cells 21. Caldwell, J. H., Schaller, K. L., Lasher, R. S., et al.: Sodium channel Na(v)1.6 is localized at nodes of Ranvier, dendrites, and synapses. Proc. Natl. Acad. Sci. U. S. A. 97:5616, 2000. 22. Campomanes, C. R., Carroll, K. I., Manganas, L. N., et al.: Kv beta subunit oxidoreductase activity and Kv1 potassium channel trafficking. J. Biol. Chem. 277:8298, 2002. 23. Chiu, S. Y.: Inactivation of sodium channels: second order kinetics in myelinated nerve. J. Physiol. (Lond.) 273:573, 1977. 24. Chiu, S. Y.: Asymmetry currents in the mammalian myelinated nerve. J. Physiol. (Lond.) 309:499, 1980. 25. Chiu, S. Y.: Sodium currents in axon-associated Schwann cells from adult rabbits. J. Physiol. (Lond.) 386:181, 1987. 26. Chiu, S. Y.: Changes in excitable membrane properties in Schwann cells of adult rabbit sciatic nerves following nerve transection. J. Physiol. (Lond.) 396:173, 1988. 27. Chiu, S. Y.: Patch-clamp studies on myelin-deficient Schwann cells from mutant trembler mice [abstract]. Soc. Neurosci. Abstr.14:48.2, 1988. 28. Chiu, S. Y.: Functions and distribution of voltage-gated sodium and potassium channels in mammalian Schwann cells. Glia 4:541, 1991. 29. Chiu, S. Y.: Channel function in mammalian axons and support cells. In Dyck, P. J., Thomas, P. K., Griffin, J. W., et al. (eds.): Peripheral Neuropathy, 3rd ed. Philadelphia, W. B. Saunders, p. 94, 1993. 30. Chiu, S. Y.: Differential expression of sodium channels in acutely isolated myelinating and non-myelinating Schwann cells of rabbits. J. Physiol. (Lond.) 470:485, 1993. 31. Chiu, S. Y., and Ritchie, J. M.: Potassium channels in nodal and internodal axonal membrane of mammalian myelinated fibres. Nature 284:170, 1980. 32. Chiu, S. Y., and Ritchie, J. M.: Evidence for the presence of potassium channels in the paranodal region of acutely demyelinated mammalian single nerve fibres. J. Physiol. (Lond.) 313:415, 1981. 33. Chiu, S. Y., and Ritchie, J. M.: On the physiological role of internodal potassium channels and the security of conduction in myelinated nerve fibres. Proc. R. Soc. Lond. B Biol. Sci. 220:415, 1984. 34. Chiu, S. Y., Ritchie, J. M., Rogart, R. B., and Stagg, D.: A quantitative description of membrane currents in rabbit myelinated nerve. J. Physiol. (Lond.) 292:149, 1979. 35. Chiu, S. Y., Scherer, S. S., Blonski, M., et al.: Axons regulate the expression of Shaker-like potassium channel genes in Schwann cells in peripheral nerve. Glia 12:1, 1994. 36. Chiu, S. Y., Schrager, P., and Ritchie, J. M.: Neuronal-type Na⫹ and K⫹ channels in rabbit cultured Schwann cells. Nature 311:156, 1984. 37. Chiu, S. Y., and Schwarz, W.: Sodium and potassium currents in acutely demyelinated internodes of rabbit sciatic nerves. J. Physiol. (Lond.) 391:631, 1987. 38. Chiu, S. Y., and Wilson, G. F.: The role of potassium channels in Schwann cell proliferation in Wallerian degeneration of explant rabbit sciatic nerves. J. Physiol. (Lond.) 408:199, 1989. 39. Chiu, S. Y., Zhou, L., Zhang, C. L., and Messing, A.: Analysis of potassium channel functions in mammalian axons by gene knockouts. J. Neurocytol. 28:349, 1999. 40. Chouinard, S. W., Wilson, G. F., Schlimgen, A. K., and Ganetzky, B.: A potassium channel beta subunit related to

41.

42.

43.

44.

45.

46.

47. 48.

49.

50.

51. 52.

53.

54. 55. 56.

57. 58.

59.

109

the aldo-keto reductase superfamily is encoded by the Drosophila hyperkinetic locus. Proc. Natl. Acad. Sci. U. S. A. 92:6763, 1995. Dodge, F. A., and Frankenhaeuser, B.: Membrane currents in isolated frog nerve fibre under voltage clamp conditions. J. Physiol. (Lond.) 143:76, 1958. Doyle, D. A., Morais, C. J., Pfuetzner, R. A., et al.: The structure of the potassium channel: molecular basis of K⫹ conduction and selectivity. Science 280:69, 1998. Duchen, L. W., and Stefani, E.: Electrophysiological studies of neuromuscular transmission in hereditary “motor end-plate disease” of the mouse. J. Physiol. (Lond.) 212:535, 1971. Einheber, S., Zanazzi, G., Ching, W., et al.: The axonal membrane protein Caspr, a homologue of neurexin IV, is a component of the septate-like paranodal junctions that assemble during myelination. J. Cell Biol. 139:1495, 1997. Feltri, M. L., D’Antonio, M., Quattrini, A., et al.: A novel P0 glycoprotein transgene activates expression of lacZ in myelinforming Schwann cells. Eur. J. Neurosci. 11:1577, 1999. Feltri, M. L., Graus, P. D., Previtali, S. C., et al.: Conditional disruption of beta 1 integrin in Schwann cells impedes interactions with axons. J. Cell Biol. 156:199, 2002. Goldin, A. L.: Diversity of mammalian voltage-gated sodium channels. Ann. N. Y. Acad. Sci. 868:38, 1999. Goldin, A. L., Barchi, R. L., Caldwell, J. H., et al.: Nomenclature of voltage-gated sodium channels. Neuron 28:365, 2000. Gong, B., Rhodes, K. J., Bekele-Arcuri, Z., and Trimmer, J. S.: Type I and type II Na(⫹) channel alpha-subunit polypeptides exhibit distinct spatial and temporal patterning, and association with auxiliary subunits in rat brain. J. Comp. Neurol. 412:342, 1999. Grossfeld, R. M., Klinge, M. A., Lieberman, E. M., and Stewart, L. C.: Axon-glia transfer of a protein and a carbohydrate. Glia 1:292, 1988. Hille, B., Armstrong, C. M., and MacKinnon, R.: Ion channels: from idea to reality. Nat. Med. 5:1105, 1999. Hodgkin, A. L., and Huxley, A. F.: A quantitative description of membrane current and its application to conduction and excitation in nerve. J. Physiol. (Lond.) 117:500, 1952. Horackova, M., Nonner, W., and Stampfli, R.: Action potentials and voltage-clamp currents of single rat Ranvier nodes [abstract]. Proc. Int. Union Physiol. 7:198, 1968. Isom, L. L., and Catterall, W. A.: Na⫹ channel subunits and Ig domains. Nature 383:307, 1996. Isom, L. L., De Jongh, K. S., and Catterall, W. A.: Auxiliary subunits of voltage-gated ion channels. Neuron 12:1183, 1994. Isom, L. L., Ragsdale, D. S., De Jongh, K. S., et al.: Structure and function of the beta 2 subunit of brain sodium channels, a transmembrane glycoprotein with a CAM motif. Cell 83:433, 1995. Jan, L. Y., and Jan, Y. N.: Cloned potassium channels from eukaryotes and prokaryotes. Annu. Rev. Neurosci. 20:91, 1997. Jonas, P., Brau, M. E., Hermsteiner, M., and Vogel, W.: Single-channel recording in myelinated nerve fibers reveals one type of Na channel but different K channels. Proc. Natl. Acad. Sci. U. S. A. 86:7238, 1989. Jongsma, H. J., and Wilders, R.: Channelopathies: Kir2.1 mutations jeopardize many cell functions. Curr. Biol. 11:R747, 2001.

110

Function of the Peripheral Nervous System

60. Kamleiter, M., Hanemann, C. O., Kluwe, L., et al.: Voltagedependent membrane currents of cultured human neurofibromatosis type 2 Schwann cells. Glia 24:313, 1998. 61. Kaplan, M. R., Meyer-Franke, A., Lamber, S., et al.: Induction of sodium channel clustering by oligodendrocytes. Nature 386:724, 1997. 62. Kearney, J. A., Plummer, N. W., Smith, M. R., et al.: A gainof-function mutation in the sodium channel gene Scn2a results in seizures and behavioral abnormalities. Neuroscience 102:307, 1901. 63. Kocsis, J. D., Ruiz, J. A., and Waxman, S. G.: Maturation of mammalian myelinated fibers: changes in action-potential characteristics following 4-aminopyridine application. J. Neurophysiol. 50:449, 1983. 64. Kohrman, D. C., Smith, M. R., Goldin, A. L., et al.: A missense mutation in the sodium channel Scn8a is responsible for cerebellar ataxia in the mouse mutant jolting. J. Neurosci. 16:5993, 1996. 65. Konishi, T.: Voltage-dependent potassium channels in mouse Schwann cells. J. Physiol. (Lond.) 411:115, 1989. 66. Konishi, T.: cAMP-mediated expression of inwardly rectifying potassium channels in cultured mouse Schwann cells. Brain Res. 594:197, 1992. 67. Konishi, T.: Activity-dependent regulation of inwardly rectifying potassium currents in non-myelinating Schwann cells in mice. J. Physiol. (Lond.) 474:193, 1994. 68. Kuffler, S. W., Nicholls, J. G., and Orkand, R. K.: Physiological properties of glial cells in the central nervous system of amphibia. J. Neurophysiol. 29:768, 1966. 69. Lambert, S., Davis, J. O., and Bennett, V.: Morphogenesis of the node of Ranvier: co-clusters of ankyrin and ankyrinbinding integral proteins define early developmental intermediates. J. Neurosci. 17:7025, 1997. 70. Li, M., Jan, Y. N., and Jan, L. Y.: Specification of subunit assembly by the hydrophilic amino-terminal domain of the Shaker potassium channel. Science 257:1225, 1992. 71. Liu, C. J., Dib-Hajj, S. D., Black, J. A., et al.: Direct interaction with contactin targets voltage-gated sodium channel Na(v)1.9/NaN to the cell membrane. J. Biol. Chem. 276:46553, 2001. 72. Low, P. A.: Endoneurial potassium is increased and enhances spontaneous activity in regenerating mammalian nerve fibers-implications for neuropathic positive symptoms. Muscle Nerve 8:27, 1985. 73. Lustig, M., Zanazzi, G., Sakurai, T., et al.: Nr-CAM and neurofascin interactions regulate ankyrin G and sodium channel clustering at the node of Ranvier. Curr. Biol. 11:1864, 2001. 74. MacFarlane, S. N., and Sontheimer, H.: Modulation of Kv1.5 currents by Src tyrosine phosphorylation: potential role in the differentiation of astrocytes. J. Neurosci. 20:5245, 2000. 75. Malhotra, J. D., Kazen-Gillespie, K., Hortsch, M., and Isom, L. L.: Sodium channel beta subunits mediate homophilic cell adhesion and recruit ankyrin to points of cell-cell contact. J. Biol. Chem. 275:11383, 2000. 76. Manganas, L. N., Akhtar, S., Antonucci, D. E., et al.: Episodic ataxia type-1 mutations in the Kv1.1 potassium channel display distinct folding and intracellular trafficking properties. J. Biol. Chem. 276:49427, 2001. 77. Manganas, L. N., and Trimmer, J. S.: Subunit composition determines Kv1 potassium channel surface expression. J. Biol. Chem. 275:29685, 2000.

78. McCormack, K., Connor, J. X., Zhou, L., et al.: Genetic analysis of the mammalian K⫹ channel beta-subunit Kvbeta 2 (Kcnab2). J. Biol. Chem. 277:13219, 2002. 79. McCormack, T., and McCormack, K.: Shaker K⫹ channel ␤ subunits belong to an NAD(P)H-dependent oxidoreductase superfamily. Cell 79:1133, 1994. 80. Menegoz, M., Gaspar, P., Le, B. M., et al.: Paranodin, a glycoprotein of neuronal paranodal membranes. Neuron 19:319, 1997. 81. Mi, H., Deerinck, T. J., Ellisman, M. H., and Schwarz, T. L.: Differential distribution of closely related potassium channels in rat Schwann cells. J. Neurosci. 15:3761, 1995. 82. Mi, H. Y., Deerinck, T. J., Jones, M., et al.: Inwardly rectifying K⫹ channels that may participate in K⫹ buffering are localized in microvilli of Schwann cells. J. Neurosci. 16:2421, 1996. 83. Miller, A.: Current and investigational therapies used to alter the course of disease in multiple sclerosis. South. Med. J. 90:367, 1997. 84. Morais-Cabral, J. H., Zhou, Y., and MacKinnon, R.: Energetic optimization of ion conduction rate by the K⫹ selectivity filter. Nature 414:37, 2001. 85. Neusch, C., Rozengurt, N., Jacobs, R. E., et al.: Kir4.1 potassium channel subunit is crucial for oligodendrocyte development and in vivo myelination. J. Neurosci. 21:5429, 2001. 86. Nonner, W.: A new voltage clamp method for Ranvier nodes. Pflugers Arch. 309:176, 1969. 87. Oh, Y., and Waxman, S. G.: The beta 1 subunit mRNA of the rat brain Na⫹ channel is expressed in glial cells. Proc. Natl. Acad. Sci. U. S. A. 91:9985, 1994. 88. Orkand, R. K.: Glial electrophysiology and transport. Ann. N. Y. Acad. Sci. 633:245, 1991. 89. Pappas, C. A., and Ritchie, J. M.: Effect of specific ion channel blockers on cultured Schwann cell proliferation. Glia 22:113, 1998. 90. Parkinson, N. J., Olsson, C. L., Hallows, J. L., et al.: Mutant beta-spectrin 4 causes auditory and motor neuropathies in quivering mice. Nat. Genet. 29:61, 2001. 91. Pedraza, L., Huang, J. K., and Colman, D. R.: Organizing principles of the axoglial apparatus. Neuron 30:335, 2001. 92. Peretz, A., Gil-Henn, H., Sobko, A., et al.: Hypomyelination and increased activity of voltage-gated K⫹ channels in mice lacking protein tyrosine phosphatase ␧. EMBO J. 19:4036, 2000. 93. Peretz, A., Sobko, A., and Attali, B.: Tyrosine kinases modulate K⫹ channel gating in mouse Schwann cells. J. Physiol. (Lond.) 519(pt. 2):373, 1999. 94. Plaster, N. M., Tawil, R., Tristani-Firouzi, M., et al.: Mutations in Kir2.1 cause the developmental and episodic electrical phenotypes of Andersen’s syndrome. Cell 105:511, 2001. 95. Poliak, S., Gollan, L., Martinez, R., et al.: Caspr2, a new member of the neurexin superfamily, is localized at the juxtaparanodes of myelinated axons and associates with K⫹ channels. Neuron 24:1037, 1999. 96. Pongs, O., Leicher, T., Berger, M., et al.: Functional and molecular aspects of voltage-gated K⫹ channel beta subunits. Ann. N. Y. Acad. Sci. 868:344, 1999. 97. Quick, D. C., Kennedy, W. R., and Donaldson, L.: Dimensions of myelinated nerve fibers near the motor and sensory terminals in cat tenuissimus muscles. Neuroscience 4:1089, 1979.

Channel Function in Mammalian Axons and Support Cells 98. Rasband, M., Trimmer, J. S., Schwarz, T. L., et al.: Potassium channel distribution, clustering, and function in remyelinating rat axons. J. Neurosci. 18:36, 1998. 99. Rasband, M. N., and Trimmer, J. S.: Developmental clustering of ion channels at and near the node of Ranvier [review]. Dev. Biol. 236:5, 2001. 100. Rasband, M. N., and Trimmer, J. S.: Subunit composition and novel localization of K⫹ channels in spinal cord. J. Comp. Neurol. 429:166, 2001. 101. Ratcliffe, C. F., Westenbroek, R. E., Curtis, R., and Catterall, W. A.: Sodium channel beta1 and beta3 subunits associate with neurofascin through their extracellular immunoglobulin-like domain. J. Cell Biol. 154:427, 2001. 102. Revenko, S. V., Timin, E. N., and Khodorov, B. I.: Nerve impulse conduction from the myelinated portion of the axon to the nonmyelinated terminal [in Russian]. Biofizika 18:1074, 1973. 103. Rhodes, K. J., Strassle, B. W., Monaghan, M. M., et al.: Association and colocalization of the Kv␤1 and Kv␤2 ␤-subunits with Kv1 ␣-subunits in mammalian K⫹ channel complexes. J. Neurosci. 17:8246, 1997. 104. Rieger, F., Pincon-Raymond, M., Lombet, A., et al.: Paranodal dysmyelination and increase in tetrodotoxin binding sites in the sciatic nerve of the motor end-plate disease (med/med) mouse during postnatal development. Dev. Biol. 101:401, 1984. 105. Ritchie, J. M.: Sodium-channel turnover in rabbit cultured Schwann cells. Proc. R. Soc. Lond. B Biol. Sci. 233:423, 1988. 106. Ritchie, J. M., and Rogart, R. B.: Density of sodium channels in mammalian myelinated nerve fibers and nature of the axonal membrane under the myelin sheath. Proc. Natl. Acad. Sci. U. S. A. 74:211, 1977. 107. Rosenbaum, C., Kamleiter, M., Grafe, P., et al.: Enhanced proliferation and potassium conductance of Schwann cells isolated from NF2 schwannomas can be reduced by quinidine. Neurobiol. Dis. 7:483, 2000. 108. Rosenbluth, J.: Intramembranous particle distribution at the node of Ranvier and adjacent axolemma in myelinated axons of the frog brain. J. Neurocytol. 5:731, 1976. 109. Schauf, C. L., and Davis, F. A.: Impulse conduction in multiple sclerosis: a theoretical basis for modification by temperature and pharmacological agents. J. Neurol. Neurosurg. Psychiatry 37:152, 1974. 110. Schwarz, J. R., Reid, G., and Bostock, H.: Action potentials and membrane currents in the human node of Ranvier. Pflugers Arch. 430:283, 1995. 111. Sherratt, R. M., Bostock, H., and Sears, T. A.: Effects of 4-aminopyridine on normal and demyelinated mammalian nerve fibres. Nature 283:570, 1980. 112. Shi, G., Nakahira, K., Hammond, S., et al.: ␤ Subunits promote K⫹ channel surface expression through effects early in biosynthesis. Neuron 16:843, 1996. 113. Shrager, P., Chiu, S. Y., and Ritchie, J. M.: Voltage-dependent sodium and potassium channels in mammalian cultured Schwann cells. Proc. Natl. Acad. Sci. U. S. A. 82:948, 1985. 114. Smart, S. L., Lopantsev, V., Zhang, C. L., et al.: Deletion of the Kv1.1 potassium channel causes epilepsy in mice. Neuron 20:809, 1998. 115. Sobko, A., Peretz, A., and Attali, B.: Constitutive activation of delayed-rectifier potassium channels by a src family tyrosine kinase in Schwann cells. EMBO J. 17:4723, 1998.

111

116. Sobko, A., Peretz, A., Shirihai, O., et al.: Heteromultimeric delayed-rectifier K⫹ channels in Schwann cells: developmental expression and role in cell proliferation. J. Neurosci. 18:10398, 1998. 117. Soltoff, S. P., and Cantley, L. C.: Mitogens and ion fluxes. Annu. Rev. Physiol. 50:207, 1988. 118. Sontheimer, H., Black, J. A., and Waxman, S. G.: Voltagegated Na⫹ channels in glia: properties and possible functions. Trends Neurosci. 19:325, 1996. 119. Spampanato, J., Escayg, A., Meisler, M. H., and Goldin, A. L.: Functional effects of two voltage-gated sodium channel mutations that cause generalized epilepsy with febrile seizures plus type 2. J. Neurosci. 21:7481, 2001. 120. Srinivasan, Y., Elmer, L., Davis, J., et al.: Ankyrin and spectrin associate with voltage-dependent sodium channels in brain. Nature 333:177, 1988. 121. Stefoski, D., Davis, F. A., Faut, M., and Schauf, C. L.: 4-Aminopyridine improves clinical signs in multiple sclerosis. Ann. Neurol. 21:71, 1987. 122. Tait, S., Gunn-Moore, F., Collinson, J. M., et al.: An oligodendrocyte cell adhesion molecule at the site of assembly of the paranodal axo-glial junction. J. Cell Biol. 150:657, 2000. 123. Targ, E. F., and Kocsis, J. D.: 4-Aminopyridine leads to restoration of conduction in demyelinated rat sciatic nerve. Brain Res. 328:358, 1985. 124. Torebjork, H. E., Ochoa, J. L., and McCann, F. V.: Paresthesiae: abnormal impulse generation in sensory nerve fibres in man. Acta Physiol. Scand. 105:518, 1979. 125. Trimmer, J. S.: Regulation of ion channel expression by cytoplasmic subunits. Curr. Opin. Neurobiol. 8:370, 1998. 126. Tytell, M., and Lasek, R. J.: Glial polypeptides transferred into the squid giant axon. Brain Res. 324:223, 1984. 127. Vabnick, I., Novakovic, S. D., Levinson, S. R., et al.: The clustering of axonal sodium channels during development of the peripheral nervous system. J. Neurosci. 16:4914, 1996. 128. Vabnick, I., Trimmer, J. S., Schwarz, T. L., et al.: Dynamic potassium channel distributions during axonal development prevent aberrant firing patterns. J. Neurosci. 19:747, 1999. 129. Verkhratsky, A., Hoppe, D., and Kettenmann, H.: K⫹ channel properties in cultured mouse Schwann cells: dependence on extracellular K⫹. J. Neurosci. Res. 28:210, 1991. 130. Vogel, W., and Schwarz, J. R.: Voltage-clamp studies in axons: macroscopic and single-channel currents. In Waxman, S. G., Kocsis, J. D., and Stys, P. K. (eds.): The Axon: Structure, Function and Pathophysiology. New York, Oxford University Press, p. 257, 1995. 131. Wang, H., Allen, M. L., Grigg, J. J., et al.: Hypomyelination alters K⫹ channel expression in mouse mutants shiverer and Trembler. Neuron 15:1337, 1995. 132. Wang, H., Kunkel, D. D., Martin, T. M., et al.: Heteromultimeric K⫹ channels in terminal and juxtaparanodal regions of neurons. Nature 365:75, 1993. 133. Waxman, S. G.: Voltage-gated ion channels in axon: localization, function, and development. In Waxman, S. G., Kocsis, J. D., and Stys, P. K. (eds.): The Axon: Structure, Function and Pathophysiology. New York, Oxford University Press, p. 218, 1995. 134. Waxman, S. G., and Brill, M. H.: Conduction through demyelinated plaques in multiple sclerosis: computer simulations of facilitation by short internodes. J. Neurol. Neurosurg. Psychiatry 41:408, 1978.

112

Function of the Peripheral Nervous System

135. Waxman, S. G., and Ritchie, J. M.: Molecular dissection of the myelinated axon. Ann. Neurol. 33:121, 1993. 136. Webster, H. D., and Favilla, J. T.: Development of peripheral nerve fibers. In Dyck, P. J., Thomas, P. K., Lambert, E. H., and Bunge, R. P. (eds.): Peripheral Neuropathy, 2nd ed. Philadelphia, W. B. Saunders, p. 329, 1984. 137. Westenbroek, R.E., Noebels, J.L., and Catterall, W.A.: Elevated expression of type II Na⫹ channels in hypomyelinated axons of shiverer mouse brain. J. Neurosci. 12:2259, 1992. 138. Wilson, G. F., and Chiu, S. Y.: Ion channels in axon and Schwann cell membranes at paranodes of mammalian myelinated fibers studied with patch clamp. J. Neurosci. 10:3263, 1990. 139. Wilson, G. F., and Chiu, S. Y.: Potassium channel regulation in Schwann cells during early developmental myelinogenesis. J. Neurosci. 10:1615, 1990. 140. Wilson, G. F., and Chiu, S. Y.: Mitogenic factors regulate ion channels in Schwann cells cultured from newborn rat sciatic nerve. J. Physiol. (Lond.) 470:501, 1993. 141. Wood, P. M., and Bunge, R. P.: Evidence that sensory axons are mitogenic for Schwann cells. Nature 256:662, 1975. 142. Zaritsky, J. J., Eckman, D. M., Wellman, G. C., et al.: Targeted disruption of Kir2.1 and Kir2.2 genes reveals the essential role of the inwardly rectifying K(⫹) current in K(⫹)-mediated vasodilation [see comments]. Circ. Res. 87:160, 2000.

143. Zerr, P., Adelman, J. P., and Maylie, J.: Episodic ataxia mutations in Kv1.1 alter potassium channel function by dominant negative effects or haploinsufficiency. J. Neurosci. 18:2842, 1998. 144. Zhang, C. L., Messing, A., and Chiu, S. Y.: Specific alteration of spontaneous GABAergic inhibition in cerebellar Purkinje cells in mice lacking the potassium channel Kv1.1. J. Neurosci. 19:2852, 1999. 145. Zhou, D., Lambert, S., Malen, P. L., et al.: AnkyrinG is required for clustering of voltage-gated Na channels at axon initial segments and for normal action potential firing. J. Cell Biol. 143:1295, 1998. 146. Zhou, L., and Chiu, S. Y.: Computer model for action potential propagation through branch point in myelinated nerves. J. Neurophysiol. 85:197, 2001. 147. Zhou, L., Messing, A., and Chiu, S. Y.: Determinants of excitability at transition zones in Kv1.1-deficient myelinated nerves. J. Neurosci. 19:5768, 1999. 148. Zhou, L., Zhang, C. L., Messing, A., and Chiu, S. Y.: Temperature-sensitive neuromuscular transmission in Kv1.1 null mice: role of potassium channels under the myelin sheath in young nerves. J. Neurosci. 18:7200, 1998. 149. Zhou, Y., Morais-Cabral, J. H., Kaufman, A., and MacKinnon, R.: Chemistry of ion coordination and hydration revealed by a K⫹ channel-Fab complex at 2.0 Å resolution. Nature 414:43, 2001.

5 Nerve Excitability Measures: Biophysical Basis and Use in the Investigation of Peripheral Nerve Disease MATTHEW C. KIERNAN, DAVID BURKE, AND HUGH BOSTOCK

Early Excitability Studies Nerve Excitability and Nerve Conduction Studies Compared Currently Used Measures of Nerve Excitability and Their Biophysical Basis Strength-Duration Behavior Recovery Cycles: Refractoriness, Superexcitability, and Late Subexcitability Threshold Electrotonus and Current-Threshold Relationship

Automated Recording of Multiple Nerve Excitability Properties Nerve Excitability Properties and Membrane Potential Effects of Ischemia on Axonal Excitability Modality and Regional Differences between Different Axons Changes in Axonal Membrane Potential in Peripheral Neuropathies

Nowadays, electrical stimulation of peripheral nerves for the purpose of clinical diagnosis nearly always employs supramaximal stimuli, to measure the velocity and amplitude of compound sensory or motor action potentials (see Chapter 35). This chapter is concerned with the alternative approach of using submaximal stimuli to determine the ease with which axons can be excited. Such studies of electrical excitability may seem novel, but they actually have a much longer history than velocity measurements, albeit a history tarnished with confusion caused by erroneous theories. More recent excitability studies are based on (and have contributed to) a much better understanding of axonal biophysics, and it is argued that they have the potential to make important contributions to our understanding of the pathophysiology of peripheral neuropathy.

EARLY EXCITABILITY STUDIES Fifty years ago, Hodgkin and Huxley’s revolutionary theory of nerve excitability36 and conduction was paralleled by an

Multifocal Motor Neuropathy and Axonal Hyperpolarization Uremic Neuropathy and Axonal Depolarization Impulse Conduction in Demyelinating Neuropathies Potential Role for Excitability Studies in the Neuropathy Clinic

independent revolution in the clinical electrophysiology of peripheral nerve, as measurements of excitability gave way to measurements of velocity. A 1952 textbook84 lists five methods of electrodiagnosis in neurology: electromyography, electroencephalography, psychogalvanic reflex, galvanic vertigo, and “the assessment of tissue excitability and what is termed ‘accommodation’.” Conduction velocity measurements are not mentioned (although an early clinical study35 had been published in 1948). The then “classical method” of excitability testing, which had been “the mainstay of electro-diagnosis for over a century,” involved testing nerves and muscles with “faradic” current from an induction coil, and “galvanic” current from a battery, to give short and long pulses, respectively.84 With the improvements of electronics during the Second World War, the simple galvanic-faradic test gave way to accurate determination of strength-duration (or intensitytime) curves using electronically generated rectangular pulses.87 For durations beyond the temps utile, threshold was a constant minimum (rheobase), while chronaxie was defined as the duration of a threshold current that was twice the rheobase. Strength-duration curves determined at the motor 113

114

Function of the Peripheral Nervous System

agreement as to standards has so far been reached.”84 It never was. The most important pre-1952 studies of accommodation in human nerves were probably Kugelberg’s thesis of 194453 and his demonstration of a dramatic increase in accommodation during ischemia, and decrease following release of ischemia, and the loss of accommodation associated with tetany or latent tetany.53–55 These early excitability studies helped to classify neuromuscular disorders as myopathic, neuropathic, or myelopathic, depending on whether they primarily affected muscle, nerve fiber, or nerve cell. During the 1950s, pioneering measurements of motor nerve conduction by Lambert and colleagues at Mayo Clinic, and of sensory nerve action potentials by Dawson, Scott, Sears, and Gilliatt at Queen Square, London (reviewed by Fowler30), set the scene for a rapid development of nerve conduction studies, and by the mid-1960s velocity measurements were being used to divide neuropathies into axonal and demyelinating types. Within 20 years, excitability studies were almost entirely displaced by conduction studies, which have held sway ever since.

point of a healthy muscle differ drastically from the responses of a denervated muscle, and were used in the clinical evaluation of the degree of muscle innervation (Fig. 5–1). The fact that normal motor point strength-duration curves resemble those obtained by stimulating the nerve trunk is simply due to the lower threshold of myelinated axon than muscle at the motor point. Unfortunately, however, the observation was misconstrued by Lapique, in his 1926 treatise on chronaxie.59 He formulated an electrical theory of neuromuscular transmission in which successful transmission required isochronism, or matching chronaxie between nerve and muscle. Lucas had already shown that nerve and muscle are not normally isochronic as early as 1906,67 and the principle of isochronism was soon soundly criticized.29,90 Nevertheless, Lapique’s fallacy lent measurements of chronaxie a spurious significance for many years, and isochronism persisted in medical texts as late as 1945.83 The other type of excitability measurement described in 1952 was accommodation, the tendency of a subthreshold current to slowly reduce nerve excitability. Studies of accommodation were also misguided by faulty theory. Hill’s theory of accommodation, published in 1936,33 was based in part on experiments on depolarized frog nerves, which led him to assume that accommodation to a steady current was always complete (so that the threshold to a superimposed test pulse should return to its original value) if the current was maintained long enough. According to this theory, accommodation was described by a single time constant, which could be deduced equally well from excitability tests made with current ramps, condenser discharges, or sinusoidal currents. For human nerves with normal resting potentials, however, accommodation is never more than partial (e.g., see Fig. 5–5 below) and Hill’s equations do not work well.4 It is therefore not surprising that the 1952 text stated: “Several methods of measuring and expressing the degree of accommodation present in a tissue have been suggested, but for technical reasons no

NERVE EXCITABILITY AND NERVE CONDUCTION STUDIES COMPARED Over the last 30 years there have been few significant changes to the neurophysiologic investigation of patients with suspected neuromuscular disorders. Motor and sensory nerve conduction studies, in combination with electromyography, have remained the methods of choice for the clinician investigating peripheral nerve function. Routine nerve conduction studies can document the presence of a neuropathy with or without resting conduction block, but they may provide little further insight into disease pathophysiology. Measurements of action potential amplitude and latency provide information only on the

100 16

we

De

80

ee

w

60 Volts

24

we e

40 N or m a

20

at

rv

20

ne

ek s

ks

ks

ed

Chronaxie v o lu

nta r y

move m

e nt l contra la teral muscle

Rheobase

4 1 0.2 0.2 0.08

Temps utile 0 0 0.01 0.03

0.1

0.3 1 3 Time (msec)

10

30

100 300

FIGURE 5–1 The normal strength-duration curve from a motor point is shown by the dashed line, with the other curves from the denervated muscle of the other hand. Determinations at different times during the reinnervation show the return of the strength-duration curve toward normal. For stimulus durations greater than the temps utile (arrow), the threshold is constant (i.e., the rheobase). Chronaxie is the stimulus duration at which threshold is twice the rheobase. (Adapted from Ochs, S.: Elements of Neurophysiology. New York, Wiley, 1965. Original data from Ritchie, A. R.: The electrical diagnosis of peripheral nerve injury. Brain 67:314, 1944.)

115

Nerve Excitability Measures

CURRENTLY USED MEASURES OF NERVE EXCITABILITY AND THEIR BIOPHYSICAL BASIS

number of conducting fibers and the conduction velocity of the fastest. However, conduction velocity is a nonspecific indicator of pathophysiology: It may be decreased by cooling, membrane depolarization or hyperpolarization, sodium channel blockade, axonal thinning, demyelination, or remyelination with short internodes. Measurements of axonal excitability are known to be sensitive to and can provide an indirect measure of resting membrane potential. Therefore, these measures provide complementary information to conventional nerve conduction studies. The currently preferred technique for assessing nerve excitability relies on the stimulus-based method of threshold tracking.18 With threshold tracking, changes in the intensity of the stimulus current required to generate a test potential of fixed amplitude are measured on line by a computer that in turn adjusts stimulus intensity to keep the amplitude of the subsequent test potential constant. “Threshold” in this context indicates the stimulus current required to produce a nerve or muscle action potential of target size (e.g., 40% of maximum). Threshold can be measured on line (“tracked”) during different maneuvers that alter nerve excitability. For example, when axons become hyperpolarized, the test potential will be smaller, and the computer will gradually increase stimulus intensity until the test potential has returned to its target size. The target is usually around 40% to 50% of maximum (because this is on the fast rising phase of the S-shaped stimulusresponse curve for the compound potential). “Threshold” is inversely related to axonal excitability, and can therefore provide an indirect measure of membrane potential.

The recent revival of nerve excitability studies by the authors and others has stemmed from basic research on the roles of different ion channels in myelinated axon.5 The principal ion channel types required to account for the electrical behavior of large myelinated axons in human peripheral nerves are indicated schematically in Figure 5–2. The contribution of the internodal axon compartment to action potential repolarization, afterpotentials, and slow excitability changes was first demonstrated by the microelectrode studies on amphibian axons of Ellen and John Barrett.7 A quantitative electrical model, showing how the nodal and internodal ion channels could interact to generate slow changes in excitability, was subsequently proposed on the basis of excitability measurements on human axons. 15 That model incorporated all the channel types shown in Figure 5–2 except for the persistent Na channels, which later excitability studies on human axons showed to be important in accounting for their surprisingly long strength-duration time constants, and for some important differences between motor and sensory axons.20 For further information on the complex electrical structure of myelinated axons, the reader is referred to recent reviews6,18,23,88 and papers describing equivalent electrical circuits and mathematical models.15,18,20,93,95 A detailed review of ion channel function is found in Chapter 4.

Na+

Nat Nap Ks

Kf

ATP

Ks

Ih

Lk

Kf

Ca2+ Na+

K+

Na+

K+

Na+

FIGURE 5–2 Schematic diagram of myelinated axon, showing principal components affecting electrical excitability. Nodal Na channels contribute both transient (Na t) and persistent (Na p) currents on depolarization: Na t is responsible for the regenerative upstroke of the action potential, and Na p affects subthreshold excitability. The nodal slow K conductance (Ks) is activated too slowly to contribute to action potential repolarization, but produces accommodation to subthreshold depolarizing currents and spike-frequency adaptation. Fast K channels (Kf) are concentrated in the juxtaparanodal region under the myelin, limit internodal depolarization during the spike, and so prevent nodal reexcitation by the depolarizing afterpotential. Hyperpolarization-activated cation current (Ih) limits post-tetanic hyperpolarization by the electrogenic sodium pump (Na,K-ATPase). There are also voltage-independent or leak conductances (Lk). The passive Na,Ca2 ion exchanger normally contributes to Ca2 homeostasis, but can be reversed with pathologic consequences during anoxia/ischemia. Ks, Kf, Lk, Na p, and Na,K-ATPase contribute to the resting potential.

116

Function of the Peripheral Nervous System

Here we focus on excitability properties, and introduce the electrical components in Figure 5–2 as they are reflected in excitability measurements. In addition to the measures featured in early excitability studies, namely strength-duration behavior and accommodation, another useful technique involves estimation of the “excitability cycle” or recovery cycle of excitability changes following an impulse. The phenomena of refractoriness and supernormality were described before the First World War.1 Clinical studies were initiated in the 1960s,32 and a trickle of subsequent studies have bucked the trend to concentrate exclusively on nerve conduction measurements.52 Of the many possible ways of measuring accommodation, the use of subthreshold rectangular current pulses to measure “threshold electrotonus”13 has been the most useful because the responses are normally closely related to electrotonus (the subthreshold voltage changes), and hence to the activity of voltage-dependent ion channels.12,13

Strength-Duration Behavior If a nerve membrane behaved simply as a passive electrical circuit before the threshold for excitation was reached, then the strength-duration measurements should result in

exponential membrane potential trajectories, and the strength-duration relationship should be described by Lapique’s59 equation: I  /[1  exp(t/)] where I is the threshold current for a pulse of duration t,  is the rheobase, and the strength-duration time constant  (defined as the ratio of threshold charge to rheobase for very brief stimuli76) is equal to chronaxie  loge2. However, the nerve membrane does not behave in this way, because sodium channels are activated before threshold is reached, giving rise to a local response. Instead the strength-duration relationship is better described by the earlier equation of Weiss103: Q  (t  ) where Q is the threshold charge (i.e., Q  It), and   chronaxie. This simple empirical relationship cannot be justified from first principles, but works remarkably well both for model axons following Hodgkin-Huxley type equations and for real rat and human axons11,21,71 (Fig. 5–3). The linearity of the relationship is such that estimates of the strength-duration time constant based on just two

50

A

Current (mA)

40 Sensory Motor

30 20 10 0 0.0

0.2

0.4

0.6

0.8

1.0

Stimulus duration (ms)

4

B

Sensory Motor

y = 0.76372 + 1.4122x r =0.997 y = 0.90940 + 2.1814x r =0.999

1

417 μs

2 540 μs

Charge (μC)

3

0 –1.0 –0.8 –0.6 –0.4 –0.2 0.0 0.2 0.4 Stimulus duration (ms)

0.6

0.8

1.0

FIGURE 5–3 A, Strength-duration curves for sensory and motor potentials of 30% of maximum recorded in the same experiment from the median nerve at the wrist. Arrows indicate chronaxies determined from the straight-line fits of the data in B to Weiss’s formula, using plots of threshold charge against stimulus duration. The time constant is given by the (negative) intercept on the duration axis, and the rheobase current is given by the slope of the regression line. (From Mogyoros, I., Kiernan, M. C., and Burke, D.: Strength-duration properties of human peripheral nerve. Brain 119:439, 1996, with permission from Oxford University Press.)

117

Nerve Excitability Measures

stimulus durations (echoing the earlier galvanic-faradic test) have been found to be just as reliable as estimates based on fitting a line to multiple points.71 Nevertheless, it should be remembered that Weiss’s equation is only approximate, and breaks down beyond the temps utile (see Fig. 5–1) when threshold is constant. It also overestimates threshold for currents of very short duration.66 The modeling of strength-duration behavior11 showed that  is linearly related to the passive nodal time constant, and therefore increases with the increased nodal capacitance in demyelination. However, because of its dependence on the subthreshold local response,  is also strongly dependent on sodium channels activated by subthreshold depolarization. Later modeling and excitability measurements20 provided evidence that differences in a persistent sodium conductance, and not differences in passive nodal properties, were responsible for the puzzling observation71,78 that sensory axons in human peripheral nerves have longer strength-duration time constants than the motor axons. The modeling accurately predicted an increase in , resulting from increased subthreshold activation of sodium channels, when nerves are depolarized.20

absolute refractoriness, axons are relatively refractory for up to 4.5 ms, during which greater current is required to generate an action potential. Refractoriness is dependent on fiber size and temperature: the smaller or cooler the fiber, the longer its duration. The refractory period is followed by a superexcitable (or supernormal) phase, characterized by a reduction in threshold occurring over a 10- to 15-ms interval. Finally, there is a phase of late subexcitability (or late subnormality) ending around 100 ms. These changes in threshold are associated with measurable changes in latency. Latency is initially increased during the refractory period, decreased during superexcitability, and increased during the late phase of subexcitability.8,96 The refractory period results from inactivation of transient Na channels.31 The refractory period is prolonged by cooling, as a result of slowed channel kinetics. It increases with membrane depolarization as a result of inactivation of Na channels and decreases with membrane hyperpolarization because approximately 30% of Na channels are inactivated at resting membrane potential. Consequently, the refractory period can be used as an indicator of membrane potential. The superexcitable period is due to a depolarizing afterpotential that results from the capacitative charging of the internode by the action potential, with subsequent discharge occurring through or under the myelin sheath.7 The superexcitable period changes reciprocally with membrane potential through effects on paranodal K channels, which open with membrane depolarization and close with membrane hyperpolarization. As a result, the superexcitable period reflects the status of the internodal membrane and can also be used as an indicator of membrane potential.

Recovery Cycles: Refractoriness, Superexcitability, and Late Subexcitability Axonal excitability oscillates in a stereotypical fashion following conduction of a single impulse, a process referred to as the recovery cycle8,49 (Fig. 5–4). Following conduction of a single action potential, the axon is initially absolutely refractory, so that it cannot produce further action potentials regardless of stimulus strength. Following this phase of

A

B 100

100

Threshold increase (%)

Threshold increase (%)

29 C

Refractoriness

0

32 C

Superexcitability

0 Late subexcitability 35 C

0

100 Inter-stimulus interval (ms)

10 100 Inter-stimulus interval (ms)

FIGURE 5–4 A, Recovery cycles for motor axons in the median nerve, averaged over 10 subjects at three different skin temperatures. B, Same data replotted with logarithmic time axis to show early part of recovery more clearly, and approximately uniform horizontal displacement of curves with temperature. (Data from Kiernan, M. C., Cikurel, K. and Bostock, H.: Effects of temperature on the excitability properties of human motor axons. Brain 124:816, 2001.)

118

Function of the Peripheral Nervous System

The final late subexcitable period reflects a corresponding membrane hyperpolarization and results from activation of a nodal slow K conductance.5,8 The driving force for K currents is the difference between the potassium equilibrium potential and the resting potential (ER  EK), so that subexcitability is sensitive to changes in the extracellular potassium concentration as well as the membrane potential.

Threshold Electrotonus and Current-Threshold Relationship Changes in potential of the nodal membrane spread to the internode, but slowly because of the resistance of the myelin sheath and consequently the slow charging of the internodal capacitance. This results in slow activation/deactivation of voltage-dependent channels on the internodal membrane. Although Na channel density is insufficient for the internodal membrane to generate an action potential, the changes in resistance of the internodal membrane and in the current stored on it will affect the behavior of the node.6 Approximately 99.9% of the axonal membrane is internodal so that, despite the lower channels densities, there are many more channels on the internode than on the node. The only technique that provides some insight into internodal conductances in human subjects in

A

vivo is the technique of threshold electrotonus. This term describes the changes in threshold produced by longlasting direct current (DC) pulses, which under most circumstances parallel the underlying electrotonic change in membrane potential. In response to depolarizing current pulses, there is an initial fast phase that is almost instantaneous and is proportional to the applied current (the “F” phase) (Fig. 5–5). This is followed by further depolarization that develops slowly over some tens of milliseconds as the current spreads to and depolarizes the internodal membrane (the “S1” phase). The threshold decrease (i.e., the extent of depolarization) reaches a peak approximately 20 ms after the onset of the current pulse, depending on its strength, and then threshold starts to return slowly toward the control level. This lessening of the degree of depolarization is due to activation of a hyperpolarizing conductance with slow kinetics. The use of K channel blocking agents indicates that the slow accommodative process is due to activation of slow K channels, which are located on both the node and the internode.5,13,18 When the DC pulse is terminated, threshold increases rapidly and there is then a slow overshoot before it gradually recovers to control level. The slow overshoot is due to the slow deactivation of a hyperpolarizing conductance, and is further evidence that the current pulse activated a slow K conductance.

B 100

100

F

0

F S1 –100

0 –100 Sensory

–200

S3 Motor

–300 Current (% threshold)

Threshold reduction (%) Current (% threshold)

Threshold reduction (%)

S2

S1

50 0 –50 0

100 Delay (ms)

200

0 –100 0

100 200 300 400 500 Delay (msec)

FIGURE 5–5 Normal threshold electrotonus waveforms and components. A, Averaged motor nerve responses to 100-ms polarizing currents ( 40% of threshold) from 38 subjects, showing fast (F) and slow (S1, S2) components (see text) and small intersubject variability (thick lines, mean; thin lines, mean  SD). Stimuli applied to ulnar nerve at wrist, and compound muscle action potential recorded from hypothenar muscle. (From Bostock, H., Cikurel, K., and Burke, D.: Threshold tracking techniques in the study of human peripheral nerve. Muscle Nerve 21:137, 1998, with permission.) B, Comparison between mean motor and sensory responses to 300-ms current pulses ( 40%, 80% of threshold) from eight subjects, showing additional component of threshold electrotonus (S3) resulting from inward rectification activated by hyperpolarization. (From Bostock, H., Burke, D., and Hales, J. P.: Differences in behaviour of sensory and motor axons following release of ischaemia. Brain 117:225, 1994, with permission of Oxford University Press.)

119

Nerve Excitability Measures

With long-lasting hyperpolarizing DC pulses, there is again a fast threshold change that increases threshold proportionally to the applied current, analogous to the comparable phase with depolarizing currents (the “F” phase). Then threshold continues to increase as the hyperpolarization spreads to the internode. This “S1” phase starts as a mirror image of the S1 phase with depolarizing current but soon diverges, because hyperpolarization closes K channels (slow nodal and fast and slow internodal K channels), and this increases the amplitude and time constant of S1. At approximately 150 ms after the start of DC hyperpolarization, S1 reaches a maximum and threshold begins to decrease toward the control level. This accommodative phase (“S3”) (see Fig. 5–5B) produces “inward rectification” and is due to activation of the hyperpolarization-activated current (IH).79 Although IH has slow activation and deactivation kinetics, it is activated and affects threshold earlier than 100 ms (where S1 peaks): Without IH, the hyperpolarizing threshold increase would be even greater. On termination of the hyperpolarizing DC pulse, threshold rapidly decreases and then undergoes a slow, depolarizing undershoot as IH is slowly deactivated and the slow K conductance is reactivated. Just as the activation of different voltage-dependent channels in a cell can be revealed by plotting a currentvoltage (I/V) relationship, it can be convenient to plot the threshold changes at the ends of a series of long current pulses of different amplitude as a current-threshold relationship (e.g., see Fig. 5–6C below). This is conventionally done with threshold increase (hyperpolarization) to the left, threshold decrease (depolarization) to the right, depolarizing current to the top, and hyperpolarizing current to the bottom, to correspond to a conventional I/V plot. Outward rectification resulting from K channel activation causes a steepening of the curve in the top right quadrant, and the steepening of the relationship in the bottom left quadrant indicates inward rectification resulting from activation of IH.

AUTOMATED RECORDING OF MULTIPLE NERVE EXCITABILITY PROPERTIES A considerable spur to the reintroduction of nerve excitability testing into clinical neurophysiology has been the development of a computer program for automatic tracking of threshold changes, and a protocol that sequentially measures a number of different excitability parameters in a motor nerve within a recording period of about 10 minutes44 (Fig. 5–6). The program first measures stimulus-response curves with two different stimulus widths (Fig. 5–6A and B), from which the strength-duration time constant can be estimated for fibers recruited at different fractions of the compound muscle action potential (CMAP) (Fig. 5–6D). A target response of 40% maximal CMAP amplitude is then set, and threshold tracking is implemented, using the slope of the stimulus-response curve to anticipate changes in threshold efficiently. Threshold electrotonus is recorded for

100-ms polarizing currents  40% of the control threshold (Fig. 5–6E), the current-threshold relationship for 200-ms pulses of currents from 50% to 100% of threshold (Fig. 5–6C), and the recovery cycle for interstimulus intervals from 2 to 200 ms (Fig. 5–6F). Many of the excitability recordings referenced in this chapter have been obtained with this convenient protocol. A similar program has been described for measuring excitability properties of sensory axons.48 Unfortunately there is currently no stimulator available commercially for making these automatic measurements, and this has limited the number of centers performing nerve excitability studies.

NERVE EXCITABILITY MEASURES AND MEMBRANE POTENTIAL The extent of the dependence of each of the excitability parameters on membrane potential was recently explored in studies that measured the effect of DC polarizing currents and ischemia on nerve excitability.43 Depolarizing the axonal membrane by 1 mA resulted in significant changes in the relatively refractory period (increased by 37%), supernormality (decreased by 82%), and threshold electrotonus (“fanning in” by 18% in the depolarizing phase and 71% in the hyperpolarizing phase). A smaller change was documented in strength-duration time constant (increased by 5%). These findings illustrate the sensitivity of axonal excitability measures to changes in membrane potential and establish quantitative relationships that have proved useful in the interpretation of the axonal abnormalities recorded in diseases affecting peripheral nerves, such as uremia (see below).

Effects of Ischemia on Axonal Excitability Different disease processes produce focal compression or ischemia and, from a clinical perspective, the changes in excitability associated with ischemia and its release, which indirectly alter resting membrane potential, are more relevant than the “pure” changes in potential induced by polarizing currents. Ischemia is a common cause of peripheral nerve dysfunction, and the excitability changes demonstrated in such studies may well be similar to those responsible for the generation of ectopic activity leading to the symptoms of paresthesias, fasciculation, and myokymia in patients suffering from peripheral nerve disease.15,17,70,72 Experimental models of ischemia have evolved from initial studies on the effects of inflation of a blood pressure cuff around the arm, with the subsequent development of postischemic paresthesias in the digits and palm.61 Ischemia paralyzes energy-dependent processes, particularly the Na,K pump (Na,K-ATPase), and this results in accumulation of Na ions inside the axon and K ions outside the axon and thereby depolarization of sensory and

120

Function of the Peripheral Nervous System

CMAP (% max)

B CMAP (mV)

A 10 1

.1 1

100

0

10 Stimulus (mA)

0

C

100 200 Stimulus (% mean threshold)

D CMAP (% max)

Current (% threshold)

100 0

–100

0 –500 0 Threshold reduction (%)

0

1 Strength-duration time constant (ms)

E

F 100 Threshold change (%)

Threshold reduction (%)

100

0

–100

0

–200 0

100 Conditioning-test delay (ms)

10

200

100

Inter-stimulus interval (ms)

motor axons. There is first an increase and then a reduction in nerve excitability.14,17,43,62,66,72 With release of ischemia, axons rapidly hyperpolarize as a result of a rebound increase in function of the pump. A series of recordings made before (control), during (at 5 and 15 minutes), and after ischemia are illustrated for a single subject in Figure 5–7, using threshold tracking to record multiple measures of axonal excitability. It can be seen that ischemia produces marked and distinctive changes in all measures of axonal excitability—changes consistent with axonal depolarization. The stimulusresponse curve, a nonspecific measure of excitability, shifts to the left as the stimulus current required to evoke the target response becomes less (Fig. 5–7A). The slope of the current-threshold relationship reflects the rectifying properties of the axon, both nodal and internodal, and it becomes steeper as potassium conductances are activated (except with strong hyperpolarizing currents, when the slope is reduced because of a reduction in inward rectification) (Fig. 5–7B). There is a fanning in of threshold

FIGURE 5–6 Multiple nerve excitability measures, recorded with a standard 10-minute protocol. Solid lines and filled circles, recording from a single subject; thin broken lines, means and 95% confidence limits for 29 normal subjects. A, Stimulus-response curves for stimulus durations of 1 ms (thick line) and 0.2 ms (thin line). Filled circle indicates 50% maximum response to 1-ms stimulus, and broken ellipse indicates 95% confidence limits for this point. B, Normalized stimulus-response curves. C, Current-threshold relationship. D, Strengthduration time constants, estimated from data in A for axons recruited at different fractions of the maximum CMAP. E, Threshold electrotonus. F, Recovery cycle. (From Kiernan, M. C., Burke, D., Andersen, K., and Bostock, H.: Multiple measures of axonal excitability: a new approach in clinical testing. Muscle Nerve 23:399, 2000, with permission.)

electrotonus curves as a result of increased resting activation of K channels and short-circuiting of the internodal axon (Fig. 5–7C). The short-circuiting of the internodal axon also reduces superexcitability (Fig. 5–7D), and this, together with increased sodium channel inactivation, prolongs the relative refractory period. Taken together, these changes are those expected from axonal depolarization.43 In some sensory axons, the increased activation of persistent sodium conductances temporarily outweighs the effects of transient sodium channel inactivation and increased potassium conductance, resulting in spontaneous activity and mild, short-lasting paresthesias. In contrast, release of ischemia reverses the changes in excitability described above, such that the axon becomes hyperpolarized14,15,17,43,63,72: the stimulus-response curve shifts to the right (Fig. 5–7A), the current-threshold relationship becomes less steep (Fig. 5–7B), threshold electrotonus curves “fan out” (Fig. 5–7C), and the relative refractory period shortens and gives way to enormous superexcitability (Fig. 5–7D). Postischemic axonal

121

Nerve Excitability Measures

A

B I1

C PI

Current (% threshold)

CMAP (mV)

10

1

.1

100

0

I2 I1 C PI

–100

–200 0

100 200 Conditioning-test delay (ms)

I1 I2

–500 0 Threshold reduction (%)

D Threshold change (%)

C Threshold reduction (%)

C

10 Stimulus (mA)

hyperpolarization is induced by rebound overactivity of the Na,K pump, attempting to normalize the ionic gradients that have developed during the period of ischemia when the pump was paralyzed. The pump is electrogenic and generates a net outward, hyperpolarizing current, because it exports three Na ions for each two K ions imported. While paresthesias experienced during the period of ischemia are typically mild, it is with the release of ischemia, when axons become hyperpolarized, that paresthesias paradoxically become most intense.72 In motor axons, the postischemic hyperexcitability has been attributed to the combination of increased Na,K pump activity and the high level of extracellular K ions accumulated during the paralysis of the pump. The pump is able to make the axonal membrane hyperpolarized relative to the equilibrium potential for K. This reversal of the normal electrochemical gradient for K means that K currents become inward and depolarizing, and, like normal Na currents, they therefore become regenerative because they cause more K channels to open. This results in spontaneous depolarizations toward EK, which produce bursts of action potentials. Following release of ischemia of sufficiently long duration to produce fasciculation, a bimodal distribution of thresholds has been demonstrated for motor axons, with a small population of axons existing in a depolarized state while the majority are hyperpolarized as a result of overactivity of the Na,K pump.15 Small changes in pump current then lead to rapid transition of axons from a depolarized to a hyperpolarized condition, generating spontaneous discharges.

PI –100

1

FIGURE 5–7 The effects of ischemia on nerve excitability properties in a single subject (C  control; I1  5 minutes of ischemia; I2  15 minutes of ischemia; PI  5 minutes postischemia). A, Absolute stimulus-response relationship. B, Current-threshold relationship. C, Threshold electrotonus. D, Recovery cycle. (From Kiernan, M. C., and Bostock, H.: Effects of membrane polarization and ischaemia on the excitability properties of human motor axons. Brain 123:2542, 2000, with permission from Oxford University Press.)

0

100

I2 0

I1 C PI 10 100 Inter-stimulus interval (ms)

Similar studies in sensory axons have failed to document such clear evidence for a bimodal distribution of thresholds in the postischemic state, as was demonstrated for motor axons.17,63 It appears that the high-threshold state is too short lived in sensory axons, because of the greater inward rectification17 (see below).

MODALITY AND REGIONAL DIFFERENCES BETWEEN DIFFERENT AXONS Axons function to transmit an impulse train securely from one end to the other and, presumably, this is done with minimal expenditure of energy. However, the pattern of activity varies widely for axons of different modality. They are subjected to impulse patterns that have very different peak and sustained discharge rates, regularity of discharge, and train length. It might be expected that different axons require slightly different strategies to achieve secure, energy-efficient conduction. In human subjects, there are a number of differences between sensory and motor axons at the same site in the same nerve and between sensory axons innervating the upper and lower limbs. For example, sensory axons in the ulnar nerve17 and in the median nerve63 undergo greater accommodation to hyperpolarizing stimuli than motor axons in the same nerves, presumably reflecting a greater expression of the hyperpolarization-activated cation conductance, IH. Sensory axons in the median nerve have a longer strength-duration time constant71,78 and shorter rheobase than motor axons in the same nerve,71 differences

122

Function of the Peripheral Nervous System

that probably relate to a greater expression of persistent Na conductance (INaP) on sensory axons.20 Sensory axons undergo less refractoriness, supernormality, and late subnormality as they recover excitability following a single discharge.49 Finally, the available evidence suggests that resting membrane potential is similar for median sensory and motor axons,24 but it is possible that median sensory axons are more dependent on the electrogenic Na,K pump to maintain resting membrane potential,63 something that might be necessary to offset their greater persistent Na channel activity.20 Either way, median sensory axons appear to undergo greater ischemic depolarization and a more complex but probably greater postischemic hyperpolarization than median motor axons.63 There are also regional differences between cutaneous afferents in the median and sural nerves that cannot be explained by the slightly smaller size of the sural afferents. For example, median afferents accommodate better to both depolarizing and hyperpolarizing stimuli than do sural afferents, differences that probably reflect greater expression of nodal and internodal slow K conductances and of the internodally located IH on median afferents.65,66 In addition, sural afferents undergo less ischemic depolarization and postischemic hyperpolarization than median afferents, but the precise mechanisms underlying these differences remain uncertain.66 In addition to these differences, there is considerable evidence that the excitability of an axon population changes along its length. For example, the changes in excitability are greater in more proximal segments, and fasciculation occurs more readily following release of ischemia when the site of the ischemia is proximal.14,53 The biophysical basis for this length-dependent difference in axonal excitability has not been the subject of exhaustive study, and the few preliminary studies have not clarified the issue.56,70 The modality-dependent and regional differences in biophysical properties (but not any length-dependent differences) could well represent appropriate differences for the different physiologic roles played by those axons (i.e., an adaptation designed for optimal function). However, this may not be the case when excitability is disturbed and the ability to conduct impulses is jeopardized. The greater expression of two essentially depolarizing conductances (INaP and IH) on sensory than on motor axons would render them more excitable when the nerve is stressed (e.g., during and after ischemia), and it is entirely logical that sensory axons would become ectopically active more readily under such conditions. Conversely, when impulse conduction is jeopardized, hyperpolarizing stresses would be more likely to precipitate conduction failure in motor axons,24 and it is perhaps not surprising that inflammatory demyelinating neuropathies commonly present with predominantly motor deficits even when sensory axons are demonstrably involved. The regional differences in those conductances responsible for accommodation to depolarizing and

hyperpolarizing stimuli imply that median sensory axons are better placed to react to depolarizing and hyperpolarizing stresses than are sural sensory axons. It is conceivable that this could be a factor in the explanation for why the clinical manifestations of many acquired polyneuropathies are more prominent in the distal lower limb.

CHANGES IN AXONAL MEMBRANE POTENTIAL IN PERIPHERAL NEUROPATHIES To date a number of clinical studies have been completed using the new protocol developed to allow rapid measurement of multiple parameters of nerve excitability.44 Studies in patients with uremic neuropathy and with multifocal motor neuropathy (MMN) are described to illustrate the effects that changes in membrane potential can have on axonal function and how these changes can be documented by studies of axonal excitability. This section concludes with a consideration of how physiologic changes in membrane potential can produce conduction block in chronically demyelinated axons.

Multifocal Motor Neuropathy and Axonal Hyperpolarization MMN is a disease process characterized by motor nerve involvement with sparing of sensory function (other than tendon jerks), producing a syndrome in affected patients of slowly progressive muscle atrophy, weakness, and fasciculation.80,82,89,99 A critical diagnostic feature in MMN is the demonstration of conduction block in multiple peripheral nerves on electrophysiologic investigation.98 Conduction block involves only motor axons, with sensory conduction spared across the lesion.42,81 The presence of “positive” symptoms and signs such as cramp, myokymia, and fasciculation in the context of predominantly negative features such as depressed tendon jerks, muscle atrophy, and weakness remains unexplained in these patients. Neither has the mechanism of the conduction block itself been elucidated. Limited information is available from pathologic studies, some of which provide evidence in favor of demyelination.42,81 Axonal properties remote from foci of conduction block are normal,25 findings that support the contention that MMN is not a generalized disease with a focal presentation. However, excitability studies have also been undertaken in MMN patients just distal to the site of conduction block in affected nerves, with strikingly abnormal findings.45 The most significant alterations in excitability properties in the MMN patients were (1) a reduction in minimum slope of the current-threshold (I/V) relationship (indicating a reduction in input conductance); (2) a “fanning out” of threshold

123

Nerve Excitability Measures

despite the changes consistent with hyperpolarization in the other excitability parameters. This discrepancy between ischemia and depolarization in Figure 5–9D presumably reflects the dependence of late subexcitability on (ER  EK), rather than just on ER (see above), and that endoneurial K increases during ischemia. Similarly, the small change in late excitability on release of ischemia probably indicates that the endoneurial fluid becomes hypokalemic as a result of the pump activity. These comparisons suggest that the nerve distal to the site of conduction block in these patients with MMN was hyperpolarized, in a way similar to what occurs in postischemic nerve. For a critical test of whether hyperpolarization was responsible for the abnormal membrane properties, a depolarizing current of 0.5 mA was applied to the nerves of the MMN patients, to test whether depolarization could reverse the abnormalities. This was indeed the case (see Fig. 5–9): All the excitability abnormalities were reduced by depolarization. It is notable that, in these MMN patients, the hyperpolarization seems to represent a stable steady state. Excitability recordings taken 10 weeks apart showed almost identical excitability abnormalities. Persistent overactivity of the Na,K pump would require a persistent intra-axonal source of Na ions to drive the pump. These Na ions could not be continuously entering the axon at the recording site, or there would be a net membrane depolarization.

electrotonus; and (3) an increase in superexcitability recorded during the recovery cycle. The changes in threshold electrotonus and recovery cycle are illustrated for six patients in Figure 5–8A and B. All these excitability parameters, highly abnormal in the MMN patients, depend on the resting conductance of the paranodal and internodal axon membrane. One way that the resting conductance of axonal membrane can be reduced is by membrane hyperpolarization, and the excitability properties in Figure 5–8A and B are very similar to the changes previously seen during hyperpolarization after release from ischemia (see Fig. 5–7C and D). To explore this possibility in more detail, more excitability parameters of the patient nerves were compared to those from polarized and ischemic nerves in healthy subjects43 (Fig. 5–9). For simplicity, the data points plotted are restricted to the mean values for both control groups, and for the effects of 1-mA depolarizing and hyperpolarizing currents, and for recordings started 5 minutes after applying a pressure cuff and 5 minutes after its release. The patient axons behaved like postischemic axons.43 In Figure 5–9D, superexcitability is plotted against late subexcitability. Whereas both hyperpolarization and recovery from ischemia increase superexcitability, late subexcitability is decreased by hyperpolarizing currents but remains unchanged when the nerve is postischemic, and unchanged or even increased in MMN patients,

A

B Threshold change (%)

100

0

–100

0

–200 0

100 Conditioning-test delay (ms)

200

C

10 100 Inter-stimulus interval (ms)

D 100 Threshold change (%)

100 Threshold reduction (%)

FIGURE 5–8 Pathologic changes in nerve excitability measures related to membrane potential. A and B, Threshold electrotonus and recovery cycle recorded distal to the site of conduction block in 6 patients with multifocal motor neuropathy (circles and error bars, means  SEM) compared with 29 normal controls (thin solid and broken lines, means  SEM). (Data from Kiernan, M. C., Guglielmi, J.-M., Kaji, R., et al.: Evidence for axonal membrane hyperpolarization in multifocal motor neuropathy with conduction block. Brain 125:664, 2002.) C and D, Threshold electrotonus and recovery cycle recorded prior to dialysis in 9 patients with chronic renal failure (circles and error bars, means  SEM) compared with 29 normal controls (thin solid and broken lines, means  SEM). (Data from Kiernan, M. C., Walters, J. L., Andersen, K. V., et al.: Nerve excitability changes in chronic renal failure indicate membrane depolarization due to hyperkalaemia. Brain 125:1366, 2002.)

Threshold reduction (%)

10

0

–100

0

–200 0

100 Conditioning-test delay (ms)

200

10 Inter-stimulus interval (ms)

100

124

Function of the Peripheral Nervous System

A

B .5

10

D

RD R

H

I

MD

M

R D

RRP (ms)

Minimum I/V slope

I

RD

PI H

0 0

1 Resting I/V slope

PI

MD M

30

35 Temperature ( C)

C

D 30 H M PI MD D

RD 50

R

Subexcitability (%)

TEd (10–20ms)

100

D 20 M PI

I RD

10

MD

H

R

I 0

–200

–100 TEh (90–100ms)

0

–50 Superexcitability (%)

Longitudinal diffusion or transport along the axon could provide an intra-axonal source of Na ions. Given that the lesion in these patients is focal, Na,K pump activity could be blocked by, for example, edema, or by antibodies directed against the protein components of the pump. In either case, interference with Na,K pump function could produce a depolarizing block with intracellular accumulation of Na at the site of the lesion. Disruption of the blood-nerve barrier42 might increase the potassium concentration in the endoneurial fluid, further aggravating any depolarization block. If Na influx continued, a steady state would be achieved only by Na ions moving intracellularly along the axon to a site where the pump was still working, and this would result in overactivity and membrane hyperpolarization. Therefore, depolarization at the site of the lesion would coexist with chronic hyperpolarization on one or both sides of this site. Such a lesion, with adjacent ischemic and postischemic lengths of nerve, would be likely to generate ectopic activity, a hallmark of MMN, with patients experiencing fasciculation or myokymia in the presence of conduction block. Other clinical symptoms and studies provide further support for this hypothesis. Cold paralysis, described in monomelic amyotrophy,34,47 also occurs in patients with MMN.41 The electrogenic Na,K pump is temperature sensitive, with slower kinetics at lower temperatures. Cooling would exacerbate the effect of the already compromised Na,K pump function, increasing the depolarizing conduc-

FIGURE 5–9 Changes in multiple excitability parameters related to changes in membrane potential. Changes with polarization (H  1-mA hyperpolarizing current; D  1-mA depolarizing current) and ischemia (I  5 minutes of ischemia; PI  5 minutes postischemia) compared with mean data from three MMN patients, including the normalizing effect of depolarization (M  MMN control; MD  MMN with 0.5-mA depolarizing current) and with data from six patients with chronic renal failure (R  before dialysis; RD  after dialysis). A, Minimum versus resting current-threshold slope. B, Relative refractory period versus skin temperature. C, Early depolarizing versus late hyperpolarizing threshold electrotonus. D, Superexcitability versus late subexcitability. Dotted lines and ellipses indicate 95% confidence limits for normal control values.

tion block. This contrasts with the expected effect of cooling in alleviating conduction block caused by demyelination.86,94 Similarly digitalis, a known blocker of Na,K pump activity, can paradoxically exacerbate the fanning out changes seen in threshold electrotonus in MMN.41 Unfortunately, attempts to track excitability changes more proximally toward the site of the focal lesion in patients with MMN have failed because the nerve becomes too inexcitable.105

Uremic Neuropathy and Axonal Depolarization Uremic neuropathy is a distal, symmetrical, mixed sensorimotor, predominantly axonal polyneuropathy affecting legs more than arms.85 The pathologic findings are similar to those in other toxic neuropathies, but the mechanism of nerve damage is unknown. Uremic neuropathy can be minimized by early dialysis68 or reversed by renal transplantation.9,75,77 Because uremic neuropathy improves with dialysis, it is attributed to the accumulation of dialyzable metabolites, but the nature of these uremic toxins remains obscure.10 Substances in the 300 to 5000 range of molecular weights have been implicated on the basis of the supposed superiority of peritoneal dialysis in preventing neuropathy,91 and this “middle molecule” theory3 is still the most popular, although all attempts to identify the putative middle-molecular-weight neurotoxins have been

Nerve Excitability Measures

inconclusive.10,69,85 It has been proposed that uremic neurotoxins cause nerve damage by inhibiting Na,K-ATPase, leading to membrane depolarization,74 because maintenance of a normal membrane potential and ionic gradients is essential for axonal survival.97 In a group of uremic patients treated by hemodialysis, the most striking abnormalities in excitability parameters were revealed by the recordings of threshold electrotonus and recovery cycle (see Fig. 5–8C and D), both properties that are strongly sensitive to membrane potential.51 Axonal excitability was highly abnormal prior to dialysis, with increased refractory period, increased accommodation in threshold electrotonus, and reduced superexcitability. These changes indicate axonal depolarization, and are the opposite of those described above in MMN.45 Hemodialysis produced rapid and significant normalization of these excitability parameters (see Fig. 5–9). Compared with nerves depolarized by applied currents or by ischemia, the excitability changes in chronic renal failure (CRF) were more like those during ischemia, but differences in subexcitability suggested that the increase in endoneurial potassium was more important in CRF than in ischemia (see Fig. 5–9D). The excitability indices of membrane potential in the predialysis recordings were strongly correlated with serum potassium, but not with other electrolytes or serum markers of renal dysfunction. Moreover, CRF patients with highly abnormal serum levels of urea and creatinine, but with normal potassium, had normal nerve excitability properties. These data all pointed to hyperkalemia as the primary cause of membrane depolarization in these subjects. In summary, axons in many patients with CRF appear to be chronically depolarized, and it is likely that this membrane depolarization is due to hyperkalemia.51 It is a reasonable conjecture that this hyperkalemic membrane depolarization contributes to the etiology of uremic neuropathy, but this has not yet been investigated.

Impulse Conduction in Demyelinating Neuropathies Acute demyelination lowers the safety margin for impulse conduction, such that axons can become sensitive to shifts in membrane potential, even when those shifts occur through normal physiologic mechanisms.19 In critically conducting axons, impulse conduction can be impaired by the effects of heating and activity and probably by any mechanism that produces a significant shift in membrane potential, whether depolarizing or hyperpolarizing. Raising temperature by 0.5 °C can be sufficient to precipitate conduction failure in a critically conducting axon,86 and warming commonly accentuates the deficit in multiple sclerosis, even to the point that warming is sometimes used as a provocative test. Conduction failure occurs because warming speeds up channel gating, affecting both activation

125

and inactivation, and this decreases the time integral of the Na current at the node of Ranvier. In a critically conducting axon, the duration of the driving current at the blocking node can reach 1.0 ms,38 and conduction block can be precipitated or relieved by maneuvers that manipulate the time course of the driving current, such as changing temperature or administration of agents that interfere with Na channel inactivation. Activity hyperpolarizes axons. Different mechanisms are responsible for hyperpolarization by short-and long-lasting trains of impulses. With brief, high-frequency trains ( 10 to 20 impulses at 100 to 200 Hz), there is cumulative activation of the nodal slow K conductance.5,8,100 The resulting hyperpolarization can increase threshold by approximately 40%, more than sufficient to jeopardize conduction in impaired axons, but the return to control excitability occurs over 100 to 150 ms,64 such that this mechanism might disrupt the discharge pattern and limit the discharge rate, but it could not produce conduction failure by itself. When an axon conducts long impulse trains, particularly at high frequency, there is an accumulation of Na ions within the axon, and this activates the electrogenic Na,K pump, as occurs on release of ischemia, to restore ionic balance. This activity-dependent hyperpolarization has been demonstrated in human sensory2,50 and motor8,16 axons and, importantly, it can be produced by natural activity.57,101 The extent and duration of the hyperpolarization depend on discharge rate and train length, and can result in an increase in threshold of approximately 40% that takes many tens of minutes to decay to control excitability. Voluntary contractions lasting as little as 15 seconds can increase the threshold of motor axons by 10% to 15%, and this increase decays over some 5 to 10 minutes.101 Such changes are likely to be clinically relevant: A conservative estimate of the safety margin for impulse conduction in a series of patients with chronic inflammatory demyelinating polyneuropathy (CIDP) suggested that significant conduction failure would occur if the axons hyperpolarized by approximately 14%.26 For the same impulse load, the extent of the activitydependent hyperpolarization seems to be greater for motor than for sensory axons101 (M. C. Kiernan, C. S.-Y. Lin, and D. Burke, unpublished findings). An important factor in this difference is probably the difference in IH (see above). In patients with inflammatory demyelinating polyneuropathies, the normal activity-dependent hyperpolarization can precipitate conduction failure at sites of impaired function. This was demonstrated first in MMN40 and subsequently in CIDP.26 There is nothing special about activity: Any process that produces sufficient hyperpolarization will produce clinically significant conduction block at pathologic sites in these disorders, provided that a sufficient number of axons are critically conducting. The release of ischemia results in a postischemic hyperpolarization, and this too can precipitate conduction failure.28 Paradoxically, in CIDP conduction failure may also occur during ischemia, during

126

Function of the Peripheral Nervous System

the depolarizing shift in membrane potential.28 This probably occurs because depolarization inactivates transient Na channels, thus decreasing the availability of functioning channels in axons that are critically dependent on the size of the Na current. It is also possible that ischemia produces an ischemic metabolite that blocks Na channels,62 a mechanism that would further limit the number of Na channels available for the action current. The important message is that critically conducting axons are delicately poised. Conduction may block if membrane potential is too far from threshold (i.e., the axon is hyperpolarized) or if the Na current becomes inadequate (because of heating or because of a limitation on the number of functioning Na channels). Significant changes in activity or significant shifts in membrane potential, whether depolarizing or hyperpolarizing, may be sufficient to produce a transient worsening of symptoms. Clinical fluctuations are well documented in multiple sclerosis; they are as likely in demyelinating polyneuropathies if a sufficient number of axons can only just maintain conduction.

POTENTIAL ROLE FOR EXCITABILITY STUDIES IN THE NEUROPATHY CLINIC Automated nerve excitability testing is still in its infancy, and it is not yet clear whether it will earn a place in the clinic, back alongside nerve conduction studies, which displaced the earlier excitability tests in the 1960s. So far, although many different abnormalities in nerve excitability measures have been found in a range of diseases affecting peripheral nerve, none has been sufficiently sensitive or specific to warrant use as a routine diagnostic tool. Excitability testing is, however, undoubtedly capable of yielding unique information about the biophysical state of nerve fibers, and this information can reasonably be expected to lead to improvements in patient treatment. Although we have concentrated above on the new information it provides about resting membrane potentials, excitability testing can also give warning of a wide variety of other changes in axonal membrane properties. Indeed, measurements of an excitability parameter that is only weakly potential dependent, the strength-duration time constant, have led to the first example of a new treatment initiated as a result of excitability testing. Kanai et al.39 recorded increased strength-duration time constants in Machado-Joseph disease, and this led the authors to introduce treatment with mexiletine to reduce persistent sodium currents. This treatment only partially normalized the strength-duration time constants, but dramatically relieved disabling muscle cramps in eight patients. It is this type of use, as a research tool for comparing groups of patients with defined pathology, in whom excitability testing can provide interesting new insights into pathophysiology, that it has proved most useful so far.

In addition to the changes in membrane potential described above, excitability changes documented in disease include resistance to ischemic depolarization in diabetics102 and a reduction in hyperpolarization-activated current (IH) associated with diabetic neuropathy,37 which in rats is prevented by blocking the polyol pathway with aldose reductase inhibitor.104 Abnormal nerve excitability in amyotrophic lateral sclerosis may involve both reduced K channel and increased persistent Na currents, both leading to membrane instability, but there is a high variability among patients.18,22,37,73 In contrast to nerve conduction studies, demyelination often fails to result in detectable excitability abnormalities,18 either because the demyelination is too distal (see below) or because the density of affected nodes is too low, and affected nodes will have high thresholds and not contribute to the 40% maximal responses that are usually tracked. Thus a study of CIDP patients found high thresholds and prominent changes in stimulus-response curves, but no consistent changes in threshold electrotonus.27 However, it has recently been found that the diffuse demyelination in Charcot-Marie-Tooth disease type 1A does produce characteristic and consistent abnormalities in threshold electrotonus (H. Nodera, S. Kuwabara, and R. Kaji, personal communication). A major limitation of current nerve excitability testing methods is that they are restricted to sites where the peripheral nerve passes close to the skin, such as the ulnar and median nerves at the elbow and wrist. However, in many diseases affecting nerve excitability, it is the most distal portions of the nerve where membrane abnormalities are most pronounced, as inferred from the predominantly distal origin of most ectopic discharges.60 Thus a study of motor axons in neuromyotonia46 failed to find evidence for a membrane abnormality responsible for the hyperexcitability, and a study of Guillain-Barré patients58 found normal membrane properties at the wrist, despite prolonged distal latencies and an increased refractory period of transmission in patients with the axonal form of the disease. Excitability testing is therefore most likely to find clinical application in conditions in which axons are affected rather uniformly along their length, such as in giving early warning of neuropathy in patients being treated with neurotoxic drugs.92

REFERENCES 1. Adrian, E. D., and Lucas, K.: On the summation of propagated disturbances in nerve and muscle. J. Physiol. (Lond.) 44:68, 1912. 2. Applegate, C., and Burke, D.: Changes in excitability of human cutaneous afferents following prolonged high frequency stimulation. Brain 112:147, 1989. 3. Babb, A. L., Ahmad, S., Bergström, J., and Scribner, B. H.: The middle molecule hypothesis in perspective. Am. J. Kidney Dis. 1:46, 1981.

Nerve Excitability Measures 4. Baker, M., and Bostock, H.: Depolarization changes the mechanism of accommodation in rat and human motor axons. J. Physiol. (Lond.) 411:545, 1989. 5. Baker, M., Bostock, H., Grafe, P., and Martius, P.: Function and distribution of 3 types of rectifying channel in rat spinal root myelinated axons. J. Physiol. (Lond.) 383:45, 1987. 6. Baker, M. D.: Axonal flip-flops and oscillators. Trends Neurosci. 25:514, 2000. 7. Barrett, E. F., and Barrett, J. N.: Intracellular recording from vertebrate myelinated axons: mechanism of the depolarizing afterpotential. J. Physiol. (Lond.) 323:117, 1982. 8. Bergmans, J.: The physiology of single human nerve fibres. Louvain, Vander, 1970. 9. Bolton, C. F., Baltzan, M. A., and Baltzan, R. G.: Effects of renal transplantation on uremic neuropathy. N. Engl. J. Med. 284:1170, 1971. 10. Bolton, C. F., and Young, G. B.: Neurological Complications of Renal Disease. Boston, Butterworths, 1990. 11. Bostock, H.: The strength-duration relationship for excitation of myelinated nerve: computed dependence on membrane parameters. J. Physiol. (Lond.) 341:59, 1983. 12. Bostock, H.: Mechanisms of accommodation and adaptation in axons. In Waxman, S. G., Stys, P. K., and Kocsis, J. D. (eds.): The Axon. Oxford, UK, Oxford University Press, p. 311, 1995. 13. Bostock, H., and Baker, M.: Evidence for two types of potassium channel in human motor axons in vivo. Brain Res. 462:354, 1988. 14. Bostock, H., Baker, M., Grafe, P., and Reid, G.: Changes in excitability and accommodation of human motor axons following brief periods of ischaemia. J. Physiol. (Lond.) 441:513, 1991. 15. Bostock, H., Baker, M., and Reid, G.: Changes in excitability of human motor axons underlying post-ischaemic fasciculations: evidence for two stable states. J. Physiol. (Lond.) 441:537, 1991. 16. Bostock, H., and Bergmans, J.: Post-tetanic excitability changes and ectopic discharges in a human motor axon. Brain 117:913, 1994. 17. Bostock, H., Burke, D., and Hales, J. P.: Differences in behaviour of sensory and motor axons following release of ischaemia. Brain 117:225, 1994. 18. Bostock, H., Cikurel, K., and Burke, D.: Threshold tracking techniques in the study of human peripheral nerve. Muscle Nerve 21:137, 1998. 19. Bostock, H., and Grafe, P.: Activity-dependent excitability changes in normal and demyelinated rat spinal root axons. J. Physiol. (Lond.) 365:239, 1985. 20. Bostock, H., and Rothwell, J. C.: Latent addition in motor and sensory fibres of human peripheral nerve. J. Physiol. (Lond.) 498:227, 1997. 21. Bostock, H., Sears, T. A., and Sherratt, R. M.: The spatial distribution of excitability and membrane current in normal and demyelinated mammalian nerve fibres. J. Physiol. (Lond.) 341:41, 1983. 22. Bostock, H., Sharief, M. K., Reid, G., and Murray, N. M. F.: Axonal ion channel dysfunction in amyotrophic lateral sclerosis. Brain 118:217, 1995. 23. Burke, D., Kiernan, M. C., and Bostock, H.: Excitability of human axons. Clin. Neurophysiol. 112:1575, 2001.

127

24. Burke, D., Kiernan, M. C., Mogyoros, I., and Bostock, H.: Susceptibility to conduction block: differences in biophysical properties of cutaneous afferents and motor axons. In Kimura, J., and Kaji, R. (eds.): Physiology of ALS and Related Diseases. Amsterdam, Elsevier Science, p. 43, 1997. 25. Cappelen-Smith, C., Kuwabara, S., Lin, C. S.-Y., and Burke, D.: Abnormalities of axonal excitability are not generalized in early multifocal motor neuropathy. Muscle Nerve 26:769, 2002. 26. Cappelen-Smith, C., Kuwabara, S., Lin, C. S.-Y., et al.: Activity-dependent hyperpolarization and conduction block in chronic inflammatory demyelinating polyneuropathy. Ann. Neurol. 48:826, 2000. 27. Cappelen-Smith, C., Kuwabara, S., Lin, C. S.-Y., et al.: Membrane properties in chronic inflammatory demyelinating polyneuropathy. Brain 124:2439, 2001. 28. Cappelen-Smith, C., Lin, C. S., Kuwabara, S., and Burke, D.: Conduction block during and after ischaemia in chronic inflammatory demyelinating polyneuropathy. Brain 125:1850, 2002. 29. Davis, H., and Forbes, A.: Chronaxie. Physiol. Rev. 16:407, 1936. 30. Fowler, C.: Early history of nerve conduction studies and electromyography. In Osselton, J. W. (ed.): Clinical Neurophysiology. Oxford, UK, Butterworth-Heinemann, p. 45, 1995. 31. Frankenhaeuser, B., and Huxley, A. F.: The action potential in myelinated nerve fibres of Xenopus laevis as computed on the basis of voltage clamp data. J. Physiol. (Lond.) 171:302, 1964. 32. Gilliatt, R. W., and Willison, R. G.: The refractory and supernormal periods of the human median nerve. J. Neurol. Neurosurg. Psychiatry 26:136, 1963. 33. Hill, A. V.: Excitation and accommodation in nerve. Proc. R. Soc. B 119:305, 1936. 34. Hirayama, K., Tsubaki, T., Toyokura, Y., and Okinaka, S.: Juvenile muscular atrophy of unilateral upper extremity. Neurology 13:373, 1963. 35. Hodes, R., Larrabee, M.G., and German, W.: The human electromyogram in response to nerve stimulation and the conduction velocity of motor axons. Arch. Neurol. Psychiatry 60:240, 1948. 36. Hodgkin, A.L., and Huxley, A.F.: A quantitative description of membrane current and its application to conduction and excitation in nerve. J. Physiol. (Lond.) 117:500, 1952. 37. Horn, S., Quasthoff, S., Grafe, P., et al.: Abnormal axonal inward rectification in diabetic neuropathy. Muscle Nerve 19:1268, 1996. 38. Inglis, J. T., Leeper, J. B., Wilson, L. R., et al.: The development of conduction block in single human axons following a focal nerve injury. J. Physiol. (Lond.) 513:127, 1998. 39. Kanai, K., Kuwabara, S., Arai, K., et al.: Muscle cramp in Machado-Joseph disease: altered motor axonal excitability properties and mexiletine treatment. Brain 126(Pt. 4):965, 2003. 40. Kaji, R., Bostock, H., Kohara, N., et al.: Activity-dependent conduction block in multifocal motor neuropathy. Brain 123:1602, 2000.

128

Function of the Peripheral Nervous System

41. Kaji, R., and Kojima, Y.: Pathophysiology and clinical variants of multifocal motor neuropathy. In Kimura, J., and Kaji, R. (eds.): Physiology of ALS and Related Diseases. Amsterdam, Elsevier, p. 85, 1997. 42. Kaji, R., Oka, N., Tsuji, T., et al.: Pathological findings at the site of conduction block in multifocal motor neuropathy. Ann. Neurol. 33:152, 1993. 43. Kiernan, M. C., and Bostock, H.: Effects of membrane polarization and ischaemia on the excitability properties of human motor axons. Brain 123:2542, 2000. 44. Kiernan, M. C., Burke, D., Andersen, K., and Bostock, H.: Multiple measures of axonal excitability: a new approach in clinical testing. Muscle Nerve 23:399, 2000. 45. Kiernan, M. C., Guglielmi, J.-M., Kaji, R., et al.: Evidence for axonal membrane hyperpolarization in multifocal motor neuropathy with conduction block. Brain 125:664, 2002. 46. Kiernan, M. C., Hart, I. K., and Bostock, H.: Excitability properties of motor axons in patients with spontaneous motor activity. J. Neurol. Neurosurg. Psychiatry 70:56, 2001. 47. Kiernan, M. C., Lethlean, A. K., and Blum, P. W.: Monomelic amyotrophy: juvenile muscular atrophy of the upper limb. J. Clin. Neurosci. 6:353, 1999. 48. Kiernan, M. C., Lin, S. Y., Andersen, K. V., et al.: Clinical evaluation of excitability measures in sensory nerve. Muscle Nerve 24:883, 2001. 49. Kiernan, M. C., Mogyoros, I., and Burke, D.: Differences in the recovery of excitability in sensory and motor axons of human median nerve. Brain 119:1099, 1996. 50. Kiernan, M. C., Mogyoros, I., Hales, J. P., et al.: Excitability changes in human cutaneous afferents induced by prolonged repetitive axonal activity. J. Physiol. (Lond.) 500:255, 1997. 51. Kiernan, M. C., Walters, J. L., Andersen, K. V., et al.: Nerve excitability changes in chronic renal failure indicate membrane depolarization due to hyperkalaemia. Brain 125:1366, 2002. 52. Kimura, J.: Refractory period measurements in the clinical domain. In Waxman, S. G., and Ritchie, R. M. (eds.): Demyelinating Disease: Basic and Clinical Electrophysiology. New York, Raven Press, p. 239, 1981. 53. Kugelberg, E.: Accommodation in human nerves and its significance for the symptoms in circulatory disturbances and tetany. Acta Physiol. Scand. 8(Suppl. 24):1, 1944. 54. Kugelberg, E.: Neurologic mechanism for certain phenomena in tetany. Arch. Neurol. Psychiatry 56:507, 1946. 55. Kugelberg, E.: Activation of human nerves by hyperventilation and hypocalcemia: neurologic mechanism of symptoms of irritation in tetany. Arch. Neurol. Psychiatry 60:153, 1948. 56. Kuwabara, S., Cappelen-Smith, C., Lin, C. S.-Y., et al.: Differences in accommodative properties of median and peroneal motor axons. J. Neurol. Neurosurg. Psychiatry 70:372, 2001. 57. Kuwabara, S., Lin, S.-Y., Mogyoros, I., et al.: Voluntary contraction impairs the refractory period of transmission in healthy human axons. J. Physiol. 531:265, 2001. 58. Kuwabara, S., Ogawara, K., Sung, J. Y., et al.: Differences in membrane properties of axonal and demyelinating GuillainBarre syndromes. Ann. Neurol. 52:180, 2002.

59. Lapique, L.: l’Excitabilité en Fonction du Temps: La Chronaxie, Sa Signification et Sa Mesure. Paris, Presses Univ. de Paris, 1926. 60. Layzer, R. B.: The origin of muscle fasciculations and cramps. Muscle Nerve 17:1243, 1994. 61. Lewis, T., Pickering, G. W., and Rothschild, P.: Centripetal paralysis arising out of arrested blood flow to the limb, including notes on a form of tingling. Heart 16:1, 1931. 62. Lin, C. S.-Y., Grosskreutz, J., and Burke, D.: Sodium channel function and the excitability of human cutaneous afferents during ischaemia. J. Physiol. (Lond.) 538:435–446, 2002. 63. Lin, C. S.-Y., Kuwabara, S., Cappelen-Smith, C., and Burke, D.: Responses of human sensory and motor axons to the release of ischaemia and to hyperpolarizing currents. J. Physiol. (Lond.) 541:1025, 2002. 64. Lin, C. S.-Y., Mogyoros, I., and Burke, D.: Recovery of excitability of cutaneous afferents in the median and sural nerves following activity. Muscle Nerve 23:763, 2000. 65. Lin, C. S.-Y., Mogyoros, I., Kuwabara, S., et al.: Accommodation to depolarizing and hyperpolarizing current in cutaneous afferents of the human median and sural nerves. J. Physiol. (Lond.) 529:483, 2000. 66. Lin, C. S.-Y., Mogyoros, I., Kuwabara, S., et al.: Differences in responses of cutaneous afferents in the human median and sural nerves to ischemia. Muscle Nerve 24:1503, 2001. 67. Lucas, K.: The excitable substance of amphibian muscle. J. Physiol. (Lond.) 36:113, 1906. 68. Manis, Y., and Friedman, E. A.: Dialytic therapy for irreversible uremia. N. Engl. J. Med. 301:1260, 1979. 69. Merrill, J. P.: The search for ‘factor X’. Clin. Nephrol. 11:56, 1979. 70. Mogyoros, I., Bostock, H., and Burke, D.: Mechanisms of paresthesias arising from healthy axons. Muscle Nerve 23:310, 2000. 71. Mogyoros, I., Kiernan, M. C., and Burke, D.: Strengthduration properties of human peripheral nerve. Brain 119:439, 1996. 72. Mogyoros, I., Kiernan, M. C., Burke, D., and Bostock, H.: Excitability changes in human sensory and motor axons during hyperventilation and ischaemia. Brain 120:317, 1997. 73. Mogyoros, I., Kiernan, M. C., Burke, D., and Bostock, H.: Strength-duration properties of sensory and motor axons in amyotrophic lateral sclerosis. Brain 121:851, 1998. 74. Nielsen, V. K.: The peripheral nerve function in chronic renal failure. V. Sensory and motor conduction velocity. Acta Med. Scand. 194:445, 1973. 75. Nielsen, V. K.: The peripheral nerve function in chronic renal failure. VIII. Recovery after renal transplantation: clinical aspects. Acta Med. Scand. 195:163, 1974. 76. Noble, D., and Stein, R. B.: The threshold conditions for initiation of action potentials by excitable cells. J. Physiol. (Lond.) 187:129, 1966. 77. Oh, S. J., Clemens, R. S. Jr., Lee, Y. W., and Diethelm, A. G.: Rapid improvement in nerve conduction velocity following renal transplantation. Ann. Neurol. 4:369,1978. 78. Panizza, M., Nilsson, J., Roth, B. J., et al.: The time constants of motor and sensory peripheral nerve fibers measured with the method of latent addition. Electroencephalogr. Clin. Neurophysiol. 93:147, 1994.

Nerve Excitability Measures 79. Pape, H. C.: Queer current and pacemaker: the hyperpolarization-activated cation current in neurons. Annu. Rev. Physiol. 58:299, 1980. 80. Parry, G. J.: Pure motor neuropathy with multifocal conduction block masquerading as motor neuron disease. Muscle Nerve 11:103, 1988. 81. Parry, G. J.: Multifocal motor neuropathy: pathology and treatment. In Kimura, J., and Kaji, R. (eds.): Physiology of ALS and Related Diseases. Amsterdam, Elsevier, p. 73, 1997. 82. Pestronk, A., Cornblath, D. R., Ilyas, A. A., et al.: A treatable multifocal motor neuropathy with antibodies to GM1 ganglioside. Ann. Neurol. 24:73, 1988. 83. Purves-Stewart, J.: The Diagnosis of Nervous Diseases, 9th ed. London, Edward Arnold, 1945. 84. Purves-Stewart, J., and Worster-Drought, C.: The Diagnosis of Nervous Diseases, 10th ed. London, Edward Arnold, 1952. 85. Raskin, R.: Neurological complications of renal failure. In Aminoff, M. J. (ed.): Neurology and General Medicine. New York, Churchill Livingstone, 2001. 86. Rasminsky, M.: The effects of temperature on conduction in demyelinated single nerve fibres. Arch. Neurol. 28:287, 1973. 87. Ritchie, A. R.: The electrical diagnosis of peripheral nerve injury. Brain 67:314, 1944. 88. Ritchie, J. M.: Physiology of axons. In Waxman, S. G., Stys, P. K., and Kocsis, J. D. (eds.): The Axon. Oxford, UK, Oxford University Press, p. 68, 1995. 89. Roth, G., Rohr, J., Magistris, M. R., and Ochsner, F.: Motor neuropathy with proximal multifocal persistent conduction block, fasciculations and myokymia: evolution to tetraplegia. Eur. Neurol. 25:416, 1986. 90. Rushton, W. A. H.: The time factor in electrical excitation. Biol. Rev. 10:1, 1935. 91. Scribner, B. H.: Discussion. Trans. Am. Soc. Artif. Intern. Organs 11:29, 1965. 92. Schilling, T., Heinrich, B., Kau, R., et al.: Paclitaxel administered over 3h followed by cisplatin in patients with advanced head and neck squamous cell carcinoma: a clinical phase I study. Oncology 54:89, 1997. 93. Schwarz, J. R., Reid, G., and Bostock, H.: Action potentials and membrane currents in the human node of Ranvier. Pflügers Arch. 430:283, 1995.

129

94. Sears, T. A., and Bostock, H.: Conduction failure in demyelination: is it inevitable? Adv. Neurology 31:357, 1981. 95. Stephanova, D. I., and Bostock, H.: A distributed-parameter model of the myelinated human motor nerve fibre: temporal and spatial distributions of electrotonic potentials and ionic currents. Biol. Cybernet. 74:543, 1996. 96. Stys, P. K., and Ashby, P.: An automated technique for measuring the recovery cycle of human nerves. Muscle Nerve 17:969, 1994. 97. Stys, P. K., Ransom, B. R., Black, J. A., and Waxman, S. G.: Anoxic/ischemic injury in axons. In Waxman, S. G., Kocsis, J. D., and Stys, P. K. (eds.): The Axon. Oxford, UK, Oxford University Press, p. 462, 1995. 98. Sumner, A.: Consensus criteria for the diagnosis of partial conduction block and multifocal motor neuropathy. In Kimura, J. and Kaji, R. (eds.) Physiology of ALS and Related Diseases. Amsterdam, Elsevier, p. 221, 1997. 99. Taylor, B. V., Wright, R. A., Harper, C. M., and Dyck, P. J.: Natural history of 46 patients with multifocal motor neuropathy with conduction block. Muscle Nerve 23:900, 2000. 100. Taylor, J. L., Burke, D., and Heywood, J.: Physiological evidence for a slow K conductance in human cutaneous afferents. J. Physiol. (Lond.) 453:575, 1992. 101. Vagg, R., Mogyoros, I., Kiernan, M. C., and Burke, D.: Activity-dependent hyperpolarization of human motor axons produced by natural activity. J. Physiol. (Lond.) 507:919, 1998. 102. Weigl, P., Bostock, H., Franz, P., et al.: Threshold tracking provides a rapid indication of ischaemic resistance in motor axons of diabetic subjects. Electroencephalogr. Clin. Neurophysiol. 73:369, 1989. 103. Weiss, G.: Sur la possibilité de rendre comparable entre eux les appareils servant à l’excitation électrique. Arch. Ital. Biol. 35:413, 1901. 104. Yang, Q., Kaji, R., Takagi, T., et al.: Abnormal axonal inward rectifier in streptozotocin-induced experimental diabetic neuropathy. Brain 124:1149, 2001. 105. Yokota, T., Saito, Y., Yuki, N., and Tanaka, H.: Persistent increased threshold of electrical stimulation selective to motor nerve in multifocal motor neuropathy. Muscle Nerve 19:823, 1996.

6 The Muscle Spindle ROBERT W. BANKS

Muscle Spindles of the Cat Soleus Abundance of Muscle Spindles and Afferent Axons

Development of Muscle Spindles Intrafusal Muscle Fibers

One of the primary functions of the peripheral nervous system is the execution of motor commands issued by the brain and spinal cord to the skeletal muscles. The complexity of this task is such that the central nervous system requires a constant stream of information about the kinematic and dynamic states of its muscles, enabling them to be continuously modified so as to complete a command successfully in a wide variety of circumstances. There are two principal kinds of encapsulated sense organ in skeletal muscles that are responsible for providing this information: the tendon organ, which senses the force of contraction (see Chapter 7); and the muscle spindle, which responds to muscle length. The afferent axons that supply their sensory endings are among the largest (and therefore most rapidly conducting) in the body. Muscle spindles and tendon organs together account for virtually all group I axons (conventionally differentiated as Ia and Ib), and muscle spindles alone account for a substantial proportion of group II axons. Ruffini’s morphologic categories of primary and secondary endings99 have long been identified as the intrafusal terminals of Ia and II afferents, respectively. Furthermore, unlike the passive tendon organ, the responses of the muscle spindle’s sensory endings are modifiable by motor input supplied by an exclusive system of small (gamma) motoneurons, as well as collateral branches of alpha motoneurons. Among the various sense organs, therefore, the eye and the ear are the only ones to exceed the muscle spindle in structural and functional complexity. Thus, although at one level it is possible to state quite simply that the function of the muscle spindle is to act as a length sensor, and this will be true wherever one occurs, at the higher level of integrated motor control systems,

Sensory Innervation Motor Innervation

such as the cyclic movement of the legs in walking or running, there is no statement that is both simple and general to describe the muscle spindle’s function. What is clear, however, is that the information required of the muscle spindle is put to various uses according to the nature of the particular motor task at issue. For readers who wish to explore further the higher level function of the muscle spindle in motor control, the recent reviews by Hultborn61 and McCrea84 provide an entry to the extensive literature. This chapter concerns the structure and function of the mammalian muscle spindle, primarily considered at the level of the individual length sensor, and focusing on the neural components. However, it also examines quantitative aspects of sensory provision in different muscles, specifically the relative abundance of muscle spindles and the richness of their afferent nerve supply. A significant proportion of the myelinated nerve fibers in a peripheral motor nerve, including the largest of these, are involved directly in motor control (as distinct from motor output) and thus constitute an expensive necessity. Consideration of adaptive processes in evolution leads us to suppose that the sensory complement of a particular muscle will tend towards an optimum in which there is a balance between the cost of its provision and the benefit gained in terms of motor control. If suitable, unbiased, measures can be found, it should be possible to obtain comparative data useful in modeling different motor control strategies. Inevitably, most of our information about mammalian muscle spindles has been derived from a very few species, including rat, mouse, and human, but in particular the cat. The muscle spindles of these species share a basic 131

132

Function of the Peripheral Nervous System

organization that may be summarized as follows: (1) with very few exceptions, each muscle spindle consists of three types of specialized muscle fiber forming a small bundle; (2) these intrafusal fibers are fully cross-striated for most of their lengths, except for the segments contacted by the sensory endings, where groups of euchromatic nuclei occur (in the equator, or equatorial region); (3) there is a variable number of sensory endings, at least one of which is always distinguished as a primary ending; (4) the striated portions of the intrafusal fibers on either side of the sensory endings (the poles, or polar regions) are innervated by a very variable motor supply; and (5) most of the innervated part of the muscle spindle is enclosed by a multilayered capsule that is expanded in the region of the sensory endings, so as to enclose the sensory endings in a gel-filled periaxial space whose contents are further individually wrapped by fibroblast-like inner capsule cells. Free afferents and a nonsomatic motor (autonomic) innervation sometimes occur in muscle spindles, but their sporadic appearance shows that they are not essential for spindle function. Neither they, nor the capsule, is considered further in the present chapter; additional information on these topics may be found in Banks and Barker.9 Some information is available on the structure of muscle spindles from a wide variety of mammalian species, often in the older literature that has been thoroughly reviewed by Barker19 and classified by Eldred et al.46,47 These descriptions are consistent with the idea that the basic organization given above is common to all mammals, though the number of types of intrafusal fiber (initially in rat, rabbit, cat, and monkey) was not finally established until the 1970s.14,89

MUSCLE SPINDLES OF THE CAT SOLEUS With this essential background in place, we may now proceed to take a particular muscle as a paradigm and give a fuller description of its complement of muscle spindles. We may take the soleus of the cat, one of very few muscles known in sufficient detail, for this purpose. The soleus is the deepest head of the ankle extensor, or triceps surae; its accessibility, simple structure, and homogeneous composition make it especially suitable for physiologic study. In the adult cat, the soleus weighs just over 4 g and is almost 85 mm long.100 Apart from the specialized intrafusal fibers, the muscle is made up entirely of slow-twitch (type S) fibers.34 These constitute the great bulk of the muscle, numbering 22,000 to 30,00037; nevertheless, it is convenient to adopt the usual fusicentric epithet of extrafusal to describe them. The extrafusal fibers have a mean length of about 42 mm,100 giving the muscle an index of architecture of 0.49 in a unipennate arrangement. There are, on average, 56 muscle spindles in the cat soleus36 (Fig. 6–1A). Individual spindles vary consider-

ably in size and complexity while usually conforming to the basic organization summarized above. This is illustrated by a representative sample of six muscle spindles shown in Figure 6–1B to F, all of which are derived from a single soleus muscle. Two of them (Fig. 6–1B) are linked end to end in tandem fashion, though only one of the intrafusal fibers (the bag2; see below) is continuous through both spindles. Their combined length overall is 10.2 mm. The remaining separate spindles (Fig. 6–1C to F) range in length from 5.3 to 8.5 mm, with a mean of 7.5 mm. The locations of the sensory endings are marked alongside the expanded portions of the capsules: three spindles (Fig. 6–1C, F, and the smaller of the tandem pair in B) each contain just a single primary (P) ending; two each contain a primary and one secondary (S1) ending (Fig. 6–1D and the larger of the tandem pair in B); and one (Fig. 6–1E) contains a primary and two secondary endings. In this case both secondary endings lie on the same side of the primary and are thus designated S1 and S2, but in general two or more secondary endings may be located on either or both sides of the primary, apparently at random. The three types of intrafusal fiber are designated as bag1, bag2 and chain.14,89 Their properties and distinguishing features are described below; for the moment we note that, in the sample shown in Figure 6–1, the intrafusal bundle of each muscle spindle consists of one bag1, one bag2, and several chain fibers, except for the smaller of the tandem pair (Fig. 6–1B), which lacks a bag1 fiber. We may thus conveniently distinguish b1b2c spindles from b2c spindles, the proportions of the two kinds differing according to muscle type.18 The lengths of the muscle spindles given above usually correspond to those of one or other bag fiber, typically the bag2. The encapsulated portion of the muscle spindle is often considerably shorter, and both bag fibers usually extend beyond the ends of the capsule. In the four separate muscle spindles of Figure 6–1, the mean length of the capsules is 3.9 mm; one spindle (Fig. 6–1F) has a particularly long extracapsular pole, whereas another (Fig. 6–1C) has one very short pole that inserts directly into tendon (note the tendon organ alongside) without any extracapsular portion whatever. Despite the similarity in the complements of intrafusal fibers of most soleus muscle spindles, the pattern of motor innervation varies widely. Two examples from the sample of Figure 6–1 are shown in greater detail in Figure 6–2 to illustrate the range from simple to complex, without in either case marking the extremes of variability. The muscle spindle shown in Figures 6–1F, 6–2B, and 6–2C is the simpler of the two. Its entire somatic innervation is enclosed within the capsule, as is commonly the case. It is supplied by a single small nerve, which contains four separate motor axons in addition to the Ia afferent. In a remarkably symmetrical arrangement, each pole receives two motor axons, one ending on the bag1 fiber in a single

The Muscle Spindle

133

FIGURE 6–1 Muscle spindles in the soleus muscle of the cat. A, Projection plan of a soleus muscle of a kitten, reconstructed from serial longitudinal sections. Lines extending between the origin (o) and insertion (i) indicate the unipennate fascicular structure. Muscle spindles (dots show the locations of their equatorial regions) are predominantly distributed close to the intramuscular branches of the nerve (n). (Modified from Chin, N. K., Cope, M., and Pang, M.: Number and distribution of spindle capsules in seven hindlimb muscles of the cat. In Barker, D. [ed.]: Symposium on Muscle Receptors. Hong Kong, Hong Kong University Press, p. 241, 1962.) B through F, Examples of whole muscle spindles, stained with silver nitrate and teased from a single soleus muscle of an adult cat, showing the locations of their primary (P) and secondary (S1, S2) sensory endings. Note that one spindle is closely associated with a tendon organ (t o). Curvature of the spindles is artifactual, resulting from the teasing process.

large end plate (Fig. 6–2Bii and Biv), and the other, after branching repeatedly, ending on the bag2 and the chain fibers (Fig. 6–2Bi and Biii). Further enlargement (Fig. 6–2C) of the axon shown in Figure 6–2Biii allows its mostly myelinated preterminal branches (blue) to be differentiated from the terminal branches (red) that constitute the separate endings. The endings of this one axon are themselves of very variable form, ranging from large, well-

defined end-plate–like structures in more polar locations, such as those on the bag2, to structures close to the sensory ending of such simplicity that it is impossible to be sure whether they are in neuroeffective contact with the associated chain fibers. In the more complex muscle spindle (Figs. 6–1E and 6–2A), the innervation of the encapsulated part enters the spindle via five small nerves. Progressing from pole to

134

Function of the Peripheral Nervous System

FIGURE 6–2 A and B, The complete innervation of two soleus muscle spindles, as revealed by silver staining and teasing. (The complete spindles are shown in Figure 6–1E and 6–1F, respectively.) In each case (Ai to Ax and Bi to Biv), the intrafusal branching and motor endings of every motor axon entering the spindles are shown at larger scale (calibration refers to these details). Six axons (Aiii, Avii, Aviii, Ax, Bii, and Biv) supply the bag1 fibers exclusively, two supply chain fibers alone (Aiv and Avi), and two (Av and Aix) supply a bag2 fiber alone. Bag2 and chain fibers are innervated together by three axons (Aii, Bi, and Biii). The small endings shown in Ai are located on a bag1 fiber and an unusually long chain fiber, but whether they were innervated by a single axon or by two separate axons is unknown. C, Further enlargement of the intrafusal branches (blue) and motor endings (red) of the axon Biii, which provides the entire motor innervation of the bag2 and chain fibers in one pole of spindle B.

pole, and beginning with that which contains the two secondary endings, we see the following axonal distributions in each small nerve: (1) two unbranched motor axons supplying separate endings to the bag1 fiber (Fig. 6–2Ax) and bag2 (Fig. 6–2Aix) fibers; (2) a group II afferent supplying the S2 sensory ending together with three motor axons, two of which are unbranched and provide separate endings to the bag1 fiber (Fig. 6–2Avii and Aviii), whereas the third branches to some three separate endings, all on the bag2 fiber (Fig. 6–2Av); (3) a group II afferent supplying the S1 sensory ending together with two motor axons, both ending only on chain fibers, one after more extensive branching (Fig. 6–2Aiv) than the other (Fig. 6–2Avi); (4) the group Ia afferent alone entering the equatorial region; and finally (5) two motor axons, one unbranched to a single large ending on the bag1 fiber (Fig. 6–2Aiii) and the other

much branched and supplying several endings of varying form to the bag2 and chain fibers (Fig. 6–2Aii). Thus, of the 13 motor axons that innervate the encapsulated portions of the two spindles, 6 supply the bag1 fibers exclusively, 2 supply a bag2 and 2 some chain fibers separately (all in the same pole), and each of the remaining 3 supply both bag2 and a group of chain fibers. The more complex muscle spindle has, in addition, a small group of endings in the extracapsular portion of one pole (Fig. 6–2Ai). They resemble, in form and size, the end plates of extrafusal fibers and are likely to be derived by collateral branching of one or two axons with an otherwise skeletomotor distribution. In this case they are located on the bag1 fiber and an unusually long and wide chain fiber, but the supplying axons are broken off close to the terminals.

The Muscle Spindle

The number of chain fibers present in each intrafusal bundle is not readily discernible in teased, whole-muscle spindles, but according to the data in Table 4 of Boyd,28 the mean number of chain fibers in soleus muscle spindles is 4.4 and the mean total number of intrafusal fibers is 6.4. In the complete muscle there are thus about 360 (56  6.4) intrafusal fibers in total. In round figures, the nerve to the soleus contains 450 myelinated axons of which perhaps 180 are sensory and 270 are motor.31,88 Most of the afferents can be accounted for by the 56 muscle spindles and 45 tendon organs,103 with the spindles receiving some 70 group II afferents (estimated from data in Banks and Stacey18); the balance would be made up of free-ending and paciniform afferents.19 Boyd and Davey31 estimated that some 115 of the motor axons are gamma efferents and therefore purely fusimotor in distribution. Thus the 360 intrafusal fibers are innervated exclusively by about 125 afferent and 115 efferent axons, or 53% of the myelinated fibers, whereas the (say) 25,000 extrafusal fibers are innervated by 155 alpha motoneurons, representing only 34% of all myelinated fibers in the nerve. Moreover, an unknown proportion (in this muscle, but see below) of the alpha efferents would in fact have a skeletofusimotor distribution, further increasing the numerical dominance of the control over the output systems in the innervation of the soleus.

ABUNDANCE OF MUSCLE SPINDLES AND AFFERENT AXONS Now that we have taken the soleus of the cat as our prime example, we must inquire how typical is its complement of muscle spindles, and the richness of their innervation, in comparison with other mammalian muscles. Of course, different exemplars of a particular muscle exhibit variability in the number of muscle spindles that they contain; nevertheless, it is clear that each muscle has a characteristic complement.36 The coefficient of variation for the number of muscle spindles in soleus of the cat is 13% (56  7 [mean  standard deviation]36). It so happens that the mean coefficient of variation was also 13% for a population of one forelimb and nine hind limb muscles of the cat, in which the mean number of muscle spindles ranged from 25 (interosseus manus V, 0.21 g) to 114 (semitendinosus, 6.41 g).36 Perhaps not surprisingly, therefore, it appears that larger muscles may possess absolutely more muscle spindles than smaller ones. But are muscle spindles relatively more abundant in larger or smaller muscles, if, indeed, there are any differences? For many authors, the answer is simply provided by dividing the number of muscle spindles by the adult weight (or mass) of the muscle to give a parameter known, following Voss,106 as spindle frequency or spindle density. By this measure, smaller muscles tend to have higher spindle densities than larger ones, so, for

135

example, the spindle densities of the human lumbricalis manus III and the gastrocnemius (both medial and lateral heads together) are 12.2 and 0.4, respectively.18 This seemingly large disparity is often taken to reflect the different functional roles of the lumbricalis in fine control of hand movements and of the gastrocnemius in gross movement of the whole body in locomotion, on the intuitive assumption that the former requires more length feedback control than the latter. Nevertheless, it is also clear that there is a nonlinear relationship between the number of muscle spindles in a particular muscle and the mass of the muscle. This prevents a straightforward comparison of the spindle densities of muscles of different sizes, even those belonging to a single species.18 That there may be some more fundamental relationship involved is indicated by comparison of the human data for the lumbricalis manus III and gastrocnemius with the corresponding data for the cat. Here the spindle densities are 173 and 6.5, respectively, although the cat is not noted for manual dexterity. But notice that both these spindle densities are about 15 times those of the human muscles. Encouraged by such observations, Banks and Stacey,18 using an allometric approach, found a linear regression relationship between the logarithms of spindle number and muscle mass. The sample consisted mainly of hind limb muscles and included most of the data then available from the rat and cat, together with an arbitrarily chosen sample of human muscles. The slope of the regression line was 0.32, a value geometrically significant as essentially the same as that to be expected if the number of muscle spindles were scaling isometrically with a parameter of length (i.e., proportional to the cube root of muscle mass, assuming constant physical density). A similar regression slope has been found for the soleus muscle alone, with data from the mouse, rat, cat, and human,8 indicating that the isometric relationship may be a common feature of homologous muscle series. Of course a regression line is a statistical abstraction, and individual muscles deviate more or less from it, but since the deviation is characteristic of a given muscle, it is presumably of functional significance, and its value can be taken as a measure of the relative abundance of spindles in that particular muscle. At first it appeared as though this musclespecific variability might show a regression against size for the different muscles of a single species similar to that for homologous muscles of different species,18 but it is now clear that it does not.8 Voss107 has tabulated an almost complete set of data for the human musculature, allometric analysis of which8 again revealed a linear regression between the logarithms of numbers of muscle spindles and muscle masses, but with a slope of 0.5. Comparison of a group of homologous feline and human muscles further indicated that a similar relationship may occur in the cat and, by extension, other mammals. Isometric scaling therefore applies only to comparisons between species; within a species, the scaling effect of muscle size appears to be

136

Function of the Peripheral Nervous System

a square-root function. Where does this leave the relative abundance of muscle spindles in human lumbricalis manus III and gastrocnemius? If we express the relative abundance as the observed number of spindles divided by the number expected from the regression relationship, that of lumbricalis manus III is 1.01 and of gastrocnemius is 0.41, so the lumbricalis emerges as almost precisely average, whereas the gastrocnemius is clearly rather poorly supplied with spindles for a muscle of its size. The muscle with the greatest relative abundance is the longissimus capitis (8.94) and that with the least is the digastric (0.13, ignoring the only two muscles recorded by Voss107 as possessing no spindles). The soleus, incidentally, has a relative abundance of 0.99. The sensory complement of the individual spindles of a single muscle varies widely, as is clear even from the small sample of Figure 6–1. In the cat soleus it has been found to range from a single primary up to a primary and four secondaries, all supplied by separate afferent axons.12 However, the overall population of spindles of a particular muscle has a characteristic pattern of sensory innervation, as shown by a specific distribution of the frequency of occurrence of the various individual complements,18 For the b1b2c spindles from a sample of 11, mostly hind limb, muscles of the cat, the distributions were adequately described by binomial statistics, the parameters of which (n, p) varied independently among the heteronymous muscles.18 Thus n varied from 3 (superficial and deep lumbrical muscles, and interosseus) to 22 (extensor digitorum longus), whereas p varied from approximately 0.07 (extensor digitorum longus) to 0.58 (complexus), for the theoretical distributions that best fitted the observed distributions. The values for soleus were n  10 and p ⬵ 0.14, corresponding to (and in part derived from) the observed mean number of 1.36 afferents per b1b2c spindle18 that were additional to the single Ia that is both necessary and sufficient for the spindle’s initial differentiation. In cat spindles these additional afferents usually form secondary endings, whereas in the rat masseter many of them contribute to multiply innervated primary endings.11 The proportion of muscle spindles of the b2c type seems to be another factor contributing to the characteristic sensory complement of different muscles. In Banks and Stacey’s18 sample this ranged from 0% in interosseus to 34% in complexus, the proportion in soleus being 11%. Almost all of these have a single sensory ending that is perhaps best recognized as a b2c primary, though its axon is intermediate in caliber between Ia and II afferents12 and its response to stretch cannot be modulated by dynamic fusimotor input (see below). A single secondary ending was additionally present in only 10% of the total number of b2c spindles in the sample of Banks et al.,12 and none of these spindles had more than two afferents. There is very little comparative information available for the sensory complements of the muscles of other species.

However, Banks and Stacey have collected some unpublished data on 98 spindles from the soleus, peroneus longus, and deep lumbrical of the rat, all of the b1b2c type. In the soleus sample, the number of afferents ranged from 1 to 4, the mean number of “additional” afferents being 1.08, and the best fitting binomial parameters were n  2 and p  0.54.

DEVELOPMENT OF MUSCLE SPINDLES In this short review there is space only for a consideration of the events that occur during normal development. (For an extensive review of abnormal development following nerve or muscle injury, as well as regeneration and reinnervation in the adult, see Zelená.109) The following account is based mainly on the rat, which has been the focus of most structural and immunohistochemical studies. In recent years the genetic basis of muscle spindle development in the mouse has become tractable through the application of knockout and transgenic technology. Gestation in the rat lasts for 21 to 23 days. At about embryonic day 15 (E15), the muscle primordia in the hind limb consist of undifferentiated primary myotubes that are immunoreactive for a tonic myosin heavy chain (MHC).75 In the absence of direct evidence to the contrary, it is simplest to assume that these myotubes are able to differentiate either as intrafusal or extrafusal muscle fibers and that they will differentiate as intrafusal fibers if, and only if, they are contacted by afferent axons (for a contradictory view, see Soukup et al.102). The arrival of the innervation, both sensory (the presumptive Ia afferents) and motor (the presumptive alpha motoneurons), is virtually coincident with the first clear differentiation of the primary myotubes at E17, when the tonic MHC is upregulated in some of the myotubes (the presumptive bag2 fibers) and downregulated in the rest (presumptive type I extrafusal fibers),75,76 although both types express similar levels of neonatal and slow-twitch MHC at this time.90 The secondary myotubes that develop in association with the presumptive bag2 myotube give rise successively to the bag1 fiber and the chain fibers, in a pattern that seems to be a modification of the extrafusal version.85,86 The presumptive bag1 fiber expresses only neonatal MHC when it first appears at E19; around 2 days later the neonatal MHC is replaced by slow-twitch and tonic MHCs.90,102 When the future chain fibers first appear, from E21 to postnatal day 4 (P4), they also express neonatal MHC, although there is disagreement as to whether it is expressed alone90,102 or together with slow- and fast-twitch isoforms.74 It may be that the myoblasts that give rise to these secondary myotubes are themselves multipotent; equally, there may be separate lineages, since the presumptive bag1 and chain myotubes appear together rather than sequentially in human spindles.91 However, it should be noted that the

137

The Muscle Spindle

secondary, unlike the primary, myotubes are assembled in the presence of the afferent innervation. If it is indeed true that at least the primary myotubes are undifferentiated until the arrival of the innervation, a relatively simple model of random instructional contact by the first afferents to enter the muscle primordium may then be postulated, thus obviating any necessity for the afferents to locate specific myotubes. The afferents themselves can then form the pathways to guide the remaining intrafusalspecific innervation (group II afferents and gamma efferents) to their appropriate targets, and such a mechanism can adequately account for many features of the pattern of sensory and motor innervation seen in adult spindles.7 Even the sequential development of recognizable bag1 and chain intrafusal muscle fibers requires the continued presence of the sensory, but not the motor, innervation, although the motor innervation does affect the differential expression of MHC isoforms along the length of the individual fibers.101,109 As yet little is known about the molecular aspects of the interaction between the afferents and the myotubes that they contact. The earliest to have been identified so far is probably the onset in the mouse of the differential expression of Egr-3, a member of the early growth response family of zinc-finger transcription factors, in presumptive intrafusal, but not extrafusal, muscle fibers.105 This occurs at about the equivalent stage of development in the mouse (E15.5) as the upregulation of tonic myosin at E17 in presumptive bag2 fibers of the rat, as described above. When Egr-3 expression is absent, spindles begin to form but the characteristic slow (tonic-like) myosin is not induced; their sensory and motor innervation is then withdrawn and the incipient spindles degenerate.104 A little later than the onset of Egr-3 expression (at E19 in the rat), the bag2 fibers begin to produce NT-3, one of the neurotrophin family of trophic factors, as judged by in situ hybridization of the messenger RNA (mRNA).39 The bag1 fibers follow suit at E21, and the bag fibers, exclusively in muscle, continue to express NT-3 through to maturity, the mRNA occurring especially in the small amounts of sarcoplasm between the nuclei of the nuclear bags. Tyrosine receptor kinase C (TrkC) is the receptor for NT-3 and it is expressed in a subfraction of dorsal root ganglion (DRG) cells, including the large muscle afferents. Absence of NT-3 or its receptor as a result of targeted disruption of the appropriate gene in the mouse results in various deficiencies, including a lack of muscle spindles and tendon organs, at least in the hind limb.52 For muscle afferents of the trigeminal mesencephalic nucleus, the situation seems more complex since 38% of masseter spindles survive NT-3 knockout and all of them survive in the absence of TrkC.53,81 This is a nice story of mutual dependency of the kind that we should expect to see operating in an integrated structure induced by instructional contact: DRG cells expressing TrkC are dependent on a supply of NT-3 for their survival, and they themselves

induce NT-3 expression in their instructed targets, an expression that continues to be nerve dependent in the adult.40 Lack of spindles (and tendon organs) in the NT-3 knockout mouse seems to be a consequence of greatly increased apoptosis from about E10.5.73 Since this is, of course, somewhat earlier than neuromuscular contact, the DRG cells are presumably sustained at first by NT-3 from a source other than the muscle itself.

INTRAFUSAL MUSCLE FIBERS Space permits only a brief consideration of the intrafusal muscle fibers; the following account is based largely on the recent reviews by Zelená,109 Soukup et al.,102 and Banks and Barker,9 where further details may be found. Typically, the chain fibers are considerably shorter than either the bag1 or bag2 fiber, and often only the bag fibers extend well beyond the ends of the capsule. Nevertheless, in almost all cases, each intrafusal muscle fiber extends from pole to pole within the spindle and thus comprises two contractile polar regions separated by a noncontractile equatorial region bearing sensory terminals. This is perhaps the most fundamental way in which intrafusal and extrafusal muscle fibers differ and therefore deserves emphasizing at the outset. Normally both contractile poles receive motor nerve endings, though in only a minority of cases are they supplied by the same axon (see Banks7 for a review). The intimate nature of the contact between the sensory terminals and the underlying muscle fibers leaves little room to doubt that length changes of the muscle in which the spindle lies are transmitted in more or less modified form to the sensory endings by the intrafusal fibers, and their passive and active mechanical properties contribute significantly to shaping the sensory responses. This is not to deny the existence of additional passive components, both parallel and series, that contribute to the sensory response, nor, indeed, additional active components, including voltage- and Ca2-dependent properties of the sensory endings themselves. When the three types of intrafusal fiber were first clearly recognized,14 the criteria for differentiation included morphology and ultrastructure, but especially enzyme histochemical profiling. Although attempts were made to correlate the profiles of intrafusal and extrafusal muscle fiber types, it soon became clear that the intrafusal fibers possessed unique combinations of properties; thus, for example, the fastest contracting fibers (the chains), while exhibiting an appropriate level of alkaline-stable ATPase activity, were also the most oxidative. Moreover, these properties showed marked regional variation with respect to the location of the primary sensory ending in particular. The introduction of immunohistochemical staining techniques further emphasized the uniqueness of the mammalian intrafusal fibers in demonstrating the

138

Function of the Peripheral Nervous System

occurrence and distribution of various isoforms of the MHC: as expected, chain fibers characteristically express a fast-twitch isoform, whereas bag2 fibers are characterized by a slow-twitch form. However, for most of their length the bag1 fibers express a slow, tonic-like myosin that shares some antigenic similarities with vertebrate tonic extrafusal fibers.80 In mammals this type of extrafusal fiber is found only in a few specialized locations, such as the “orbital rim G fibers” of extraocular muscles, where, most interestingly, these fibers may play the role of nuclear-bag fibers in muscle spindles (see Barker19 for a review). Both the bag1 and bag2 fibers of skeletal muscles show regional variation in myosin expression, the bag1 having a slow-twitch myosin like that of the bag2 in its extracapsular polar regions, whereas the tonic myosin also occurs in the juxtaequatorial region of the bag2 (i.e., its contractile portion within the periaxial space).102 Indeed, the expression of tonic myosin in the presumptive bag2 fiber is one of the earliest signs of differentiation of intrafusal from extrafusal fibers (see above). Intracellular recording20 has shown that, in response to electrical stimulation of a fusimotor axon at 1 Hz, neuromuscular transmission evokes action potentials in chain fibers, and rarely in bag2. Only junctional potentials have been recorded from bag1 fibers, and they are more common in the bag2 (seven of eight recordings) than are action potentials. Direct observation of living spindles in situ25 or in isolation32 has shown that the two polar regions of each intrafusal muscle fiber act as independent contractile units, so any excitation, whether junctional or action potentials, presumably fails to cross the equatorial region. Contraction of the chain fibers is rapid and distributed throughout the pole, whereas that of the bag fibers is focal and slower, especially in the bag1.25 Against expectation, the sites of these foci of contraction did not usually correspond to a motor ending, nor did they correspond to the locations from which junctional potentials had been recorded in similar preparations.10 It should be noted, however, that, to elicit these contractions, repetitive stimulation at rates of up to 110 Hz were used. The greatest discrepancies occurred in the bag1 fibers, whose focal contractions could be as much as 2.5 mm away from the nearest motor ending. The contraction sites were usually located extracapsularly where the bag1 fiber exhibits some twitch-like properties in sarcomere structure and (at least in the rat) myosin isoform content.

SENSORY INNERVATION We have seen that, most probably, the Ia afferent induces the differentiation of the intrafusal muscle fibers. Its peripheral terminals occupy, and indeed define, the equatorial region, a length of about 350 m in b1b2c muscle spindles of the cat hind limb. There is such a close correlation between

the disposition of these sensory terminals and the grouping of the underlying myonuclei into nuclear bags and nuclear chains12 that it seems likely to reflect an important mutual functional dependency. The equatorial myonuclei are frequently described in even the most elementary account of muscle spindles, but our familiarity with them should not distract us from their highly unusual nature. Where else can such an accumulation of euchromatic nuclei (about 80 per bag2 fiber, on average12) be found in such a small space? It seems unlikely that the local demands for mRNA are so high as to require these numbers. Could they, perhaps, be essentially passive elements (volumetrically incompressible springs?) that happen to possess mechanical properties suited to the overall mechanosensory transduction process of the equatorial region? From measurements of overall tension and the distribution of strain in isolated spindles both before and after removal of a portion of the capsule, Poppele et al.95 were able to show that the stiffness of the equatorial sensory region, unlike the polar regions, is not length dependent. At the “rest length” (the shortest length they used, with a just-measurable tension), Poppele et al.95 found the stiffness of the equatorial region to be eight times greater than that of the polar regions. With the spindle extended by about 15%, the equatorial region was only twice as stiff as the polar regions, as a result of the nonlinear (apparently exponential) increase in polar-region stiffness. The relative stiffness of the equatorial region is, at first sight, perhaps somewhat surprising, because the intrafusal muscle fibers, as the most conspicuous longitudinal element, exhibit a local minimum in diameter here.12 Poppele et al.95 suggested that the equatorial stiffness might be due mainly to the sensory terminals themselves; alternatively (or additionally), the fine elastic fibers that attach to the surfaces of the bag fibers over about 300 m on either side of the primary ending3 could be the main passive component involved. It is unclear whether these elastic fibers are continuous with the more robust ones that are a prominent feature surrounding the polar regions of the bag2 fiber, and whose branches are extensively dispersed among the inner and outer capsules,38,54 but if so they are also well placed to absorb tension during active or passive length changes in either pole of a bag fiber, and to provide the necessary restoring force on reversal of the change in length. The sensory terminals that occupy the equatorial region of all the intrafusal fibers constitute the primary ending of Ruffini,99 which he described as annulo-spiral in form. He noted that the primary ending of the cat was usually supplied by a single, very large nerve fiber, but that occasionally it was supplied by two fibers, which remained separate as far as they could be traced (see above). For simplicity, we consider further only those endings derived from single fibers. Reconstruction of three such primary endings from serial sections,4,12 one of which is illustrated in Figure 6–3, has shown that, whereas the overall form of the endings is remarkably constant, the contributions made to that form

The Muscle Spindle

139

FIGURE 6–3 The primary ending of a cat tenuissimus muscle spindle. Isometric reconstruction from serial, 1-m thick, toluidine-blue stained transverse sections. The sensory terminals are shaded gray. Equatorial nuclei are shown as ellipsoids containing a nucleolus (black dot), or as the nucleolus alone. Myelinated preterminal branches are unshaded. A, The entire ending. Note that the Ia afferent has already divided into branches that supply the bag1 terminals and the bag2/chain terminals separately (b1br, b2cbr). B, Removal of most of the preterminal branches to show the terminals on the bag1 (below) and bag2 (above) fibers. Note the dense packing of terminals on the bag1 fiber especially. C, Removal of the bag fibers to show the chain terminals. (Modified from Banks, R. W., Barker, D., and Stacey, M. J.: Form and distribution of sensory terminals in cat hindlimb muscle spindles. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 299:329, 1982.)

by individual terminals are very variable (Fig. 6–4). One may infer from this that the determination of terminal form is due to local interaction between the terminal and the underlying intrafusal muscle fiber, and is not purely an intrinsic property of the terminal itself. The sensory terminals are derived from a system of preterminal branches; the following description of the pattern and distribution of the myelinated preterminal branches of primary endings is based on data relating to the cat tenuissimus muscle, since no other muscle has provided so much information. The data are taken from papers by Banks et al.12,16 and Banks.4 If each ending is considered to begin at the first branching node, then the 34 endings included a total of 275 nodes, of which 101 (37%) branched dichotomously, 18 (7%) trichotomously, and 1 (0.4%) pentachotomously. The number of nodes in individual endings ranged from 3 to 14, and the proportion of branched nodes from 22% to 100%. There were a total of 177 ultimate preterminal branches in a range of 3 to 9 per ending. Nineteen of these were first, 83 second, 72 third, and 3 were fourth order. With few exceptions, each ultimate branch supplied only one type of intrafusal muscle fiber. Of the exceptions, most (seven) supplied bag2 and chain fibers together; in the only remaining case, an ultimate myelinated branch supplied a bag1 together

with a single chain fiber.15 Each of the 35 bag1 fibers received on average more ultimate preterminal branches (1.9  0.28 per fiber [mean  standard error]) than either the 36 bag2 fibers (1.5  0.18 per fiber) or the approximately 140 chain fibers (0.55  0.09 per fiber; data based on the sample of Banks et al.16), although those supplying bag1 fibers tended to be of lower order (modal value 2) than those supplying bag2 (modal values 2 and 3) or chain fibers (modal value 3). No two endings showed the same pattern of branching, but one feature did recur repeatedly: In 27 of the 34 endings, the first branching node was dichotomous and one branch exclusively supplied all the terminals on the (single) bag1 fiber, whereas the other, perforce, supplied the terminals of the bag2 and chain fibers. Moreover, in each of the three endings where the first node branched tri- or pentachotomously, one first-order branch again supplied the bag1 fiber exclusively. Unmyelinated branches form the final distributional stage of the preterminal axons. They arise almost exclusively from the heminodes of the ultimate myelinated segments, are about 1 m in diameter and 10 to 20 m in length, and are sometimes branched (Fig. 6–5). Taken together with evidence from their ultrastructure and histochemistry,98 their dimensions relative to the much larger sensory terminals strongly indicate that they are the sites

140

Function of the Peripheral Nervous System

FIGURE 6–4 Schematic diagrams to illustrate the variability in the preterminal branching pattern and terminal distribution in cat tenuissimus primary endings, based on reconstructions from serial transverse (A and B) or longitudinal (C) 1-m thick sections. Information on the distribution of chain terminals was obtained only from the longitudinal sections. Myelinated internodes are shown as elongated rectangles, and unmyelinated segments of axons are shown as simple lines. Differently shaded regions of the bag1 (b1), bag2 (b2), and chain (c) fibers represent separate terminals. Primary A is the same as that in Figure 6–3; the terminals of primary C are shown in more detail in Figure 6–6. (Modified from Banks, R. W.: Observations on the primary sensory ending of tenuissimus muscle spindles in the cat. Cell Tissue Res. 246:309, 1986. © Springer-Verlag.)

where encoding of the receptor potentials occurs. The transition from preterminal to terminal branch is marked by a sudden increase in diameter, the expanded terminals invariably being in contact with intrafusal muscle fibers. The following description of the sensory terminals is again derived mostly from the cat tenuissimus muscle and the papers of Banks et al.12 and Banks.4 Individual terminals may be simple or branched, with rami of very variable lengths. They wind around the muscle fibers, appearing to occupy the available space in a competitive manner, and so making their own, very variable, contribution to the overall form of the primary ending (see Fig. 6–4). The bag1 fiber has the most, and the most highly branched, terminals; individually, chain fibers have the fewest and the least branched (Fig. 6–6). Consequently, of the total neuromuscular contact area in a primary ending, about 35% is due to terminals on the bag1 fiber, 25% to terminals on the bag2 fiber, and 40% to those on the chain

fibers collectively. Again, the density of terminals is greatest on the bag1 fiber (with about 55% coverage of the equatorial surface of the fiber), intermediate on the bag2 (with about 45% coverage), and least on the chains (with about 42% coverage), so that the overall longitudinal extent of the terminals on each fiber is about the same among the different types. In thin (1-m) sections through the long axes of intrafusal muscle fibers, profiles of the sensory terminals are mostly lentiform in outline (Fig. 6–7), especially toward the middle of the ending, where the pitch of any spirals is small. Both the free (outer) boundary of the terminal profile and the neuromuscular (inner) boundary thus approximate to segments of circles, whose common chord is in the plane of the surface of the muscle fiber and is of fairly constant width, at about 6 m. The shape of these profiles suggests a mechanical model of tension transmission from the intrafusal fibers, whether at constant length or undergoing active or passive length change, in which the terminals are

141

The Muscle Spindle

mpt

upt

to c 1

t

to c 2 t

FIGURE 6–5 Detail of the primary ending of Figure 6–3 showing part of the terminals on the bag1 fiber. The final segment of a myelinated preterminal branch (mpt) gives rise, at the heminode, to a much narrower unmyelinated preterminal segment (upt) that immediately branches to supply two separate, greatly expanded, terminals (t) in contact with the muscle fiber. (Modified from Banks, R. W., Barker, D., and Stacey, M. J.: Form and distribution of sensory terminals in cat hindlimb muscle spindles. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 299:329, 1982.)

squeezed between inner (cellular) and outer (extracellular) tension-bearing elements. It may be supposed that the minimum-energy condition (zero tension) would correspond to a circular terminal profile, that increasing tension results in progressively greater lentiform distortion of the profile, and that consequently the sensory-terminal membrane is progressively stretched as the surface area of a volume element of the terminal membrane is increased. Moreover, differential tension distribution between the inner and outer surfaces of the sensory terminals would result in asymmetrical development of the lentiform profile, as shown by the prominently bulging terminals on the bag1 fibers in Figure 6–7, or, at the opposite extreme of form, the deeply indented ones of the chain fibers. A candidate for the extracellular tension element required by this model is the continuous basal lamina that covers the intrafusal muscle fibers and the outer (free) surfaces of the sensory terminals. There is, however, no intervening basal lamina at the neuromuscular boundary, so the terminals appear to have insinuated themselves between the cellular and extracellular components of the sarcolemma. The lack of dystrophin at this boundary, when it is present elsewhere at the intrafusal muscle fiber membrane,87 is perhaps a secondary consequence of the separation of the sarcolemmal components, but nonetheless interesting in attesting to highly localized variations in mechanical properties. In the absence of any evidence to the contrary, it is simplest to assume that the mechanical properties of the basal lamina and the sensory terminals are constant throughout

1 2 3

FIGURE 6–6 Semischematic arrangement of the terminals of primary C in Figure 6–4. Individual terminals (darker shades) are shown above and below their corresponding locations in the overall arrangement (paler shades) on the bag1 (red), bag2 (green), and chain (blue) fibers. Note that there are only two separate terminals on the three chain fibers, with chain 2 receiving its terminals directly from those of the other two fibers (arrows indicate these “sensory-cross terminals”1). Examples of the longitudinal sections used in the reconstruction are shown in Figure 6–7, spindle ii. (Modified from Banks, R. W.: Observations on the primary sensory ending of tenuissimus muscle spindles in the cat. Cell Tissue Res. 246:309, 1986. © Springer-Verlag.)

the equatorial region. If so, the range of differential form of the terminals (bulging on the bag1, more or less symmetrical on the bag2, and deeply indented on the chains) must reflect the mechanical properties of the different types of intrafusal muscle fiber. Whatever the precise mechanism of terminal distortion and its differential properties, we may suppose that longitudinal stretch of the equatorial region leads to a continuously varying receptor potential, generated by the sensory terminals. Hunt and Ottoson63 were able to record such a potential from single afferent axons (Ia or II) in isolated muscle spindles of the cat, when the spiking activity of the afferents had been blocked with tetrodotoxin. Hunt and

142

Function of the Peripheral Nervous System

Spindle Bag1 fibres

Sarcomere length um

(i)

2.2

(ii)

2.5

iii

2.6

Bag2 fibres (i)

2.35

(ii)

2.5

(iii)

2.5 Chain fibres

(i) 1.95 (ii) (iii)

2.05 20 um 2.25

FIGURE 6–7 Tracings of 1-m thick, longitudinal sections, stained with toluidine blue, through the bag1, bag2, and a sample chain fiber of three cat tenuissimus muscle spindles (i to iii). Mean sarcomere lengths show that the spindles were fixed under different (unknown) amounts of static tension increasing in order, i to iii. Analysis4 of the shape of the terminal profiles (colored areas) indicated that they are progressively deformed by the increasing longitudinal tension. (Modified from Banks, R. W.: Observations on the primary sensory ending of tenuissimus muscle spindles in the cat. Cell Tissue Res. 246:309, 1986. © Springer-Verlag.)

Wilkinson64 went on to show that the fundamental frequencies of both the receptor potential and the overall tension of the muscle spindle responded similarly to sinusoidal stretch. The depolarizing current is carried by Na, but a Ca2 current, which can be blocked by D600, is also discernible when Na is removed from the bathing medium.65 Presumably closure of the mechanosensory channels leads to repolarization of sensory-terminal membranes, but at least two types of K current (one tetraethylammonium sensitive, the other not) also seem to

contribute; in primary endings, the K currents can produce hyperpolarization on release of stretch, though the precise location of these currents within the primary ending remains uncertain.65 Kruse and Poppele,71 adopting a systems analytical approach, provided evidence that a Ca2activated K channel contributes to the mid-frequency (0.4 to 4 Hz) dynamics of the primary ending. Although, under normal conditions, Ca2 makes an insignificant contribution to the receptor potential, it nevertheless seems to play a key role, as yet undefined, in the proper functioning of the primary ending. Thus blockage of Ca2 channels by diltiazem or CoCl2 rapidly results in the abolition of spiking activity.71 The importance of Ca2 is underlined by the variety and abundance of Ca2-binding proteins that occur in the ending, including calbindin D-28k, calretinin (Fig. 6–8A), neurocalcin, and, at least in frog spindles, frequenin.44,48,56,66,108 All are members of the EF-hand superfamily, representing two families of proteins with either six or four EF-hands, thought to act respectively as Ca2 buffers (calbindin D-28k, calretinin), or Ca2-activated switches (neurocalcin, frequenin).33,67 Details of the distribution of the various proteins differ: Apart from the primary sensory ending, calbindin D-28k also occurs in the intrafusal muscle fibers of rat spindles, whereas calretinin occurs in the chain fibers of cat spindles and in cat tendon organs, but is absent from secondary endings and from rat tendon organs.44,48 Maintenance of the ionic composition of such large sensory terminals as those of the primary and secondary endings of muscle spindles may be expected to be metabolically demanding. It is therefore not surprising to find that mitochondria are abundant in them, often being concentrated in their central cores. What is perhaps more surprising is the abundant population of vesicles that is also present. Various types of vesicle occur commonly in many kinds of mechanoreceptor terminal, but until recently they appear to have been dismissed as having no obvious or demonstrable function related to mechanosensory transduction.2 Especially prominent among them are large numbers of small, clear vesicles resembling those of chemical synapses1 (Fig. 6–9). The apparent similarity of these “synaptic-like vesicles” to the classic presynaptic vesicles of motor neuromuscular junctions, for example, was enhanced when De Camilli et al.43 showed that synapsin 1 and synaptophysin occur in the sensory terminals of muscle spindles (see Fig. 6–8B) and tendon organs as well as in motor endings. Use of the fluorescent styryl dye FM1-43 to study the dynamics of vesicle release in the (motor) neuromuscular junction26 led directly to the observation that it also stained the sensory endings of muscle spindles (see Fig. 6–8C), but stimulation of the nerve is necessary to induce uptake of dye into the motor endings, whereas the sensory endings accumulate it spontaneously. It is supposed that the

The Muscle Spindle

143

FIGURE 6–8 A, The distribution of calretinin in the equatorial region of a spindle from the abductor digiti quinti medius muscle of the cat, demonstrated using the peroxidase-antiperoxidase technique (25 m-thick frozen section, calibration 50 m). The primary ending (preterminal and terminal branches) is the most intensely stained, but the chain fibers themselves are also positively stained. B, The distribution of synaptophysin, a protein characteristic of synaptic vesicles, in a semispinalis capitis muscle spindle of the rat, using indirect immunofluorescence and confocal microscopy (25-m-thick frozen section, calibration 50 m). Immunostaining is essentially restricted to the sensory terminals, in this case of a primary ending. Those on the bag2 fiber (above) and a chain fiber (below) are visible in this confocal plane. C, FM1-43 labeling of the sensory terminals in a primary ending of a rat lumbrical muscle spindle (calibration 20 m). The living muscle was maintained in vitro during uptake of the dye, which occurs spontaneously. The muscle was then fixed prior to confocal microscopy. The dye is thought to be incorporated into the membranes of “synaptic-like vesicles”13 within the terminals. (The preparations illustrated were made by Amal El-Tarhouni [A], Ian Maclean and Christine Richardson [B], and Guy Bewick [C].)

FIGURE 6–9 A, Electron micrograph of a transverse section through the equatorial region of a bag fiber in a muscle spindle of cat flexor digitorum longus. Note the location of the sensory terminal (t) between basal lamina (bl) and underlying muscle fiber. B, Detail of the box in A showing small, clear (“synaptic-like”) vesicles (v) in the terminal. An “” figure (arrow) indicates that the vesicles may fuse with the terminal membrane. (The preparations illustrated were made by Mohammed Adal.)

dye becomes incorporated into the vesicular membranes, and if so, it appears that the vesicles in the sensory endings are fused with the terminal membrane and recycled constitutively, though there do not seem to be any specialized release sites. We have now provided evidence that this process of recycling can be modulated in an activitydependent manner; that glutamate is relatively enriched in the sensory terminals; and that exogenously applied glutamate has an excitatory action on muscle spindle afferents.13,17,27 These results have revealed a hitherto unknown, possibly even unsuspected, level of complexity of the mechanosensory properties of muscle spindle afferents, one that will almost certainly prove to have a more general applicability among mechanoreceptors. Much

more work will be needed before its functional significance can be fully understood. We have already seen that encoding of the continuously varying potentials generated by terminal deformation (receptor potentials) is likely to occur preferentially at the heminodes of the ultimate myelinated preterminal branches. But in each primary ending there are several such potential encoding sites, and theoretical considerations indicate that the overall output of the ending (in the form of impulses successfully invading the parent afferent) will be dominated by whichever encoding site momentarily has the highest firing rate.45 This is due to resetting of the encoding sites having lower firing rates at that moment by antidromic invasion of impulses from the site with the

144

Function of the Peripheral Nervous System

highest rate. In extreme circumstances it can lead to complete occlusion of the output of a less active site by that of a more active one, a phenomenon first described from muscle spindle studies, and explained in this way, by Crowe and Matthews.41 Banks et al.16 systematically studied the interactions of encoding sites activated by pairs of dynamic and static fusimotor axons (see following section for a description of these terms). Use of the tenuissimus muscle enabled the physiologic and histologic properties of individual primary endings to be correlated. Banks et al.16 defined a coefficient of interaction (Ci), such that Ci  1 for linear summation of the separate inputs Ci  0 for complete occlusion of one input by the other Although Ci was not constant, standardization of conditions allowed a characteristic mean value to be calculated for each primary ending. Whenever the static and dynamically activated sites were separated by myelinated preterminal branches containing at least 4 complete nodes of Ranvier (12 of 16 primaries), their interactions were dominated by occlusion (Ci  0.35). With fewer than four nodes separating the encoding sites, the overall output usually showed greater summation, and the greatest amount of summation (Ci  0.69) was shown by a primary ending whose activated encoding sites were separated by branches containing just a single complete node. While recognizing that several factors could be involved in the summation, Banks et al.16 argued that electrotonic spread of receptor potentials between the encoding sites was probably the most important (at least under the conditions of their experiments), since others such as mechanical coupling of the intrafusal fibers would not be expected to correlate with primary-ending structure. Having considered the mechanosensory transduction and encoding processes, we are now in a position to understand at least some of the complexities in the final output of a sensory ending (especially of a primary) as it responds to the stimulation of changing muscle length. Quantitative study of the relationship between spindle afferent output and muscle length, including its modification during fusimotor stimulation (see following section), was greatly facilitated by the single-unit technique introduced by Kuffler et al.78 This method has been used, more or less modified, in virtually all subsequent investigations of the response properties of the sensory endings. In the present chapter, it is impossible to treat the many important contributions to this work individually. They have been reviewed by some of the principal authors of this work, including Kennedy et al.,69 Matthews,83 Hulliger,59 Boyd,29 Hunt,62 and Prochazka.97 We conclude this section with a brief summary of the passive responses of sensory endings, before, in the final section, seeing how these are modified by fusimotor stimulation. Linear systems analysis of continuous length changes, both sinusoidal94 and random,93 has provided quantitative descriptions of the responses of primary and second-

ary endings that are applicable within their linear ranges. For example, if the amplitude of stretch does not exceed about 0.1% of muscle length, the response of a primary ending may be fully described by the transfer function: Ks(s+0.44)(s+11.3)(s+44) H(s)=

(s+0.44)(s+0.8) where s  j is the Laplace transform variable and K is a parametric gain constant. A comparison of responses to random and sinusoidal stretches outside the linear range revealed at least two types of nonlinearity93: a static one affecting both primary and secondary endings in which sensitivity (impulses/s/mm) decreases with stretch amplitude for stretches up to about 1% of muscle length; and a dynamic one affecting only primary endings, which, for example, produces a fractional power-law relationship between Ia responses and velocity of stretch during ramp stretches.57,58 There is fairly general agreement that the transition from linearity to (static) nonlinearity corresponds with the limit at which attached cross-bridges in the intrafusal sarcomeres are no longer able to absorb their proportion of the applied stretch. It is quite possible that the dynamic nonlinearity is also due to the mechanical properties of the intrafusal fibers, in this case mainly the bag1. We have seen that this fiber is characterized by a tonic myosin, and that it does not twitch, but there is also evidence that stretch activation may contribute to its dynamic behavior.96 The interaction of separate encoding sites located in the first-order branches of the Ia afferent, and thus associated with the bag1 and bag2  chain fibers, respectively, results in a further nonlinearity that we have already considered (see above). Attempting to identify the source of the nonlinearities in the sensory responses is important for understanding the internal working of the spindle, but this knowledge is much less important in trying to model the behavior of the muscle spindle as part of a general motor control system. For this purpose, it may be sufficient to have a nonlinear transfer function available that accurately predicts the output of the spindle under all stimulation conditions. Attempts to derive such a function using white noise analysis to estimate Wiener kernels have made some progress, but face some important technical difficulties (see Kröller70).

MOTOR INNERVATION Examples of primary endings responding to various combinations of muscle stretch and fusimotor stimulation are shown in Figure 6–10, using unpublished results of Banks and Hulliger. They have been chosen to illustrate several important aspects of the functional organization of fusimotor innervation. The stretches, whether trapezoidal or sinusoidal, are of large amplitude and consequently outside the linear range of the primary.

145

The Muscle Spindle

FIGURE 6–10 Examples of the responses of primary endings to a variety of conditions of muscle stretch and fusimotor stimulation in the peroneus digiti quinti of cat. A and B, Responses of the primary endings (instantaneous frequency plots, upper traces) of two spindles in the same muscle to a series of five trapezoidal stretches (lower traces) in the presence (broad line, middle traces) or absence (narrow line, middle traces) of fusimotor stimulation. Pseudorandom intervals of fusimotor stimulation were used with a mean rate of 100 Hz and 20% coefficient of variation. In each horizontal pair (i to iii, A and B) are shown the effects of stimulating the same fusimotor axon with the same pseudorandom pattern on the responses of the two primary endings: i, a dynamic gamma axon; ii, a static gamma axon that activates the bag2 fiber only in each spindle; and iii, a static gamma axon that activates both bag2 and chain fibers in each spindle. Note the strength of the fusimotor action in Aiii; in the first active stretch cycle of Biii, cross-correlation analysis showed a clear tendency to 1:1 driving (Ia fires an impulse for every motor stimulus), indicating a predominantly chain action. C, The same endings as Ai to Aiii except that the muscle was stretched sinusoidally at 1 Hz and 0.5 mm amplitude. Phase advance for passive cycles was 75 degrees. Phase advance for active cycles was bag1 active (Ci), 64 degrees; bag2 active (Cii), 54 degrees; and bag2 and chains active (Ciii), 47 degrees. D, The same endings as Ai to Aiii except that the muscle was held at constant mean length (trace not shown) while the fusimotor axons were stimulated with three cycles of a rate that varied from 0 to 150 to 0 Hz. Both primary response (upper trace) and stimulation rate (lower trace) are shown as instantaneous frequency plots. Note the weak effect of bag1 activation at constant length (dynamic axon, Di), the high gain (and hysteresis) of bag2 activation at relatively low rates of stimulation (static axon, Dii), and a tendency to drive when the chain fibers are strongly activated (static axon, Diii), but recall that the bag2 fiber is also active here. Calibrations: A to C upper trace ordinate 0 to 300 Hz, lower trace ordinate 5 to 5 mm, and abscissa 0 to 40 s; D ordinates (both traces) 0 to 200 Hz and abscissa 0 to 40 s. (The data were obtained together with Manuel Hulliger while on research leave in his laboratory in Calgary.)

(a)

(b)

(i)

(ii)

(iii)

(c)

(i)

(ii)

(iii)

(d)

146

Function of the Peripheral Nervous System

Figures 6–10A and B show the responses of two primary endings (afferent conduction velocities of 76 and 73 m/s, respectively) in the same peroneus digiti quinti muscle of a cat to a series of trapezoidal stretches. Each primary ending could be activated by several fusimotor axons, three of them in common, as shown here. All three are gamma axons, which are exclusively fusimotor in distribution, as shown by Leksell.79 Fusimotor effects are classified into two principal categories: dynamic, in which the sensitivity of the primary to the rate of stretch is increased; and static, in which the mean firing rate of the primary is increased but the dynamic sensitivity is virtually unaltered or reduced.82 Moreover, single fusimotor axons may be described as dynamic or static in that they induce similar effects in each of the muscle spindles that they innervate.24,42 In the examples shown in Figure 6–10A and B, the axons were stimulated at a mean rate of 100 Hz during the periods of the broad horizontal bars. Axon i (conduction velocity 26 m/s) was dynamic, whereas axons ii (conduction velocity 26 m/s) and iii (conduction velocity 23 m/s) were both static. Figure 6–10C shows the same arrangement and combination of primary ending and fusimotor axons as Figure 6–10A except that the muscle is now stretched sinusoidally at 1 Hz. Despite the nonlinearities (especially that of the encoder when the primary output falls silent), it is possible to fit pure sines with minimum error to the responses expressed as probability density functions. When averaged, the passive-cycle sines have the following properties: peak-to-peak modulation, 57 impulses/s, and phase advance, 75 degrees. The corresponding values when the fusimotor axons are active are, for a mean stimulation rate of 30 Hz (not shown), dynamic gamma (i), 68 impulses/s and 70 degrees; static gamma (ii), 50 impulses/s and 55 degrees; and static gamma (iii), 49 impulses/s and 52 degrees. For a mean stimulation rate of 100 Hz (illustrated), the values are dynamic gamma (i), 84 impulses/s and 64 degrees; static gamma (ii), 56 impulses/s and 54 degrees; and static gamma (iii), 50 impulses/s and 47 degrees. These values are quite typical of those obtained from the soleus muscle spindles in the systematic study by Hulliger et al.60 Note that in each case the peak-to-peak modulation is increased by dynamic action with respect to the passive response, whereas it is unaltered or reduced by static action, but both dynamic and static activation reduce the phase advance. Finally, Figure 6–10D shows the effects of steadily increasing, and then decreasing, the rate of stimulation of the fusimotor axons on the primary response while the muscle is held at constant length. This is the same primary ending and set of motor axons as in Figure 6–10A and C. Note that the primary ending has a marked tendency to fire an impulse for every stimulus pulse (“driving”) when static gamma (iii) is activated, indicating that a fast intrafusal mechanism is responsible.

Elucidation of these and other phenomena of fusimotor action has been the work of several decades and has involved some of the most elegantly conceived and executed experiments in muscle spindle physiology. They deserve to be included here, but space considerations forbid it once again; however, details may be found in the review by Banks.7 Accordingly, we summarize only the main conclusions. The functional classification of gamma axons into dynamic and static types corresponds to their segregated distribution, examples of which we encountered in Figure 6–2. There is no distinction between the conduction velocities of the two types, both ranging from 15 to 55 m/s in the cat hind limb. Dynamic axons are distributed to bag1 fibers, whereas static axons supply bag2 and chain fibers. In the cat these distributions are usually exclusive, but in the rat and monkey several instances of co-innervation of the bag1 with either or both of the other types of intrafusal fiber have been reported.72,77 Even in these species, however, the overall distribution of the axons concerned may be predominantly of one kind or the other, so that the distinction between dynamic and static axons still holds. In any particular muscle spindle, the static axons may be distributed to the bag2 and chain fibers together or separately. The variability in the distribution pattern of static gamma axons accounts for a similar variability in the details of their actions,49 such as those illustrated in Figure 6–10, where static axon ii activates the bag2 fiber of both spindles (A) and (B), and static axon iii activates, in different proportions in the two spindles, the bag2 together with the chain fibers. The possibility that there might be two types of static gamma axon30 was refuted by Banks,5 who showed that there was, nevertheless, a statistical tendency for the faster conducting axons to be more widely distributed (i.e., to more spindles) than the more slowly conducting, and that a driving action, when present, tended to be due to the slowest axons supplying the spindle concerned. Apart from these correlations, the complete distribution of individual static axons could be accounted for by chance. Banks6 went on to demonstrate by histologic analysis that the occurrence of a segregated supply of the bag2 and chain fibers in any particular spindle depended on the total number of static axons entering the spindle, and that this was closely correlated with the number of afferent axons present, which was itself apparently determined at random (see Abundance of Muscle Spindles and Afferent Axons above). Banks’5 conclusions regarding the differential distribution of static gamma axons in relation to their conduction velocities were based on his observations on the cat tenuissimus; qualitatively similar results have since been obtained from two other hind limb muscles, the peroneus digiti quinti (peroneus tertius) and peroneus longus.35,51 The results from these, perhaps more typical, muscles were necessarily obtained using only physiologic information. The criteria for identifying the intrafusal

147

The Muscle Spindle

distribution of individual static gamma axons on this basis have recently been critically examined by Petit et al.92 The complete distribution of any single static gamma axon usually includes both bag2 and chain fibers.21 When we are so used to the concept of homogeneous motor units, this may seem strange. It is true that the differential distribution in relation to conduction velocity could provide the central nervous system with a measure of separate control, but Banks6 has pointed out that, when sufficient static gamma axons supply a single spindle to make segregation possible, it is the inputs to the two poles rather than the bag2 and chain fibers that are likely to be segregated. The probable functional significance of the co-innervation of bag2 and chain fibers by single static gamma axons has been investigated by Emonet-Dénand et al.,50 who reported that coactivation of bag2 and chain fibers enables primary endings to signal changes of length continuously over a larger range of stretch velocities and more independently of the average muscle length than is possible during activation of either type of intrafusal fiber alone. In addition to the purely fusimotor (gamma) innervation, muscle spindles may receive some of their motor input as collateral branches of the axons of alpha motoneurons (see Banks7 for a review). The complete distribution of these axons is therefore often described as skeletofusimotor, and the axons are referred to as beta axons. This usage can be quite convenient, but should not be allowed to imply that they constitute a separate category of motor axon. On the contrary, all the evidence indicates that they are identical to the corresponding alpha axons, apart from the fact that they happen to have collateral input to muscle spindles.7 The number of such collateral branches received by any particular spindle is determined at random, but unlike the gamma axons, there is no correlation with the number of afferents present.6 It is likely that this reflects the necessity for gamma axons to locate their intrafusal targets, which they could most easily do by following pathways established by the foregoing afferents during development. The first physiologic observations of beta innervation in mammals were made by Bessou et al.23 The axons were relatively slowly conducting and had dynamic actions, subsequently shown to be due to bag1 activation. Barker et al.22 argued that the proportion of intrafusal motor endings that they identified as derived from beta axons was too high for them all to have been supplied by collaterals of slow-conducting axons. They suggested that the deficit was made up by fast beta axons, a prediction that was confirmed by glycogen depletion studies on peroneus digiti quinti (p. tertius) by Harker et al.55 These beta axons were static in action, selectively activating long chain fibers such as that shown in Figure 6–2Ai. Once regarded as an accident of development, or as an example of the retention of an evolutionarily primitive feature, skeletofusimotor axons are now seen as an integral feature of the motor control

system in mammals. Based on the physiologic identification of beta axons with conduction velocities of 55 m/s or more (any slower beta axons would have been excluded), Jami et al.68 have estimated that at least 50% of the spindles in peroneus digiti quinti (p. tertius) receive a beta innervation, and that one third of the supply is dynamic and two thirds static. They also found that the majority (11/17, or 69%) of alpha motoneurons that conducted at 75 m/s or less were in fact skeletofusimotor in distribution, whereas only a minority (23/82, or 28%) of those that conducted at more than 80 m/s were.

ACKNOWLEDGMENTS I would like to thank Tomasˇ Soukup for helpful comments. The preparations illustrated in Figures 6–8 and 6–9 were made by Amal El-Tarhouni (8A), Ian Maclean and Christine Richardson (8B), Guy Bewick (8C), and Mohammed Adal (9). I am most grateful to all of them for their help with these images. The data shown in Figure 6–10 were obtained together with Manuel Hulliger while on research leave in his laboratory in Calgary. I am completely indebted to Manuel for these results. I am most grateful to Hong Kong University Press, the Royal Society, and Springer-Verlag for permission to reproduce and modify Figures 6–1B and 6–3 through 6–7.

REFERENCES 1. Adal, M. N.: The fine structure of the sensory region of cat muscle spindles. J. Ultrastruct. Res. 26:332, 1969. 2. Akoev, G. N., Alekseev, N. P., and Krylov, B. V.: Mechanoreceptors. Berlin, Springer-Verlag, 1988. 3. Banks, R. W.: On the attachment of elastic fibres in cat tenuissimus muscle spindles. J. Physiol. (Lond.) 348:16P, 1984. 4. Banks, R. W.: Observations on the primary sensory ending of tenuissimus muscle spindles in the cat. Cell Tissue Res. 246:309, 1986. 5. Banks, R. W.: The distribution of static -axons in the tenuissimus muscle of the cat. J. Physiol. (Lond.) 442:489, 1991. 6. Banks, R. W.: Intrafusal motor innervation: a quantitative histological analysis of tenuissimus muscle spindles in the cat. J. Anat. 185:151, 1994. 7. Banks, R. W.: The motor innervation of mammalian muscle spindles. Prog. Neurobiol. 43:323, 1994. 8. Banks, R. W.: On the number of spindles in mammalian muscles. J. Physiol. (Lond.) 511:69P, 1998. 9. Banks, R. W., and Barker, D.: The muscle spindle. In Engel, A. G., and Franzini-Armstrong, C. (eds.): Myology, 3rd ed. New York, McGraw-Hill (in press). 10. Banks, R. W., Barker, D., Bessou, P., et al.: Histological analysis of cat muscle spindles following direct observation of the effects of stimulating dynamic and static motor axons. J. Physiol. (Lond.) 283:605, 1978.

148

Function of the Peripheral Nervous System

11. Banks, R. W., Barker, D., Saed, H. H., and Stacey, M. J.: Innervation of muscle spindles in rat deep masseter. J. Physiol. (Lond.) 406:161P, 1988. 12. Banks, R. W., Barker, D., and Stacey, M. J.: Form and distribution of sensory terminals in cat hindlimb muscle spindles. Philos. Trans. R. Soc. Lond. (Biol.) 299:329, 1982. 13. Banks, R. W., Bewick, G. S., Reid, B., and Richardson, C.: Evidence for activity-dependent modulation of sensoryterminal excitability in spindles by glutamate release from synaptic-like vesicles. In Gandevia, S. G., Proske, U., and Stuart, D. G. (eds.): Sensori-motor Control of Movement and Posture. New York, Plenum, p. 13, 2002. 14. Banks, R. W., Harker, D. W., and Stacey, M. J.: A study of mammalian intrafusal muscle fibres using a combined histochemical and ultrastructural technique. J. Anat. 123:783, 1977. 15. Banks, R. W., Hulliger, M., and Scheepstra, K. A.: Correlated histological and physiological observations on a case of common sensory output and motor input of the bag1 fibre and a chain fibre in a cat tenuissimus spindle. J. Anat. 193:373, 1998. 16. Banks, R. W., Hulliger, M., Scheepstra, K. A., and Otten, E.: Pacemaker activity in a sensory ending with multiple encoding sites: the cat muscle spindle primary ending. J. Physiol. (Lond.) 498:177, 1997. 17. Banks, R. W., Richardson, C., and Bewick, G. S.: Immunocytochemical demonstration of glutamate in the sensory terminals of rat muscle spindles. J. Physiol. (Lond.) 528:62P, 2000. 18. Banks, R. W., and Stacey, M.: Quantitative studies on mammalian muscle spindles and their sensory innervation. In Hník, P., Soukup, T., Vejsada, R., and Zelená, J. (eds.): Mechanoreceptors: Development, Structure and Function. New York, Plenum Press, p. 263, 1988. 19. Barker, D.: The morphology of muscle receptors. In Hunt, C. C. (ed): Handbook of Sensory Physiology, Vol. 3, Pt. 2: Muscle Receptors. Berlin, Springer-Verlag, p. 1, 1974. 20. Barker, D., Bessou, P., Jankowska, E., et al.: Identification of intrafusal muscle fibres activated by single fusimotor axons and injected with fluorescent dye in cat tenuissimus spindles. J. Physiol. (Lond.) 275:149, 1978. 21. Barker, D., Emonet-Dénand, F., Laporte, Y., et al.: Morphological identification and intrafusal distribution of the endings of static fusimotor axons in the cat. J. Physiol. (Lond.) 230:405, 1973. 22. Barker, D., Stacey, M. J., and Adal, M. N.: Fusimotor innervation in the cat. Philos. Trans. R. Soc. Lond. (Biol.) 258:315, 1970. 23. Bessou, P., Emonet-Dénand, F., and Laporte, Y.: Occurrence of intrafusal muscle fibres innervated by branches of slow motor fibres in the cat. Nature 198:594, 1963. 24. Bessou, P., Laporte, Y., and Pagès, B.: Similitude des effets (statiques ou dynamiques) exercés par des fibres fusimotrices uniques sur les terminaisons primaires de plusieurs fuseaux chez le chat. J. Physiol. (Paris) 58:31, 1966. 25. Bessou, P., and Pagès, B.: Cinematographic analysis of contractile events produced in intrafusal muscle fibres by stimulation of static and dynamic fusimotor axons. J. Physiol. (Lond.) 252:397, 1975.

26. Betz, W. J., Mao, F., and Bewick, G. S.: Activity-dependent fluorescent staining and destaining of living vertebrate motor nerve terminals. J. Neurosci. 12:363, 1992. 27. Bewick, G. S., Reid, B., and Banks, R. W.: Investigating the role of small clear vesicles in vertebrate mechanosensory endings using rat muscle spindles. J. Physiol. (Lond.) 528: 62P, 2000. 28. Boyd, I. A.: The structure and innervation of the nuclear bag muscle fibre system and the nuclear chain muscle fibre system in mammalian muscle spindles. Philos. Trans. R. Soc. Lond. (Biol.) 245:81, 1962. 29. Boyd, I. A.: Intrafusal muscle fibres in the cat and their motor control. In Barnes, W. J. P., and Gladden, M. H. (eds.): Feedback and Motor Control in Invertebrates and Vertebrates. London, Croom Helm, p. 123, 1985. 30. Boyd, I. A.: Two types of static -axon in cat muscle spindles. Q. J. Exp. Physiol. 71:307, 1986. 31. Boyd, I. A., and Davey, M. R.: Composition of Peripheral Nerves. Edinburgh, E. & S. Livingstone, 1968. 32. Boyd, I. A., and Ward, J.: Motor control of nuclear bag and nuclear chain intrafusal fibres in isolated living muscle spindles from the cat. J. Physiol. (Lond.) 244:83, 1975. 33. Burgoyne, R. D., and Weiss, J. L.: The neuronal calcium sensor family of Ca2-binding proteins. Biochem. J. 353:1, 2001. 34. Burke, R. E., Levine, D. N., Salcman, M., and Tsairis, P.: Motor units in cat soleus muscle: physiological, histochemical and morphological characteristics. J. Physiol. (Lond.) 238:503, 1974. 35. Celichowski, J., Emonet-Dénand, F., Laporte, Y., and Petit, J.: Distribution of static axons in cat peroneus tertius spindles determined by exclusively physiological criteria. J. Neurophysiol. 71:722, 1994. 36. Chin, N. K., Cope, M., and Pang, M.: Number and distribution of spindle capsules in seven hindlimb muscles of the cat. In Barker, D. (ed.): Symposium on Muscle Receptors. Hong Kong, Hong Kong University Press, p. 241, 1962. 37. Clark, D. A.: Muscle counts of motor units: a study in innervation ratios. Am. J. Physiol. 96:296, 1931. 38. Cooper, S., and Gladden, M. H.: Elastic fibres and reticulin of mammalian muscle spindles and their functional significance. Q. J. Exp. Physiol. 59:367, 1974. 39. Copray, J. C. V. M., and Brouwer, N.: Selective expression of neurotrophin-3 messenger RNA in muscle spindles of the rat. Neuroscience 63:1125, 1994. 40. Copray, J. C. V. M., and Brouwer, N.: Neurotrophin-3 mRNA expression in rat intrafusal muscle fibres after denervation and reinnervation. Neurosci. Lett. 236:41, 1997. 41. Crowe, A., and Matthews, P. B. C.: The effects of stimulation of static and dynamic fusimotor fibres on the response of stretching of the primary endings of muscle spindles. J. Physiol. (Lond.) 174:109, 1964. 42. Crowe, A., and Matthews, P. B. C.: Further studies of static and dynamic fusimotor fibres. J. Physiol. (Lond.) 174:132, 1964. 43. De Camilli, P., Vitadello, M., Canevini, M. P., et al.: The synaptic vesicle proteins synapsin I and synaptophysin (protein p38) are concentrated both in efferent and afferent nerve endings of the skeletal muscles. J. Neurosci. 8:1625, 1988. 44. Duc, C., Barakat-Walter, I., and Droz, B.: Innervation of putative rapidly adapting mechanoreceptors by calbindin- and

The Muscle Spindle

45. 46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

calretinin-immunoreactive primary sensory neurons in the rat. Eur. J. Neurosci. 6:264, 1994. Eagles, J. P., and Purple, R. L.: Afferent fibers with multiple encoding sites. Brain Res. 77:187, 1974. Eldred, E., Yellin, H., DeSantis, M., and Smith, C. M.: Supplement to bibliography on muscle receptors: their morphology, pathology and physiology. Exp. Neurol. 55:1, 1977. Eldred, E., Yellin, H., Gadbois, L., and Sweeney, S.: Bibliography on muscle receptors: their morphology, pathology and physiology. Exp. Neurol. Suppl 3:1, 1967. El-Tarhouni, A., and Banks, R. W.: The distribution of calretinin in muscle receptors of the cat. J. Physiol. (Lond.) 487:77P, 1995. Emonet-Denand, F., Laporte, Y., Matthews, P. B. C., and Petit, J.: On the subdivision of static and dynamic fusimotor actions on the primary ending of the cat muscle spindle. J. Physiol. (Lond.) 268:827, 1977. Emonet-Dénand, F., Laporte, Y., and Petit, J.: Functional consequences of bag2 and chain fiber coactivation by static -axons in cat spindles. J. Neurophysiol. 77:1425, 1997. Emonet-Dénand, F., Laporte, Y., and Petit, J.: Comparison of static fusimotor innervation in cat peroneus tertius and longus muscles. J. Neurophysiol. 80:249, 1998. Ernfors, P., Lee, K. F., Kucera, J., and Jaenisch, R.: Lack of neurotrophin-3 leads to deficiencies in the peripheral nervous system and loss of limb proprioceptive afferents. Cell 77:503, 1994. Fan, G. P., Copray, S., Huang, E., et al.: Formation of a full complement of cranial proprioceptors requires multiple neurotrophins. Dev. Dyn. 218:359, 2000. Gladden, M. H.: Structural features relative to the function of intrafusal muscle fibres in the cat. In Homma, S. (ed.): Understanding the Stretch Reflex. Amsterdam, Elsevier, p. 51, 1976. Harker, D. W., Jami, L., Laporte, Y., and Petit, J.: Fast conducting skeletofusimotor axons supplying intrafusal chain fibers in the cat peroneus tertius muscle. J. Neurophysiol. 40:791, 1977. Hietanen-Peltola, M., Pelto-Huikko, M., Rechardt, L., et al.: Calbindin-D-28k-immunoreactivity in rat muscle spindle: a light and electron microscopic study. Brain Res. 579:327, 1992. Holm, W., Padeken, D., and Schäffer, S. S.: Characteristic curves of the dynamic response of primary muscle spindle endings with and without gamma stimulation. Pflugers Arch. 391:163, 1981. Houk, J. C., Rymer, W. Z., and Crago, P. E.: Dependence of dynamic response of spindle receptors on muscle length and velocity. J. Neurophysiol. 46:143, 1981. Hulliger, M.: The mammalian muscle spindle and its central control. Rev. Physiol. Biochem. Pharmacol. 101:1, 1984. Hulliger, M., Matthews, P. B. C., and Noth, J.: Static and dynamic fusimotor action on the response of Ia fibres to low frequency sinusoidal stretching of widely ranging amplitude. J. Physiol. (Lond.) 267:811, 1977. Hultborn, H.: State-dependent modulation of sensory feedback. J. Physiol. (Lond.) 533:5, 2001.

149

62. Hunt, C. C.: Mammalian muscle spindle: peripheral mechanisms. Physiol. Rev. 70:643, 1990. 63. Hunt, C. C., and Ottoson, D.: Impulse activity and receptor potential of primary and secondary endings of isolated mammalian muscle spindles. J. Physiol. (Lond.) 252:259, 1975. 64. Hunt, C. C., and Wilkinson, R. S.: An analysis of receptor potential and tension of isolated cat muscle spindles in response to sinusoidal stretch. J. Physiol. (Lond.) 302:241, 1980. 65. Hunt, C. C., Wilkinson, R. S., and Fukami, Y.: Ionic basis of the receptor potential in primary endings of mammalian muscle spindles. J. Gen. Physiol. 71:683, 1978. 66. Iino, S., Kobayashi, S., and Hidaka, H.: Neurocalcinimmunopositive nerve terminals in the muscle spindle, Golgi tendon organ and motor endplate. Brain Res. 808:294, 1998. 67. Ikura, M.: Calcium binding and conformational response in EF-hand proteins. TIBS 21:14, 1996. 68. Jami, L., Murthy, K. S. K., and Petit, J.: A quantitative study of skeleto-fusimotor innervation in the cat peroneus tertius muscle. J. Physiol. (Lond.) 325:125, 1982. 69. Kennedy, W. R., Poppele, R. E., and Quick, D. C.: Mammalian muscle spindles. In Sumner, A. J. (ed.): The Physiology of Peripheral Nerve Disease. Philadelphia, W. B. Saunders, p. 74, 1980. 70. Kröller, J.: Reverse correlation analysis of the stretch response of primary muscle spindle afferent fibers. Biol. Cybern. 69:447, 1993. 71. Kruse, M. N., and Poppele, R. E.: Components of the dynamic response of mammalian muscle spindles that originate in the sensory terminals. Exp. Brain Res. 86:359, 1991. 72. Kucera, J.: Characteristics of motor innervation of muscle spindles in the monkey. Am. J. Anat. 173:113, 1985. 73. Kucera, J., Fan, G. P., Jaenisch, R., et al.: Dependence of developing group Ia afferents on neurotrophin-3. J. Comp. Neurol. 363:307, 1995. 74. Kucera, J., and Walro, J. M.: Origin of intrafusal muscle fibers in the rat. Histochemistry 93:567, 1990. 75. Kucera, J., and Walro, J. M.: Origin of intrafusal fibers from a subset of primary myotubes in the rat. Anat. Embryol. 192:149, 1995. 76. Kucera, J., Walro, J. M., and Reichler, J.: Role of nerve and muscle factors in the development of rat muscle spindles. Am. J. Anat. 186:144, 1989. 77. Kucera, J., Walro, J. M., and Reichler, J.: Neural organization of spindles in three hindlimb muscles of the rat. Am. J. Anat. 190:74, 1991. 78. Kuffler, S. W., Hunt, C. C., and Quilliam, J. P.: Function of medullated small nerve fibres in mammalian ventral roots: efferent muscle spindle innervation. J. Neurophysiol. 14:29, 1951. 79. Leksell, L.: The action potential and excitatory effects of the small ventral root fibres to skeletal muscle. Acta Physiol. Scand. Suppl. 31:1, 1945. 80. Liu, J.-X., Eriksson, P.-O., Thornell, L.-E., and PedrosaDomellöf, F.: Myosin heavy chain composition of muscle spindles in human biceps brachii. J. Histochem. Cytochem. 50:171, 2002. 81. Matsuo, S., Ichikawa, H., Silos-Santiago, I., et al.: Proprioceptive afferents survive in the masseter muscle of trkC knockout mice. Neuroscience 95:209, 2000.

150

Function of the Peripheral Nervous System

82. Matthews, P. B. C.: The differentiation of two types of fusimotor fibre by their effects on the dynamic response of muscle spindle primary endings. Q. J. Exp. Physiol. 47:324, 1962. 83. Matthews, P. B. C.: Evolving views on the internal operation and functional role of the muscle spindle. J. Physiol. (Lond.) 320:1, 1981. 84. McCrea, D. A.: Spinal circuitry of sensorimotor control of locomotion. J. Physiol. (Lond.) 533:41, 2001. 85. Milburn, A.: The early development of muscle spindles in the rat. J. Cell Sci. 12:175, 1973. 86. Milburn, A.: Stages in the development of the cat muscle spindle. J. Embryol. Exp. Morphol. 82:177, 1984. 87. Nahirney, P. C., and Ovalle, W. K.: Distribution of dystrophin and neurofilament protein in muscle spindles of normal and mdx-dystrophic mice: an immunocytological study. Anat. Rec. 235:501, 1993. 88. Olson, C. B., and Swett, C. P.: A functional and histochemical characterization of motor units in a heterogeneous muscle (flexor digitorum longus) of the cat. J. Comp. Neurol. 128:475, 1966. 89. Ovalle, W. K., and Smith, R. S.: Histochemical identification of three types of intrafusal muscle fibers in the cat and monkey based on the myosin ATPase reaction. Can. J. Physiol. Pharmacol. 50:195, 1972. 90. Pedrosa, F., and Thornell, L.-E.: Expression of myosin heavy chain isoforms in developing rat muscle spindles. Histochemistry 94:231, 1990. 91. Pedrosa-Domellöf, F., and Thornell, L.-E.: Expression of myosin heavy chain isoforms in developing human muscle spindles. J. Histochem. Cytochem. 42:77, 1994. 92. Petit, J., Banks, R. W., and Laporte, Y.: Testing the classification of static axons using different patterns of random stimulation. J. Neurophysiol. 81:2823, 1999. 93. Poppele, R. E.: An analysis of muscle spindle behavior using randomly applied stretches. Neuroscience 6:1157, 1981. 94. Poppele, R. E., and Bowman, R. J.: Quantitative description of linear behavior of mammalian muscle spindles. J. Neurophysiol. 23:59, 1970. 95. Poppele, R. E., Kennedy, W. R., and Quick, D. C.: A determination of static mechanical properties of intrafusal muscle in isolated cat muscle spindles. Neuroscience 4:401, 1979. 96. Poppele, R. E., and Quick, D. C.: Stretch-induced contraction of intrafusal muscle in cat muscle spindle. J. Neurosci. 1:1069, 1981.

97. Prochazka, A.: Proprioceptive feedback and movement regulation. In Rowell, L., and Sheperd, J. T. (eds.): Handbook of Physiology. Sect. 12. Exercise: Regulation and Integration of Multiple Systems. New York, American Physiological Society, p. 89, 1996. 98. Quick, D. C., Kennedy, W. R., and Poppele, R. E.: Anatomical evidence for multiple sources of action potentials in the afferent fibres of muscle spindles. Neuroscience 5:109, 1980. 99. Ruffini, A.: On the minute anatomy of the neuromuscular spindles of the cat, and on their physiological significance. J. Physiol. (Lond.) 23:190, 1898. 100. Sacks, R. D., and Roy, R. R.: Architecture of the hind limb muscles of cats: functional significance. J. Morphol. 173:185, 1982. 101. Soukup, T., Pedrosa, F., and Thornell, L.-E.: Influence of neonatal motor denervation on expression of myosin heavy chain isoforms in rat muscle spindles. Histochemistry 94:245, 1990. 102. Soukup, T., Pedrosa-Domellöf, F., and Thornell, L.-E.: Expression of myosin heavy chain isoforms and myogenesis of intrafusal fibres in rat muscle spindles. Microsc. Res. Tech. 30:390, 1995. 103. Swett, J. E., and Eldred, E.: Distribution and numbers of stretch receptors in medial gastrocnemius and soleus muscles of the cat. Anat. Rec. 137:453, 1960. 104. Tourtellotte, W. G., Keller-Peck, C., Milbrandt, J., and Kucera, J.: The transcription factor Egr3 modulates sensory axon-myotube interactions during muscle spindle morphogenesis. Dev. Biol. 232:388, 2001. 105. Tourtellotte, W. G., and Milbrandt, J.: Sensory ataxia and muscle spindle agenesis in mice lacking the transcription factor Egr-3. Nat. Genet. 20:87, 1998. 106. Voss, H.: Untersuchungen über Zahl, Anordnung und Länge der Muskelspindeln in den Lumbricalmuskeln des Menschen und einiger Tiere. Z. Micros. Anat. Forsch. 42:509, 1937. 107. Voss, H.: Tabelle der absoluten und relativen Muskelspindelzahlen der menschlichen Skelettmuskulatur. Anat. Anz. 129:562, 1971. 108. Werle, M. J., Roder, J., and Jeromin, A.: Expression of frequenin at the frog (Rana) neuromuscular junction, muscle spindle and nerve. Neurosci. Lett. 284:33, 2000. 109. Zelená, J.: Nerves and Mechanoreceptors. London, Chapman & Hall, 1994.

7 The Golgi Tendon Organ JON J. A. SCOTT

Tendon Organ Structure Distribution Physiologic Responses Responses to Motor Unit Contractions

Static Sensitivity Dynamic Sensitivity Recordings during Natural Movements Transduction

The control of movement is highly dependent on sensory feedback from the limbs, which arises from a variety of sensory receptors in the skeletal muscles, joints, and skin. Skeletal muscles are supplied with two encapsulated mechanoreceptors that have a significant role in providing movement-related feedback, the Golgi tendon organ and the muscle spindle, which monitor the tension and length of the muscle, respectively. The tendon organ, innervated by a single Ib afferent axon, is comparatively simple in structure and has certainly attracted less attention than the muscle spindle. As a consequence, there are aspects of its physiology that are still poorly understood. This chapter reviews the structure, distribution, and physiology of the tendon organ in normal muscles and following reinnervation. Some aspects of the role of tendon organ feedback in the spinal and supraspinal control of movement are also addressed, particularly in relation to the evolving picture regarding the reflex actions mediated via the Ib inhibitory interneuron and the significant changes in reflex action occurring during locomotion.

TENDON ORGAN STRUCTURE Tendon organs are located at the junction between skeletal muscle fibers and their associated tendon. The organs are enclosed in a connective tissue capsule and range from 0.2 to 1.5 mm in length, depending on the species and the muscle, and are approximately 0.1 mm in diameter. The majority of tendon organs have a fusiform shape, though bifid and occasional trifid forms occur. The body of the

Recovery from Nerve Lesions Reflex Actions Central Projections and Proprioception

organ comprises a bundle of collagen strands that are continuous, at their proximal end, with a group of muscle fibers and, at their distal end, merge with the collagen of the tendon itself (Fig. 7–1). Thus the inserting muscle fibers are in series with the collagen strands and their contractions will generate tension within those strands. The strands of collagen arise individually from muscle fibers but do not remain distinct; rather, they are tightly intertwined and branch and fuse together, creating a network of longitudinal interconnections.8,61,67 Overall, there is a net fusion of strands from the muscle to the tendinous end such that the number of strands merging with the tendon is fewer than that arising from the muscle fibers.61 As a consequence, individual muscle fibers do not have unique mechanical inputs through the body of the organ. This feature may have consequences for the pattern of neural output. There is considerable variation in the number of muscle fibers inserting into the body of individual tendon organs, with a range of 3 to 50 depending on the muscle and the form of the organ, though the majority of studies indicate ranges of 10 to 25.4,22,61,65,67 Given that the muscle fibers comprising each motor unit are randomly distributed throughout the muscle, each muscle fiber inserting into the tendon organ is likely to be derived from a different motor unit. Physiologic studies and the use of the glycogendepletion technique have confirmed that each motor unit that activates a tendon organ normally contributes one or two fibers only to the group of inserting muscle fibers.22,32,67 In a given muscle, the proportion of muscle fibers inserting into tendon organs is relatively small, thus the majority of fibers insert directly into the tendon and lie in 151

152

Function of the Peripheral Nervous System

A

B FIGURE 7–1 The Golgi tendon organ. A, Light micrograph of a tendon organ from human abductor pollicis brevis muscle. The distal ends of the muscle fibers insert into the collagen strands of the organ, which, in turn, merge into the distal tendon of the muscle. The Ib afferent axon enters the capsule and gives rise to myelinated branches, from which unmyelinated terminals derive and innervate the collagen strands. B, Schematic of a tendon organ at the junction of the gastrocnemius muscle and the Achilles tendon. (From Proske, U.: The Golgi Tendon Organ. 1993.)

parallel with the organs. These fibers, therefore, have the potential to reduce the mechanical tension on an organ and to diminish its response by “unloading.”32,71 In addition to these fibers, some muscle fibers that lie outside the

body of the organ insert into the outer surface of the capsule.8,67,72 Other fibers may insert into collagen strands within the organ that lack innervation.72 The modes of action, if any, of these latter groups of fibers is unclear,

153

The Golgi Tendon Organ

although they may act as if they were in parallel with the organ. The body of the organ is enclosed within a connective tissue capsule between 3 and 12 ␮m thick, comprising, in cross-section, 5 to 20 concentrically arranged capsular cells.51,61 At the point of entry of the afferent axon, the capsular cells are continuous with the perineural sheath of the axon. At the proximal end of the organ, the capsule constricts to form a tight collar around the collagen bundle, which also demarcates the most distal extent of the muscle fibers.61 A similar, distal collar marks the merging of the organ with the main tendon. The lumen of the organ is divided into three types of compartment by “septal cells.”61 These compartments are identified as neuronal compartments, which contain only the myelinated axon close to its entry into the capsule; terminal compartments, which contain collagen bundles along with unmyelinated branches and terminals of the afferent axon; and fibrous compartments, which contain only collagen bundles.51,61,72 Blood vessels enter the capsule along with the afferent axon and course between the layers of the capsule into the lumen of the organ, although they do not appear to be present in the terminal compartments.51 The innervation to the tendon organ is usually via a single, large-diameter, myelinated group I afferent axon. This afferent fiber is commonly known as a Ib afferent axon, to distinguish it from the Ia afferent axon that gives rise to the primary endings of the muscle spindle, although there is almost complete overlap in conduction velocity, except for the fastest conducting fibers. In the cat, Ib afferent axons have conduction velocities of 60 to 110 m/s and range in diameter from 7 to 15 ␮m close to the point of entry into the capsule.4,8 In some instances, Ib afferents branch and innervate two, or very occasionally three, tendon organs.20,65 The Ib afferent axon divides into two or three myelinated branches either close to the point of entry or within the neuronal compartment of the capsule (see Fig. 7–1). These myelinated branches subsequently subdivide and lose their myelin, with further branchings of the unmyelinated, preterminal branches in the terminal compartments. The terminal branches intertwine among the network of collagen strands. Close to their point of termination, the branches may become devoid of Schwann cell processes and are covered only by the basal lamina.50,61 The terminals themselves take the form of spirals and sprays that are richly supplied with mitochondria and form intimate associations with the collagen strands.50,51,61 Innervation of tendon organs by myelinated afferent axons of only 2 to 3 ␮m in diameter has also been observed,68 but their function is not known.

DISTRIBUTION Tendon organs are located at the myotendinous junction of skeletal muscles, more commonly at the insertion rather than origin of the muscle. In pennate muscles, the organs

are commonly found near the point of nerve entry in a central zone of the muscle.59,65 Occasionally, tendon organs may be found entirely within the tendon,4 but the responsiveness of such receptors is as yet unknown. Tendon organs may also be found in close association with other receptors, such as paciniform corpuscles and, in particular, muscle spindles, where they are referred to as dyads.46 Tendon organs have been found to be present in almost all skeletal muscles examined, including limb, neck, masticatory, and respiratory muscles.36 It is interesting to note that some particularly specialized muscles, such as the cat tenuissimus,52 the human orbicularis oris (J. Scott, unpublished data), and sometimes the small hand muscles such as the lumbricals,14 appear to be devoid of tendon organs. Detailed counts of tendon organs in individual muscles are rare and, unfortunately, have not been carried out systematically. In almost all instances, the number of tendon organs is lower than that of spindles and shows considerable variation: In a survey of 10 cat hind limb muscles, the proportion of tendon organs to spindles ranged from 0.33 to 0.94.36 Interestingly, the limited information available from primate and human intrinsic hand muscles indicates relatively low ratios. Thus for the abductor pollicis, in which there are tendon organ counts of seven and eight in the bonnet monkey and human, respectively, as well as for the lumbricals, which likewise have counts between zero to one and five, respectively14 (J. Scott, unpublished data), the ratios of tendon organs to spindles are below 0.3. Because tendon organs are specifically responsive to the tension generated by muscle contraction, a more meaningful way of expressing receptor complement is as a ratio of the number of motor units in the muscle. This has been done for a number of cat hind limb muscles, giving a range of 1.9 to 9.0 motor units per tendon organ.36 Given that there are typically between 10 and 25 muscle fibers inserting into each tendon organ (see above), this implies that, taken together, the population of tendon organs of the muscles should be sufficient to monitor the contractile activity of all the motor units.

PHYSIOLOGIC RESPONSES The tendon organ lies in series with the muscle fibers that insert into it. Therefore, it can be stretched by stretch or contraction of the muscle. However, the organ has a high threshold to the passive tensions generated by wholemuscle stretch and responds with low rates of firing; indeed, a significant proportion of organs have been shown to be insensitive to stretch within the physiologic range.33 By contrast, tendon organs are sensitive to active contractions of the whole muscle. Furthermore, Houk and Henneman32 first demonstrated that a given tendon organ shows a high sensitivity to the tension generated by active contraction of a small number of individual motor units and

154

Function of the Peripheral Nervous System

fails to respond to the contractions of other units. The range of numbers of response-generating motor units (4 to 15) was found to be approximately the same as that of muscle fibers inserting into the organs. Therefore, each tendon organ serves to sample the contractile tension generated by a small population of motor units,5 each of which contributes one or two muscle fibers to the group that inserts into the body of the organ.22,67 The force threshold for the response to the contraction of single motor units may be as low as a few milligrams.6 Measurements of the sensitivities of isolated cat tendon organ–muscle fiber preparations have given values of 335 impulses/s/g.17 The majority of the motor units within a muscle, when stimulated separately, do not activate an individual tendon organ because none of the muscle fibers comprising these units inserts into the organ. Such in-parallel units have the potential to “unload” the tendon organ32 by reducing the relative tension being channeled through the organ. In theory, activation of an in-parallel motor unit could reduce the response of the organ to an in-series unit. The physiologic significance of such unloading effects is unclear.36 Stuart et al.69 showed that the firing rate of a tendon organ responding to the contraction of an in-series motor unit could be reduced when an in-parallel unit was contracting at the same time.7,31,69 There is also evidence of reciprocal unloading effects elicited by two in-series motor units.31 Such effects would imply that the response of the tendon organ was dependent on the combination of motor units

contracting concurrently and not directly on the force being applied to the organ itself. In support of this, it has been shown that the firing rate of a tendon organ to simultaneous contraction of all the in-series motor units is often higher than that during contraction of the muscle as a whole.5 This phenomenon of unloading has been cited as one of the possible factors underlying the apparent nonlinearity of tendon organ responses17 (see below). However, Gregory and co-workers25–27 reported that the reductions in firing rate were only transient during tetanic contractions. They attributed this to the initial lowering of tension by the in-parallel fibers, which was then recovered as the in-series fibers shortened further.

Responses to Motor Unit Contractions Tendon organs are typically silent in the passive muscle except at extreme muscle lengths. In response to tetanic contraction of an in-series motor unit, the organ will respond with a train of action potentials. There is an initial, high-frequency burst during the rising phase of the tension, followed by adaptation to a plateau rate of firing during the steady-state tension (Fig. 7–2A). The tendon organ response may therefore be characterized by a static sensitivity, the sensitivity shown to steady-state forces, and a dynamic sensitivity, the sensitivity shown to changing force levels. Complications inherent in measuring these 80

Static discharge (impulses s–1)

Impulses s–1

60

40

20

10 g

FR FR FF

60

FR S 40

FF FF FF

20

0 0

20

30

40

50

Plateau tension (g)

0.5 s

A

10

B FIGURE 7–2 Responses of a tendon organ afferent to muscle contraction. A, Instantaneous frequency discharges of a Ib afferent in response to isometric contraction of an activating FR motor unit in the peroneus tertius muscle of the cat. The motor unit was stimulated at 60, 80, 100, and 150/s, and the responses have been overlaid. B, Response relationships between the afferent firing frequency of a tendon organ and the plateau tensions developed by the individual motor units that inserted into the organ. (From Petit, J., Davies, P., and Scott, J. J. A.: Static sensitivity of tendon organs to tetanic contraction of in-series motor units in feline peroneus tertius muscle. J. Physiol. [Lond.] 481:177, 1994, with permission of Cambridge University Press.)

155

The Golgi Tendon Organ

sensitivities and the extent of the apparent nonlinearities have given rise to significant debate and also to difficulties in interpreting the role of tendon organs in the feedback control of muscle contraction.36 Motor units in skeletal muscles vary in terms of their muscle fiber complement and their contractile properties. In terms of the contractile properties, they are classified into three groups, FF, FR, and S.10 The FF units are fast contracting and fast fatiguable, whereas the S units are slow contracting and very fatigue resistant, with the FR units being intermediate in terms of both fatigability and contraction speed. These units also differ in terms of the numbers of muscle fibers, with the FF units having the most and the S units the fewest muscle fibers. Analyses of the motor-unit composition of the muscle fibers inserting into tendon organs have revealed a heterogeneity that reflects that of the muscle as a whole.26,27,37,38,58 Thus a detailed analysis of the motor unit input to a tendon organ in the peroneus tertius muscle of the cat showed it was activated by 11 motor units: 5 FFs, 4 FRs, and 2 Ss.55

Static Sensitivity If tendon organs are to provide the central nervous system with meaningful information regarding muscle force, then it might be expected that there should be consistent relationships between the discharge frequency and the contractile force generated by activating motor units. Tetanic contractions, applied under isometric conditions, have frequently been employed to investigate the sensitivity of tendon organs, because the plateau tension provides a constant stimulus that elicits a steady discharge from individual organs. Under such conditions, a number of studies have shown that there was no consistent relationship between force and discharge frequency; thus a small, type S motor unit might elicit a higher discharge than a powerful FF unit.5,26,32,37,38,58 Furthermore, activation of a motor unit that inserts into more than one tendon organ may give rise to different frequencies of discharge from the different organs,38 and combined stimulation of pairs of activating motor units does not appear to elicit a summed response.27 Fukami17 showed that, for isolated tendon organs, there is a linear relationship between the static discharge and the contractile tension generated by a single, inserting muscle fiber. This has also been shown to hold true for the relationship between discharge frequency and the plateau tension generated by individual activating motor units55 (Fig. 7–2B), although there were differences between the sensitivities to different motor units, the sensitivity being greatest for type S and least for FF motor units. When motor units were activated in combination, linear relationships were again recorded for the discharge frequency against tension for each specific combination. The slopes of the relationships were shallower than for the single units and,

with the exception of combinations of type S units, tended to be similar.55 Measurements of the static sensitivities of a group of tendon organs to the tetanic contractions of single and combined motor units showed that the sensitivities were high for the lowest tensions and dropped rapidly to a steady level of approximately 100 impulse/s/N that was independent of the size and number of the activating units.55 The relationship between sensitivity and contractile tension may be attributable to the unloading effect of the in-parallel muscle fibers of the activating motor units in combination with the compliance of the tendon; this unloading effect will be largest for the largest motor units, and therefore, when they are stimulated singly, they may show lower sensitivities than the smaller units. Beyond a certain level of tension, for large units or as motor units are combined, the unloading resulting from the compliance of the tendon will approach a constant level because the stiffness of the tendon will increase as it is stretched until it reaches a high, constant level.57 As a consequence, the sensitivity of the organs will also approach a low, constant level as reported by Petit et al.55 A further consequence of this unloading effect is that, when motor units are added together, for the first few units the order in which that addition takes place will affect the sensitivity of the organ to the applied tension.26 Although tetanic, isometric contractions represent a simple method for investigating the steady-state input to a tendon organ, they do not constitute a representative physiologic condition, particularly because the motor units are typically stimulated at rates well above those occurring naturally. An alternative approach has been to investigate isotonic contractions, with ongoing muscle shortening or lengthening. These have been generated with groups of motor units stimulated asynchronously, so that a smooth tension profile can be elicited while stimulating the individual motor units at physiologic rates. The responses of the tendon organs were again found to be proportional to the active tension, with the same sensitivity as for the isometric contractions, indicating that the tendon organ has the capacity to provide an accurate signal of the contractile force being applied by the contracting muscle under conditions that more closely reflect natural movements.56

Dynamic Sensitivity During the onset of a contraction, the discharge rate of a tendon organ is higher than that for the equivalent steadystate tension and is related to the rate of rise of the tension. This dynamic sensitivity has particularly been investigated under isometric conditions using rates of motor unit stimulation that generate unfused contractions. A common observation under such conditions is that the tendon organ will respond with an action potential for each rising phase

156

Function of the Peripheral Nervous System

in the tension oscillation,30,39 with such 1:1 responses being referred to as “driving.” One consequence of this dynamic sensitivity and of the phenomenon of driving is that, when a motor unit is stimulated at rates below the fusion frequency for the contraction (up to approximately 40 impulses/s for FF units), the rate of tendon organ discharge will reflect the oscillation frequency rather than the underlying absolute tension. For example, increasing the stimulation frequency from 10 to 20 impulses/s will double the frequency of the force oscillations and therefore double the rate of discharge, even though the mean force level changes by a differing amount.39 The effects of 1:1 driving can also be observed when pairs of motor units are activated asynchronously.30 Because these stimulation rates reflect the main physiologic range of motor units firing, there is the implication that the dynamic sensitivity could dominate the feedback to the central nervous system and therefore, particularly at low levels of contractile force, the organs would be signaling force variation rather than the absolute force levels.36 The nature of the dynamic response appears complex. In a passive muscle at physiologic length, the tendon organs are usually silent. During the rising phase of a tetanic contraction, the sensitivity changes, with an initial high peak in discharge frequency that then declines, even though the absolute tension is still rising, until the steadystate levels are achieved. Early models fitted the response with two exponential components, the first with a fast and the second with a slower time constant.31 Likewise, during the unfused contractions, the initial peak at the onset of the first contraction is higher than the maintained driving response.39 Further analysis of the dynamic response has shown that, for each motor unit–tendon organ combination, the initial instantaneous frequency for the first two impulses (assuming that the tendon organ was not firing before the contraction onset) is linearly related to the rate of rise of the tension, the slope of the relationship being steepest for the type S units. This initial dynamic sensitivity is absent if the tendon organ is already discharging before the onset of the contraction, or if the activating motor unit is repetitively stimulated at relatively short intervals. When stretches are applied to isolated tendon organs, the receptor potential shows an initial overshoot with respect to the applied tension and the discharge rate shows a further overshoot that is dependent on the rate of rise of the receptor potential.18 This latter overshoot was attributed to the accommodative properties of the pacemaker site such that the threshold for action potential generation decreases with increasing rate of rise of the receptor potential13,18 and is absent if the tendon organ is already discharging before the stimulus onset. The second component of the dynamic response is exponentially related to the rate of rise of the tension, the sensitivity being independent of motor unit size.13

Recordings during Natural Movements Although the studies described above have enabled development of significant understanding of the patterns of response to a range of stimulation paradigms, the ideal would be to be able to analyze the responses of tendon organs during natural movements. Unfortunately, this is technically very demanding and, compared with recordings from muscle spindles, there is a paucity of recordings from tendon organs during natural movements. The majority of the available data have come from two sources, recordings made from cats using chronic implants and microneurographic recordings from human subjects. The recordings from cats have shown that the Ib afferents discharge in relation to the electromyographic (EMG) activity within the muscles and therefore to the occurrence of contractile activity.2,43 During stepping movements, Ib afferents from ankle extensors showed maximal discharge during the loading periods of the stance phase.2 During smooth increases in contractile force, the tendon organ discharges were commonly marked by stepwise increments in firing rate, which were attributed to the increments in the mechanical input to the tendon organ resulting from motor unit recruitment.2 Similar steps in discharge were also observed during the generation of crossed-extensor reflexes.12 The development of the microneurographic technique has enabled recordings to be made from single nerve fibers in human subjects. Although the numbers of tendon organs from which recordings have been made is relatively small, the findings have confirmed those from animal studies in that the organs respond to muscle contraction, as evidenced by EMG activity, and show a low sensitivity to muscle stretch.1,15 Although the firing rates appear to be relatively low, the organs display a high sensitivity to variations in the contractile activity, as demonstrated by their responsiveness to the pulsatile motor activity that has been recorded during slow finger movements.70 During progressively developing contractions, recordings from human Golgi tendon organs have also shown steps in the firing rate that coincided with increments in the contractile force and may be attributable to motor unit recruitment.15 Interestingly, some but not all of these steps showed an initial dynamic peak even though the Ib afferent was already discharging (see below).

TRANSDUCTION As has already been described, the adequate stimulus for tendon organs is contraction of the inserting muscle fibers that stretches the collagen strands. This stretching deforms the unmyelinated terminals of the Ib afferent axon that are wrapped around the collagen strands, giving rise to a receptor potential resulting from conductance

157

The Golgi Tendon Organ

changes in the deformed terminals. Early models of tendon organ function proposed that, as the tendon organ is stretched, so the axon terminals are compressed between the collagen strands, giving rise to the mechanical distortion and hence the change in membrane potential.8 Although no direct measurements of the membrane properties and ion channel openings have been made, indirect measurements of the receptor potential in isolated tendon organs indicate that stretch itself is the mechanical stimulus.18,71 Under steady-state conditions, the receptor potentials were found to be linearly related to the stretch applied to the organ and the sensitivity was also related to the compliance of the organ, indicating that stretch itself was the activating stimulus.18 During sinusoidal stretching, wherein the dynamic sensitivity becomes apparent, the relationship between the receptor potential and the applied stretch was nonlinear for all but the smallest amplitude stretches, and appeared dependent both on the mechanics of the organ and on the ionic processes within the terminals.18 The Ib afferent axon branches several times before giving rise to the terminals that are associated with the collagen strands. The first- and second-order branches of the axon are commonly myelinated,61 indicating a capacity to conduct action potentials. This raises the question as to where the site of action potential generation is located, because the terminal node of each branch could act as a pacemaker site. At one extreme, the receptor potential could spread to the most proximal branch point, so that a single train of action potentials is generated representing the summation of the receptor potentials produced in the separate branches. Alternatively, each branch could conduct action potential trains that then interact at one or more branch points. Where there are multiple pacemaker sites, an action potential arising in one branch and arriving at the branching node can be conducted up the parent branch but can also be conducted antidromically down the other daughter branch. In this case, the action potential will collide with any action potentials traveling orthodromically in the second daughter branch. As a consequence, the pattern of discharge recorded in the main ascending branch will depend on the relative firing rates in the daughter branches and will range from probabilistic mixing, giving rise to a degree of summation, to complete occlusion of one branch’s activity. Detailed analysis of the discharge patterns of the Ia afferents from muscle spindles has shown evidence of a range of competitive pacemaker interactions resulting in varying levels of occlusion.3 When two motor units are stimulated together, the frequency of action potential discharge in the Ib afferent axon to the combined stimulation is less than the arithmetic sum of the responses to the units stimulated separately. This has been proposed as evidence of pacemaker interactions,27 although it could also be due to unloading effects.17,55

Further evidence for the operation of multiple pacemaker sites derives from the occurrence of cross-adaptation. As described above, the dynamic response shows an initial peak that is dependent on the accommodative properties of the pacemaker site.13,18 Thus, if an activating motor unit is stimulated twice, tetanically, the tendon organ’s response to the second contraction lacks an initial dynamic peak as a result of adaptation of the pacemaker threshold. However, if the organ is subjected to successive contractions of two different activating motor units, the dynamic peak may or may not be present.26,27 If the two motor units stretch collagen strands that receive innervation from different branches of the Ib afferent axon, then each of these branches may have its own pacemaker site, such that adaptation of one site by the first contraction does not influence the sensitivity of the second site, which displays a full dynamic response to the contraction of the second unit. Conversely, if the two collagen strands are innervated by terminals arising from the same main axon’s branch, then the initial dynamic peak will be absent from the response to the second unit. These findings have been successfully replicated with computer models that incorporate multiple pacemaker sites.24

RECOVERY FROM NERVE LESIONS Two main types of nerve injury have been investigated experimentally, nerve crush (axonotmesis) and nerve transection (neurotmesis) with subsequent repair. These differ significantly in their outcomes. Following nerve crush, there is usually a high proportion of axons that regenerate and successfully reinnervate their original target sites. After nerve section, however, many axons fail to regenerate successfully and so a significant proportion of target sites remains uninnervated. Where reinnervation does take place, it is highly unlikely to be effected by the original axon; indeed, the reinnervation may be by an axon that previously innervated a different type of end organ. The tendon organ is relatively resistant to denervation atrophy, and after nerve crush injury there is usually successful restoration of the Ib afferent ending, although it is reduced in extent.35,62 During the early stages of recovery, the Ib afferents frequently fail to maintain discharge throughout the plateau phase of a tetanic contraction, with a reduced static sensitivity, but these gradually recover to normal levels.63 The two phases of the dynamic response appear to be affected differentially; the second phase, which shows an exponential relationship to the rate of rise of the tension (see above) is the same as normal. The initial sensitivity, which is additionally dependent on the accommodative properties of the pacemaker site, is reduced during the early stages but also recovers toward normal.63 Following nerve transection, the pattern of recovery is much poorer, with many tendon organs remaining

158

Function of the Peripheral Nervous System

uninnervated. Where reinnervation does occur, the endings are frequently abnormal in appearance.62 There are three confounding factors that affect the outcome of the reinnervation: first, the effects of the denervationreinnervation process itself on the properties of the afferent axon, as, for example, the effects observed following nerve crush injury; second, the effects of reinnervation by inappropriate axons; and third, the effects consequent on the reorganization of the motor units. Following nerve transection, there is frequently grouping of the muscle fibers comprising individual motor units.41 This can significantly alter the mechanical input to a tendon organ; for example, all the muscle fibers inserting into an organ could be derived from a single motor unit, or a group of muscle fibers comprising a reinnervated motor unit could lie immediately adjacent to an organ and thereby generate significant unloading.62 Recordings from reinnervated tendon organs display a range of abnormalities, including phasic-only responses and on-off responses, in which the afferent axon discharges one or two action potentials at the onset and relaxation of the contraction but not during the maintained phase, and off-only responses. It is noteworthy that tendon organs showing abnormal responses to some motor units could also display a normal pattern of response to the contractions of other units, suggesting that the abnormalities in the mechanical input do play the major part in determining the patterns of response following reinnervation.63,64 What is unclear at present is what impact the significant disruptions in proprioceptive feedback, resulting from axonal mislocation and abnormal patterns of response, have on the control of movement.

REFLEX ACTIONS Investigation of the spinal interactions mediated by Ib afferents has been complicated by the problems associated with trying to separate the effects of Ib actions from those of other afferents, particularly the group Ia muscle spindle afferents.36 Many early studies used the technique of graded electrical stimulation of the muscle nerves of the cat hind limb, in decerebrate preparations, to investigate the reflex actions of the muscle afferents. These studies showed that, when the stimulation intensity was progressively elevated above the threshold for Ia afferents, and therefore above the threshold for eliciting the monosynaptic stretch reflex, there was evidence of a disynaptic inhibitory action on the homonymous and some synergic muscles and of facilitation of antagonists. This disynaptic inhibition represented the classic autogenetic inhibition that was proposed to be mediated by the group Ib afferents of the tendon organ.21,42 More detailed examination showed that the Ib interactions to the spinal cord do not mirror the Ia actions, in the form of an inverse

myotatic reflex, but rather that the distribution of inhibitory actions is more widespread and the facilitation of antagonists more variable. In particular, the inhibition arising from flexor muscles tends to be weaker than that from extensors.29 The picture of a simple autogenetic inhibitory reflex, mediated via an interneuron—the “Ib inhibitory interneuron,” which provides a negative feedback regulation of contractile force—has undergone considerable revision over recent years. Detailed investigations of the inputs to the Ib inhibitory interneuron in the lumbar segments of the cat spinal cord have shown that this interneuron receives a wide range of converging inputs, which include segmental, excitatory inputs from Ia, joint, and cutaneous afferents as well as Ib afferent inputs. There are also descending excitatory inputs from the corticospinal and rubrospinal tracts and inhibitory inputs from the dorsal and noradrenergic reticulospinal systems.28,34,36 Because of the pattern of convergence on the Ib inhibitory interneuron, which includes Ia and Ib inputs, and the target projections, this pathway may be more appropriately termed nonreciprocal group I inhibition.28 In addition to the widespread capacity for modulating the Ib reflex via the converging inputs, studies of walking movements have demonstrated a more dramatic pattern of Ib action. During the stance phase of locomotion in the cat, when the extensor muscles are active, the Ib afferent feedback contributes to a positive feedback loop that reinforces the extensor muscle action during the weight-bearing phase of walking54; the Ib action shows a change in sign from inhibition to excitation. Furthermore, the Ib feedback from the ankle extensors has an inhibitory action on the flexors, preventing the initiation of the swing phase until the extensor muscles are unloaded as weight is transferred to the contralateral foot.53,54 During a lengthening contraction of the extensors in the stance phase, the positive feedback from the Ib loop can act in concert with the negative feedback Ia-mediated stretch reflex to compensate for the loading.66

CENTRAL PROJECTIONS AND PROPRIOCEPTION The afferent feedback from tendon organs is relayed to both the cerebellum and the cerebral cortex, and therefore might be predicted to have input to central processing of both motor control and proprioception. The cerebellum receives tendon organ inputs from the hind limb via the dorsal spinocerebellar tract (DSCT) and the ventral spinocerebellar tract, which enter the cerebellum via the inferior peduncle and terminate as mossy fibers in the anterior lobe and paramedian lobule. Forelimb input is relayed via the cuneocerebellar and rostral spinocerebellar tracts. Climbing fiber input is routed via the inferior olive.16

159

The Golgi Tendon Organ

The best studied of the cerebellar projections relaying tendon organ feedback is that of the DSCT in the cat, originating in Clarke’s column. The DSCT receives inputs from a variety of muscle, joint, and cutaneous afferents, but the different inputs are segregated to different groups of DSCT cells.44 The input to the DSCT neurons takes a variety of forms. In a study of the effects of weak muscle contraction, just over 50% of the DSCT neurons tested were contraction sensitive, indicating a wide distribution of the muscle-afferent inputs. However, the effects were strongly contrasting, being separable into four main types. The first group of DSCT cells showed excitation that persisted throughout the contraction and was attributed to monosynaptic inputs from Ib afferents. The second group showed declining excitation during the contraction, probably as a consequence of presynaptic inhibition. The third and largest group showed declining inhibition that could be attributed to disynaptic inputs via the Ib inhibitory interneuron, and the fourth group displayed mixed effects.73 Clearly, the cerebellum receives a complex variety of information about ongoing muscle contractions, but how this information is processed within the cerebellum in terms of regulation of movement and muscle tone is yet to be elucidated. In addition to the cerebellar projection, the Ib afferent feedback is also relayed to area 3a, in the somatosensory cortex, of the cerebral cortex. This projection in the cat is routed via collaterals of the dorsal spinocerebellar axons that synapse in nucleus Z and are relayed through the thalamus.48,49 The question therefore arises as to whether the Ib feedback contributes to conscious proprioception and generates some form of sense of force or effort or whether these arise from corollary discharges. Stimulation of populations of low-threshold muscle afferents from the hand gives rise to a short-latency cortical potential and illusions of muscle stretch.19 At the singlefiber level such perceptions are very rare, though in a sample of three Ib afferents, microstimulation of one of the afferents did elicit a sensation of movement.45 Roland and Ladegaard-Pedersen60 demonstrated that a “sense of tension” exists that persists during skin anesthesia and during partial muscle paralysis, which disturbs the relationship between motor output and achieved contractile force. Such a sense of tension should therefore originate from the muscle receptors. However, a number of other studies suggest that the preferential source of information is derived centrally; for example, the perception that the weight of an object increases with developing fatigue40,47 and the sense of force is unaffected by changing the initial muscle length for maximal isometric contractions, thereby altering the actual achieved force.11 Eccentric exercise also results in errors of force estimation,9 but these changes are not matched by changes in tendon organ sensitivity,23 again implying that the dominant information contributing to the sense of force is that derived from the centrally generated

motor commands rather than the peripheral feedback from the tendon organs.

ACKNOWLEDGMENTS My thanks are extended to Dr. Revers Donga for his helpful criticisms of the manuscript.

REFERENCES 1. Al-Falahe, N. A., Nagaoka, M., and Vallbo, A. B.: Response profiles of human muscle afferents during active finger movements. Brain 113:325, 1990. 2. Appenteng, K., and Prochazka, A.: Tendon organ firing during active muscle lengthening in awake, normally behaving cats. J. Physiol. (Lond.) 353:81, 1984. 3. Banks, R. W., Hulliger, M., Scheepstra, K. A., and Otten, E.: Pacemaker activity in a sensory ending with multiple encoding sites: the cat muscle spindle primary ending. J. Physiol. (Lond.) 498:177, 1997. 4. Barker, D.: The morphology of muscle receptors. In Hunt, C. C. (ed.): Handbook of Sensory Physiology, Vol. 3. Berlin, Springer, p. 1, 1985. 5. Binder, M. D.: Further evidence that the Golgi tendon organ monitors the activity of a discrete set of motor units within a muscle. Exp. Brain Res. 43:186, 1981. 6. Binder, M. D., Kroin, J. S., Moore, G. P. and Stuart, D. G.: The response of Golgi tendon organs to single motor unit contractions. J. Physiol. (Lond.) 271:337, 1977. 7. Binder, M. D., and Osborn, C. E.: Interactions between motor units and Golgi tendon organs in the tibialis posterior muscle of the cat. J. Physiol. (Lond.) 364:199, 1985. 8. Bridgman, C. F.: The structure of tendon organs in the cat: a proposed mechanism for responding to muscle tension. Anat. Rec. 162:209, 1968. 9. Brockett, C., Warren, N., Gregory, J. E., et al.: A comparison of the effects of eccentric exercise on force and position sense at the human elbow joint. Brain Res. 771:251, 1997. 10. Burke, R. E., Levine, D. W., Tsairis, P., and Zajac, F. E.: Physiological types and histochemical profiles in motor units of the cat gastrocnemius. J. Physiol. (Lond.) 234:723, 1973. 11. Caferelli, E., and Bigland-Ritchie, B.: Sensation of static force in muscles of different length. Exp. Neurol. 65:511, 1979. 12. Crago, P. E., Houk, J. C., and Rymer, W. Z.: Sampling of total muscle force by tendon organs. J. Neurophysiol. 47:1069, 1982. 13. Davies, P., Petit, J., and Scott, J. J. A.: The dynamic response of Golgi tendon organs to tetanic contraction of in-series motor units. Brain Res. 690:82, 1995. 14. Devanandan, M. S., Ghosh, S., and John, K. T.: A quantitative study of muscle spindles and tendon organs in some intrinsic muscles of the hand in the bonnet monkey (Macaca radiata). Anat. Rec. 207:263, 1983. 15. Edin, B., and Vallbo, A. B.: Muscle afferent responses to isometric contractions and relaxations in humans. J. Neurophysiol. 63:1307, 1990.

160

Function of the Peripheral Nervous System

16. Ekerot, C. F., Larson, B., and Oscarsson, O.: Information carried by the spinocerebellar paths. Prog. Brain Res. 50:79, 1979. 17. Fukami, Y.: Responses of isolated Golgi tendon organs of the cat to muscle contraction and electrical stimulation. J. Physiol. (Lond.) 318:429, 1981. 18. Fukami, Y., and Wilkinson, R. S.: Responses of isolated Golgi tendon organs of the cat. J. Physiol. (Lond.) 265:673, 1977. 19. Gandevia, S. C., Burke, D., and McKeon, B.: The projection of muscle afferents from the hand to cerebral cortex in man. Brain 107:1, 1984. 20. Goldfinger, M. D., and Fukami, Y.: Distribution, density and size of muscle receptors in cat tail dorsolateral muscles. J. Anat. 135:371, 1982. 21. Granit, R.: Reflexes of self-regulation of muscle contraction and autogenetic inhibition. J. Neurophysiol. 13:351, 1950. 22. Gregory, J. E.: Relations between identified tendon organs and motor units in the medial gastrocnemius muscle of the cat. Exp. Brain Res. 81:602, 1990. 23. Gregory, J. E., Brockett, C. L., Morgan, D .L., et al.: Effect of eccentric muscle contractions on Golgi tendon organ responses to passive and active tension in the cat. J. Physiol. (Lond.) 538:209, 2002. 24. Gregory, J. E., Morgan, D. L., and Proske, U.: Site of impulse initiation in tendon organs of cat soleus muscle. J. Neurophysiol. 54:1383, 1985. 25. Gregory, J. E., Morgan, D. L., and Proske, U.: The discharge of cat tendon organs during unloading contractions. Exp. Brain Res. 61:222, 1986. 26. Gregory, J. E., and Proske, U.: The responses of Golgi tendon organs to stimulation of different combinations of motor units. J. Physiol. (Lond.) 295: 251, 1979. 27. Gregory, J. E., and Proske, U.: Motor unit contractions initiating impulses in a tendon organ in the cat. J. Physiol. (Lond.) 313:251, 1981. 28. Harrison, P. J., and Jankowska, E.: Sources of input to interneurones mediating group I non-reciprocal inhibition of motoneurones in the cat. J. Physiol. (Lond.) 361:403, 1985. 29. Harrison, P. J., Jankowska, E., and Johannison, T.: Shared reflex pathways of group I afferents in different hindlimb muscles. J. Physiol. (Lond.) 338:113, 1983. 30. Horcholle-Bossavit, G., Jami, L., Petit, J., et al.: Activation of tendon organs by asynchronous contractions of motor units in cat leg muscles. Neurosci. Lett. 103:44, 1989. 31. Horcholle-Bossavit, G., Jami, L., Petit, J., et al.: Unloading of tendon organ discharges by in-series motor units in cat peroneal muscles. J. Physiol. (Lond.) 408:185, 1989. 32. Houk, J. C., and Henneman, E.: Responses of Golgi tendon organs to active contractions of the soleus muscle of the cat. J. Neurophysiol. 30:466, 1967. 33. Houk, J. C., Singer, J. J., and Henneman, E.: Adequate stimulus for tendon organs with observations on mechanics of ankle joint. J. Neurophysiol. 34:1051, 1971. 34. Hultborn, H.: State-dependent modulation of sensory feedback. J. Physiol. (Lond.) 533:5, 2001. 35. Ip, M. C., Vrbova, G., and Westbury, D.: The sensory reinnervation of hind-limb muscles after denervation and deefferentation. Neuroscience 2:423, 1977.

36. Jami, L.: Golgi tendon organs in mammalian skeletal muscle: functional properties and central actions. Physiol. Rev. 72:623, 1992. 37. Jami, L., and Petit, J.: Frequency of tendon organ discharges elicited by the contraction of motor units in cat leg muscles. J. Physiol. (Lond.) 261:633, 1976. 38. Jami, L., and Petit, J.: Heterogeneity of motor units activating single Golgi tendon organs in cat leg muscles. Exp. Brain Res. 24:485, 1976. 39. Jami, L., Petit, J., Proske, U., and Zytnicki, D.: Responses of tendon organs to unfused contractions of single motor units. J. Neurophysiol. 53:32, 1985. 40. Jones, L. A., and Hunter, I. W.: Perceived force in fatiguing isometric contractions. Percept. Psychophys. 33:369, 1983. 41. Karpati, G., and Engel, W. K.: Type grouping in skeletal muscles after experimental reinnervation. Neurology 18: 447, 1968. 42. Laporte, Y., and Lloyd, D. P. C.: Nature and significance of the reflex connections established by large afferent fibers of muscular origin. Am. J. Physiol. 169:609, 1952. 43. Loeb, G. E.: Somatosensory unit input to the spinal cord during normal walking. Can. J. Physiol. Pharmacol. 59:627, 1981. 44. Lundberg, A., and Oscarsson, O.: Functional organisation of the dorsal spinocerebellar tract in the cat. IV. Synaptic connections of afferents from Golgi tendon organs and muscle spindles. Acta Physiol. Scand. 38:53, 1956. 45. Macefield, G., Gandevia, S. C., and Burke, D.: Perceptual responses to microstimulation of single afferents innervating joints, muscles and skin of the human hand. J. Physiol. (Lond.) 429:113, 1990. 46. Marchand, E. R., Bridgman, C. F., Shumpert, E., and Eldred, E.: Association of tendon organs with spindles in muscles of the cat’s leg. Anat. Rec. 169:23, 1971. 47. McCloskey, D. I., Ebeling, P., and Goodwin, G. M.: Estimation of weights and tensions and apparent involvement of a “sense of effort”. Exp. Neurol. 42:220, 1974. 48. Mcintyre, A. K., Proske, U., and Rawson, J. A.: Central projection of afferent information from tendon organs in the cat. J. Physiol. (Lond.) 354:395, 1984. 49. Mcintyre, A. K., Proske, U., and Rawson, J. A.: Pathway to the cerebral cortex for impulses from tendon organs in the cat’s hindlimb. J. Physiol. (Lond.) 369:115, 1985. 50. Merrillees, N. C. R.: Some observations on the fine structure of a Golgi tendon organ of a rat. In Barker, D. (ed.): Symposium on Muscle Receptors. Hong Kong, Hong Kong University Press, p. 199, 1962. 51. Nitatori, T.: The fine structure of human Golgi tendon organs as studied by three-dimensional reconstruction. J. Neurocytol. 17:27, 1988. 52. Palmer, J. M., and Sitwell, D. L.: Analysis of innervation of tenuissimus muscle of the cat. Anat. Rec. 130:434, 1958. 53. Pearson, K. G.: Could enhanced reflex function contribute to improving locomotion after spinal cord repair? J. Physiol. (Lond.) 533:75, 2001. 54. Pearson, K. G., and Collins, D. F.: Reversal of the influence of group Ib afferents from plantaris on activity in medial gastrocnemius muscle during locomotor activity. J. Neurophysiol. 70:1009, 1993.

The Golgi Tendon Organ 55. Petit, J., Davies, P., and Scott, J. J. A.: Static sensitivity of tendon organs to tetanic contraction of in-series motor units in feline peroneus tertius muscle. J. Physiol. (Lond.) 481:177, 1994. 56. Petit, J., Scott, J. J. A., and Reynolds, K. J.: Tendon organ sensitivity to steady-state isotonic contraction of in-series motor units in feline peroneus tertius muscle. J. Physiol. (Lond.) 500:227, 1997. 57. Proske, U., and Morgan, D. L.: Tendon stiffness: methods of measurement and significance for the control of movement. A review. J. Biomech. 20:75, 1987. 58. Reinking, R. M., Stephens, J. A., and Stuart, D. G.: The tendon organs of cat medial gastrocnemius: significance of motor unit type and size for the activation of Ib afferents. J. Physiol. (Lond.) 250:491, 1975. 59. Richmond, F. J. R., and Stuart, D. G.: Distribution of sensory receptors in the flexor carpi radialis muscle of the cat. J. Morphol. 183:1, 1985. 60. Roland, P. E., and Ladegaard-Pedersen, H.: A quantitative analysis of sensations of tension and of kinaesthesia in man. Brain 100:671, 1977. 61. Schoultz, T. W., and Swett, J. E.: The fine structure of the Golgi tendon organ. J. Neurocytol. 1:1, 1972. 62. Scott, J. J. A.: The functional recovery of muscle proprioceptors after peripheral nerve lesions. J. Peripher. Nerv. Syst. 1:19, 1996. 63. Scott, J. J. A., Davies, P., and Petit, J.: The static sensitivity of tendon organs during recovery from nerve injury. Brain Res. 697:225, 1995.

161

64. Scott, J. J. A., Davies, P., and Petit, J.: The dynamic response of feline Golgi tendon organs during recovery from nerve injury. Neurosci. Lett. 207:179, 1996. 65. Scott, J. J. A., and Young, H.: The number and distribution of muscle spindles and tendon organs in the peroneal muscles of the cat. J. Anat. 151:143, 1987. 66. Sinkjaer, T., Andersen, J. B., Ladouceur, M., et al.: Major role for sensory feedback in soleus EMG activity in the stance phase of walking in man. J. Physiol. (Lond.) 523:817, 2000. 67. Spielmann, J. M., and Stauffer, E. K.: Morphological observations of motor units connected in-series to Golgi tendon organs. J. Neurophysiol. 55:147, 1986. 68. Stacey, M. J.: Free endings in skeletal muscle of the cat. J. Anat. 105:231, 1969. 69. Stuart, D. G., Mosher, C. C., Gerlach, R. L., and Reinking, R. M.: Mechanical arrangement and transducing properties of Golgi tendon organs. Exp. Brain Res. 14:274, 1972. 70. Wessberg, J., and Vallbo, A. B.: Coding of pulsatile motor output by human muscle afferents during slow finger movements. J. Physiol. (Lond.) 485:271, 1995. 71. Wilkinson, R. S., and Fukami, Y.: Responses of isolated Golgi tendon organs of cat to sinusoidal stretch. J. Neurophysiol. 49:976, 1983. 72. Zelena, J., and Soukup, T.: The in-series and in-parallel components in rat hindlimb tendon organs. Neuroscience 9:899, 1983. 73. Zytnicki, D., Lafleur, J., Kouchtir, N., and Perrier, J.-F.: Heterogeneity of contraction-induced effects in neurons of the cat dorsal spinocerebellar tract. J. Physiol. (Lond.) 487:761, 1995.

8 The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons SALLY N. LAWSON

Overview Large Light (Neurofilament-Rich) and Small Dark (Neurofilament-Poor) DRG Neurons Soma Size, Conduction Velocity, and Neurofilament Soma Size in Relation to Sensory Properties Afferent Receptive Properties Terms Used to Define Properties of Afferents Functional Classes of Cutaneous Afferent Neurons Cutaneous, Muscle, and Visceral Afferent Receptive Properties Electrical Membrane Properties of DRG Neurons Nociceptors versus LTMs

Non-nociceptive Neurons Changes in Electrophysiology in Chronic Pain Models Chemical Phenotype of DRG Neurons Neurotrophic Factors and Their Receptors Inflammatory Mediators and Tissue Damage Other Receptors Oligosaccharides Expressed by DRG Neurons Transmitters, Neuromodulators, Peptides, and Their Receptors Excitatory Amino Acid Transmitters Neuropeptides Cytochemistry of Cutaneous, Visceral, and Muscle Afferents

OVERVIEW This chapter attempts to relate the morphology, electrophysiology, and cytochemistry of dorsal root ganglion (DRG) neurons to each other and to their sensory receptive properties. The classification of primary afferent neurons in relation to their afferent receptive properties was covered in depth in the previous edition182 and is merely summarized here. Primary afferent neurons are pseudo-unipolar. That is, in their mature form they have only one process leaving the soma. This process is called the initial segment. It branches at its T junction into a peripherally and a centrally projecting process. The initial segment is short in neurons with unmyelinated fibers, and much longer in neurons with large myelinated fibers. The sensory neuron or unit includes the cell body, or soma, plus fibers and terminals. The cell bodies (somata) of primary somatosensory neurons are segmentally arranged in DRGs and in cranial root ganglia. Somatosensory neurons in DRGs may be nocicep-

Membrane Receptors and Sensory Transduction Thermoreceptive Molecules Mechanoreceptive Molecules Chemoreceptive Molecules Acid-Sensing Molecules Ion Currents and Channels Na Currents and Channels K Currents and Channels Hyperpolarization-Activated Currents and Channels Ca2 Currents and Channels Chemical Phenotype of Nociceptive and LTM Neurons Summary

tive or non-nociceptive; may have unmyelinated (C) fibers or myelinated (A) fibers; may respond to mechanical, thermal, and/or chemical stimuli; and may project to skin, muscle, and blood vessels of the trunk and limbs or to visceral organs in the thorax and abdomen. Glial cells in DRGs and cranial root ganglia are of neural crest origin, as are all DRG and jugular ganglion neurons and some neurons in the trigeminal ganglion (Vth cranial nerve) neurons. The rest of the trigeminal ganglion neurons and all neurons in the geniculate (VII), petrosal (IX), and nodose (X) ganglia are of placodal origin.174 This chapter focuses on DRG neurons.

Large Light (Neurofilament-Rich) and Small Dark (Neurofilament-Poor) DRG Neurons Studies of the morphology and histology of DRG neurons from the late 19th century mostly found two main subgroups of neurons (see Lawson164 and Rambourg et al.254), known as LL (large light) or A-type neurons and SD (small dark) or B-type neurons. LL and SD neurons have overlapping 163

164

Function of the Peripheral Nervous System

distributions of cell size (cross-sectional area through the middle of the cell); in mouse and rat the sizes of each are normally distributed,164,171 leading to bimodal distributions of cell size. LL neurons have a large mean size but include neurons of all sizes; SD neurons are small. Bimodal distribu-

A

B

C

tions of DRG neuronal size occur in many species, including mouse, rat, cat,243,248 and human (S. Border and S. N. Lawson, unpublished data) (Fig. 8–1) as well as chick. These populations are not evenly distributed within the length of the DRG. In chick DRGs the larger neurons tend to be more

D

E

F

FIGURE 8–1 Size distributions of dorsal root ganglion (DRG) neuronal profiles that contain nuclei in frozen sections (A–D) and wax sections (E and F). Sections are stained to show neurofilament (NF) immunoreactivity (filled grey histograms show NF-rich neurons and hatched histograms show NF-poor neurons) in all but E, which shows horseradish peroxidase (HRP)–injected cells with no immunoreactivity. Different antibodies against NF were used. Antibodies were RT97 against a highly phosphorylated epitope on the 200-kDa NF subunit (NF200) (A, D, and F), against a phosphorylation-independent epitope (B), and against the 68-kDa subunit (C). A small-diameter population of NF-poor neurons and an NF-rich population with a much larger mean diameter are apparent in all species, but cell sizes depend on species. A, Adult rat DRG, plotted on same scale as cat ganglia (B) and as human (C) ganglia from a human postmortem L5 DRG of a 70-year-old man. D, Same data as in A, plotted for comparison with E and F and showing size ranges (small, medium, and large), which differ slightly as a result of differences in processing in D, E, and F. E and F, Sizes of dye-injected neurons with C-, A-, or A-fiber neurons. After intracellular recording, dye was injected into the soma, enabling cell size to be plotted in relation to conduction velocity. E, Cells injected with HRP, with no immunocytochemistry. F, Fluorescent dye injection followed by immunocytochemistry for NF200. NF-rich neurons have A fibers and NF-poor neurons have C fibers. (B: Data from Perry, M. J., and Lawson, S. N.: Neurofilament in feline primary afferent neurons: a quantitative immunocytochemical study. Brain Res. 607:307, 1993. C: Courtesy of Sandra Border. E: Data replotted from Harper, A. A., and Lawson, S. N.: Electrical properties of rat dorsal root ganglion neurons with different peripheral conduction velocities. J. Physiol. [Lond.] 359:47, 1985. F: Data replotted from Lawson, S. N., and Waddell, P. J.: Soma neurofilament immunoreactivity is related to cell size and fibre conduction velocity in rat primary sensory neurons. J. Physiol. [Lond.] 435:41, 1991.)

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons

peripheral (ventrolateral), and the later developing smaller neurons are mesiodorsal.118 Similarly, in rodent (unpublished data) and cat308 DRGs, the earlier developing167,168 LL neurons tend to be concentrated more peripherally. LL DRG neurons have clear Nissl substance (patches of organelles that pick up conventional stains) separated by unstained (with conventional stains) swathes of neurofilament (NF), whereas SD neurons have little NF and more even, darker staining.80,168,281,344 Thus LL neuronal somata are NF rich and SD somata are NF poor.243,245 Immunostaining for NF (e.g., Fig. 8–1A to D; see also Fig. 8–4 later) can distinguish between these populations in several species (e.g., rat, mouse, cat, human) as originally demonstrated in rat.171

Soma Size, Conduction Velocity, and Neurofilament Neuronal cell body size and conduction velocity are loosely related. In adult mice, rats, and cats, DRG neuronal somata with C fibers are small, those with A fibers are small to medium sized, and those with A-fiber neurons mainly medium or large121,175,347 (Fig. 8–1E and F) if these words are defined as follows: small  within size range of SD cell distribution, large  right-hand side of LL cell distribution, and medium  between small and large ranges. With this definition, small neurons have mainly C and fewer A fibers, medium-sized neurons may have C, mainly A, or fewer A fibers, and large neurons A (few) and mainly A fibers. In rat, the NF-poor/SD and NF-rich/LL lumbar DRG neurons have C fibers and A fibers, respectively (see Fig. 8–1F),173 and NF content of the soma is thus frequently used to distinguish between these groups. For deductions to be made about conduction velocity from cell size alone, the size ranges of small, medium, and large neurons need to be established for the neurons under study, because actual sizes vary with species (Fig. 8–1A to C) (also see Gerke and Plenderleith100), postnatal age,164,168 histologic processing, and whether studied in sections, or in vitro, where some neurons may flatten. An additional complication is that cell size (both LL and SD) reduces by about 30% after spinal nerve axotomy, coupled with a loss of DRG neurons. In one study in rat the neuronal loss was 20% after 1 week increasing to 35% by 45 days327; in mouse DRGs 24% and 54% neuronal loss, respectively, was reported 7 and 28 days after midsciatic section.283 Furthermore, the loss after nerve transection appeared to be limited to cutaneous, not muscle, afferents.128 Thus comparisons between axotomized and nonaxotomized neurons should take into account possible changes in cell size; selective cell loss; and altered small, medium, and large cell size boundaries.

Soma Size in Relation to Sensory Properties It is frequently stated that small neurons in DRGs are nociceptive. It is the case that a higher proportion of small neurons than of large neurons is nociceptive. However, some

165

A-fiber nociceptive neurons are large, especially those with A fibers (see later), and some small neurons are C- or A-fiber low-threshold mechanoreceptors (LTMs) (unpublished observations). Thus, although many small neurons are likely to be nociceptive, and many large neurons are likely to be LTMs, there are important exceptions to this.

AFFERENT RECEPTIVE PROPERTIES Although DRG neurons are all afferent neurons, only those that give rise to a conscious sensation are, strictly speaking, sensory. Thus visceral afferent neurons that may not evoke conscious sensation are usually called afferent, not sensory. Primary afferent neurons have classically been described in terms of their modality or adequate stimulus, threshold, conduction velocity, adaptation properties, and sensory receptive properties. These terms are expanded below.

Terms Used to Define Properties of Afferents Modality. The term sensory modality of a primary afferent neuron is used to describe the type of sensation asociated with stimulation of that neuron. Classically, the modalities of human sensation included sight, hearing, taste, smell, and touch. For the somatosensory system as well as touch, it is useful to include temperature (heat, cold), and pain. More precise definition can be achieved by referring to the quality of the evoked sensation, such as “gentle pressure” to further define touch, or “burning pain.” Adequate Stimulus. Adequate stimulus has been defined as the “type of naturally occurring event that most effectively excites” the afferent unit.185 Thus a low-intensity mechanical stimulus is the adequate (most effective) stimulus for a LTM fiber. Threshold. The threshold stimulus is the weakest that can activate the fiber. Low-threshold units respond to lowintensity (usually nondamaging and nonpainful) stimuli. Thus cutaneous LTMs usually give rise to sensations such as touch, vibration or pressure, position, or movement, whereas low-threshold thermoreceptors give rise to sensations of warmth or cooling of the skin. High-intensity stimuli are likely to cause damage to the tissues. Nociceptors and Pain. Stimuli that are damaging or potentially damaging to the tissues are said to be noxious, and primary afferent neurons that respond only, or preferentially, to such stimuli are called nociceptors. To quote Light and Perl, “Nociceptors are defined as primary afferent units that uniquely signal stimuli intense enough to cause damage to the tissue.”185 Nociceptors usually have high thresholds for activation, but these thresholds may be lower where tissues are particularly vulnerable, such as the surface of the cornea. Activation of nociceptors usually

166

Function of the Peripheral Nervous System

results in pain once the information they transmit reaches the higher centers of the brain in a conscious animal. Some fibers may be selectively activated by noxious mechanical, noxious thermal, or noxious chemical stimuli, others by more than one of these types of stimulus. The last group are called polymodal nociceptors.185 In contrast to cutaneous afferents, in which there are clearly separate groups of nociceptive and LTM neurons,185 there is debate about whether afferents projecting to some visceral organs (heart, colon, bladder) may be involved in “organ regulation and non-painful sensations as well as pain,”133 whereas innervation of other organs clearly involves separate lowthreshold and nociceptive afferents (see later). Transduction. Stimuli are transduced into electrical signals at receptive regions of the fiber. The type(s) of natural stimulus (mechanical, thermal, or chemical) that can evoke electrical signals in a given neuron depend on the type(s) of protein receptor molecules (an ion channel or associated with an ion channel) in its receptive membrane. For instance, activation of a Na channel or closure of a K channel may cause a depolarizing receptor potential; if this reaches threshold, one or more action potentials will propagate along the fiber. Transduction usually occurs at the receptive membrane at the peripheral terminals of the neuronal fiber but may occur in receptive non-neuronal cells, such as the Merkel cell, whose possible role in mechanical transduction has been the subject of much debate.117 Possible molecular transduction mechanisms are discussed in a later section. Conduction Velocity. Afferents are divided according to their fiber conduction velocity into C, A, and A fibers (cutaneous afferents) and group IV, III, II, and I fibers (muscle afferents).185 The conduction velocity ranges of these groups of fibers can be determined with compound action potential recordings,185 but need to be determined for each experimental situation. This is because the velocity ranges differ with age, species, temperature, the nerve used, and the proximodistal position along the nerve since some fibers slow toward the periphery.122 For example, the upper limits of 2.5 m/s for C fibers and 30 m/s for A fibers, appropriate for adult cat peripheral nerve,185 are sometimes applied to rat studies, although values determined with compound action potentials for young adult rats are much lower (1 m/s for C fibers and 1 to 6 or 1 to 8 m/s for A fibers).73,85 Adaptation. A stimulus of constant strength applied to the sensory receptive field results in firing of action potentials at a rate that decreases with time. This decrease is known as adaptation, and may be rapid or slow. The terms rapidly adapting (RA) and slowly adapting (SA) were first coined in relation to mechanosensitive primary afferent neurons. RA mechanoreceptors tend to fire only during the initial application or removal of a constant mechanical stimulus. That is, they have a phasic or dynamic pattern of firing and

detect change or movement, and can signal such information to the central nervous system (CNS). The most rapidly adapting mechanoreceptors are those innervating pacinian corpuscles, which signal information about acceleration or rapid vibration.185 Conversely, SA mechanoreceptors continue to fire (with a static or phasic firing pattern) in response to a sustained mechanical stimulus such as stretch or light pressure, and signal information about the amount of pressure or degree of stretch to the CNS. There is some evidence that intrinsic membrane properties contribute to adaptation and that these may be similar at the receptive membrane and in the soma.120

Functional Classes of Cutaneous Afferent Receptive Neurons Non-nociceptive neurons include low-threshold thermoreceptor (cool and warm receptor) neurons, low-threshold chemoreceptive afferents such as taste receptors, and a variety of visceral afferent chemoreceptors (Table 8–1). However, the largest group of non-nociceptive afferent DRG neurons projecting to skin and skeletal muscle are LTMs. Cutaneous LTMs conduct in A-, A-, and C-fiber conduction velocity ranges, and muscle spindle afferent LTMs conduct in group II and group I ranges. Fibers of cutaneous and muscle LTMs241 make up a large proportion of fast-conducting myelinated fibers (A/types I and II). The different types of unit are summarized in Table 8–1, and their properties are covered in greater depth by Light and Perl.185 An overview follows. Low-Threshold Mechanoreceptors (LTMs) Cutaneous A␣␤-Fiber LTM Units. Cutaneous A-fiber LTM units show a variety of adaptation rates and firing patterns. SA units are said to be of two types. The SA type I (SA I) have an irregular discharge and are thought to make contact with Merkel cells in skin (see review by Halata et al.117). The SA II fire more regularly127 and are classically described as contacting Ruffini endings in dermis, although there are few such endings in human glabrous skin.234 RA units in skin are important in the detection of stimuli of changing intensity and movement of mechanical stimuli. They include fibers (glabrous RA units) in glabrous skin that are thought to innervate Meissner’s corpuscles and fibers in hairy skin that innervate guard hair (G hair) follicles (known as G hair units). Afferent units that are best activated by light mechanical stimuli to the skin between hairs or to movement of groups of hairs are known as Field units. G hair and Field units are subdivided according to their sensitivity to rate of movement such that the more slowly adapting G hair and Field units are called type II and the more rapidly adapting units activated only by faster rates of movement are called type I.127 Thus GII hair units respond to sustained deflection of a hair, whereas GI hair units respond best to sudden deflection, or flicking, of a hair. This leads to the following

Phasic/ Slower than A/

Tonic/Slow

Slow movement

Pressure, stretch, position

Phasic/ Rapid and variable

A/ A/

SAI (Merkels cells); irregular firing SAII (? Ruffini-end organs), regular firing, responds to stretch and cooling

C

C mechanoreceptor (C LTM), very slow (not fast) movement over skin, cooling

A/

A/

A

Glabrous RA (Meissners corpuscle)/ movement G Hair follicle/ field unit, variable adaptation to hair deflection or skin indentation

Movement

A/

CV Groups

D hair/extremely sensitive to slow hair movement, also stretch, cooling

Pacinian corpuscles in dermis

Vibration, Phasic/ Very acceleration, rapid tap, jerk

Phasic/ Rapid

Fiber Type/ Responds to

Type of Stimulus

Firing/ Adaptation Rate

Skin (Exteroceptors)

Table 8–1. Classes of Cutaneous Afferent Receptive Neurons*

?

Sustained pressure

? emotional touch

? insect detection

Movement

Movement

Tap, buzz, vibration

Sensation

Pressure on muscle, Contraction sensitive units

Muscle spindle primary and secondary endings, Golgi tendon organs. Ruffini corpuscles in joint capsules & ligaments

?

Primary endings respond with dynamic firing component to rate of change of stretch (muscle spindle intrafusal nuclear bag fibers)

Paciniform endings in ligaments, joint capsules and facia

Fiber Type/ Responds to

III, IV

Ia, Ib, II

Ia

A/

CV Group

Muscle/Tendon/Joint (Proprioceptors)

A, A/ or unknown

A

A

A/

CV Group

Table continued on following page

Organ capsules, walls of GI tract & lung, baroreceptors, renal pelvis

Lung SA stretch receptors

Paciniform corpuscles in mesentery (in cat) & large arteries Lung RA stretch receptors

Fiber Type/ Responds to

Viscera (Enteroceptors)

?

Firing/ Adaptation Rate

Probably Slow

?

Unknown

Polymodal nociceptor

Fibers responsive to heat or cold

C/A

Mostly C also A

C & A

A A/ C

C/A

CV Groups

?

burning pain/ache

Heat/cold

Sharp pricking or dull pain

Warm, cool

Sensation

Unknown

Polymodal nociceptor

?

Joint capsules, ligaments skeletal muscles: compression or stretch.

Thermoreceptor (cool or warm) in muscle Free nerve endings

Fiber Type/ Responds to

IV

III & IV

III & IV

III/IV

IV

CV Group

Muscle/Tendon/Joint (Proprioceptors)

Unknown

Polymodal nociceptor

?

Bladder pressure 60 cm H2O

Free nerve endings

?

e.g., Low pH, glucose, various chemoreceptors, ion concentrations, ischaemia, irritant receptors in lung

Fiber Type/ Responds to

C

C A

C & A

C & A

Vary

CV Group

Viscera (Enteroceptors)

CV  conduction velocity; GI  gastrointestinal tract; HTMs  high-threshold mechanoreceptors; LTMs  low-threshold mechanoreceptors; RA  rapidly adapting; SA  slowly adapting. *See text and references 34, 52, 116, 117, 133, 176, 185, 205, 242, 260, and 273. This table is not intended to be a complete listing but to provide an overview of some of the types of units innervating different types of tissue. The most complete section relates to skin.

Silent nociceptors Normally insensitive become sensitized by, (e.g., inflammatory mediators, heat, skin damage)

Polymodal nociceptors One or more of: noxious Depends on mechanical, noxious stimulus heat, cold or chemical type (e.g., pH, ischemia, inflammatory mediators)

Noxious thermal heat or cold Heat 44˚ C or cold 17˚ C

HTMs mainly A, some A in dermis and in epidermis.

Noxious mechanical

Some are slow

Free nerve endings

Cutaneous, presumed free nerve endings

Taste and olfaction are excluded as DRG neurons are not involved

Fiber Type/ Responds to

NOCICEPTORS

THERMO-RECEPTORS Warm or cool Variable

CHEMORECEPTORS Internal chemical environment

Type of Stimulus

Skin (Exteroceptors)

Table 8–1. Classes of Cutaneous Afferent Receptive Neurons*—Continued

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons

groupings: GI hair, GII hair, Field I, and Field II.127 Thus, although Field and G hair units are commonly thought of as rapidly adapting, in reality their rates of adaptation vary. The most rapidly adapting units are those innervating pacinian corpuscles in the dermis112,235 and the mesentery of some species such as cat.224 These units respond best to rapidly changing stimuli such as a tap to the skin, or rapid vibration. Cutaneous A-Fiber LTMs. Most cutaneous A-fiber LTMs projecting to hairy skin are called “D hair” units after the very fine down hairs. D hair units are extremely sensitive to slow movement of hair, but also respond both to stretch and frequently to cooling of the skin.185 Their sensitivity to small movements of fine hairs makes detection of insects in the fur one possible role. Cutaneous C-Fiber LTMs. Cutaneous C-fiber LTMs (C mechanoreceptors) respond to extremely slow (1 mm/s) movement across the skin.185 They have been described in a number of species (rat, cat, rabbit, guinea pig). Recent evidence in the human shows the presence of these units323 and indicates that information from C mechanoreceptors reaches the insula cortex and is associated with a pleasant emotional response (limbic touch).231 Thermoreceptors Cooling and warming receptors have C-fiber conduction velocities in subprimates but may have C or A fibers in primates.185 Cooling receptors fire at higher frequencies as temperature is decreased from about 30° C, and human psychophysical experiments show that sensations of cold/cool result from temperatures below 30° C, ache below 17.5° C, and cold pain below 14° C.65,185 Peripheral warm receptors fire in the range from 30° to 46° C.65,185 Many C-fiber polymodal nociceptors have a thermal threshold of about 43° C, and some A-mechanoheat units have a higher threshold (53° C). Nociceptors C-Fiber Nociceptors. C-fiber nociceptive neurons are the most numerous of the nociceptive fibers and probably have receptive fields in all types of tissues. The majority of those with superficial receptive fields respond to both noxious mechanical and noxious heat stimuli and can thus be classed as C-mechanoheat or C-polymodal nociceptors.185 A smaller proportion of units with superficial receptive fields are C-fiber high-threshold mechanoreceptor (HTM) units, because they respond to noxious mechanical stimuli but not to a single application of heat or cold; a few respond to noxious heat only. A-Fiber Nociceptors. Cutaneous A-fiber nociceptors have A or A fibers and respond to noxious mechanical stimuli or noxious mechanical and noxious heat stimuli (mechanoheat units). Superficial receptive fields are punctate. For a discussion of firing properties and sensitization of these

169

neurons, see Light and Perl.185 A-fiber nociceptors in primate (monkey) have different responses to noxious heat. Type I units (two-thirds A, one-third A, conducting up to 60 m/s) have a higher heat threshold with a later onset, long-lasting response to heat and a lower mechanical threshold, whereas type II units (exclusively A, 30 m/s) have a lower heat threshold with a prompt, early, and short response to heat and a higher mechanical threshold.322 First pain sensation to heat may be served by the type II A fibers, and first pain to noxious mechanical stimuli by type I (A and A fibers).322 Although sometimes overlooked in recent literature, some A-fiber nociceptors conduct in the A range. For a review of the evidence for A nociceptors, see Djouhri et al.73 Briefly, conduction velocities of A-fiber nociceptors range from 5 to 60 m/s in cat33,35 and 5 to 60 m/s in primate (monkey),322 clearly including units with A fibers because in both species the upper end of the A range is less than 30 m/s. The earliest descriptions of properties of cutaneous nociceptive A fibers in cat33,35 and monkey240 included nociceptors with A fibers. Some of these A units were insensitive mechanoreceptors or moderate-pressure receptors, with lower thresholds than many A nociceptors, but because both types fired more enthusiastically to noxious than to nonnoxious pressure, their adequate stimulus is in the noxious range and they are classed as nociceptors. Burgess and Perl33 found in cat that the percentages of nociceptive A fibers that conducted above the A range (30 m/s) were 7% of highthreshold nociceptors, 35% of insensitive mechanoreceptors, and 100% of moderate-pressure receptors. Intracellular recordings in DRG neurons show that about 20% to 65% of A-fiber nociceptive neurons have A fibers in cat,157 rat263 (X. Fang et al., unpublished data), and guinea pig.71 For more details on these percentages, see Djouhri et al.73 Nonetheless, at the time of writing, it is frequently assumed that nociceptors conduct only in the C- and A-fiber range, perhaps because (1) the distribution of conduction velocities of A-fiber nociceptors tends to peak in the upper part of the A range33,35 and (2) the A wave of the compound action potential is dominated by LTM fibers. (For further discussion, see Lawson166 and Djouhri et al.73) Unresponsive DRG Neurons: Possible “Silent” Nociceptors. Inexcitable afferent units are hard to identify with certainty because (1) the appropriate search stimulus may not have been used and (2) in fiber recordings sympathetic C fibers have to be excluded. Nonetheless, apparently inexcitable units have been described in several species projecting not only to joint tissues but also to skin and visceral organs.71,133,273 These units may respond after repeated stimuli or after acute inflammation of the tissues206 to stimuli that did not previously excite them. Such units may include very-high-threshold nociceptive neurons and/or units that are either sensitized or excited by substances released during tissue damage, including inflammatory mediators (see later). The latter may act as detectors of inflammation/tissue damage.133,185

170

Function of the Peripheral Nervous System

ELECTRICAL MEMBRANE PROPERTIES OF DRG NEURONS

Cutaneous, Muscle, and Visceral Afferent Receptive Properties Table 8–1 provides a summary of some of the different types of primary afferent neurons, linking stimulus type, conduction velocity range, adaptation properties, and associated sensations where known. Information about sensations is largely dependent on microneurography in the conscious human. Microneurography employs electrode penetration of a nerve to record from, and stimulate, single afferent fibers, enabling correlation between fiber properties and reported sensations. Unlike cutaneous and muscle afferents, visceral afferents do not all fall into clear nociceptive and non-nociceptive categories. For example, firing of afferent fibers encodes distention pressures from nonpainful through to painful in both bladder and colon, whereas other organs, such as gallbladder and ureter, are innervated only by high-threshold afferents that are mechano- and chemosensitive and are clearly visceral nociceptors; still others (esophagus) are innervated by both high- and low-threshold afferents.133 Only a few examples of visceral afferents are listed in the table. For a detailed list and extensive bibliography of the earlier literature, see the tables in Perl and Burgess.242

A

C HTM

D

Nociceptors versus LTMs There are clear differences in soma membrane properties between nociceptive and LTM neurons. Nociceptive neurons have longer action potential and longer afterhyperpolarization durations in cat, rat, and guinea pig A-fiber neurons71,156,263 (Fig. 8–2), and in rat and guinea pig in C-fiber neurons71 (X. Fang et al., unpublished data). To summarize, in each of the conduction velocity ranges (C, A, and A), nociceptive neurons have significantly longer afterhyperpolarization durations,71,156,263 and larger action

Aδ HTM

B

B

C LTM

E

ns

1

B

I Aα/β LTM G H

H Aα/β LTM MS J

Aα/β LTM T HAIR

ns

ns

G Aα/β HTM

* *

10

C Unresp 20 mV 2 ms

***

ns

Aδ Unresp

*

ns

F C

Aδ LTM

AP durn at base (ms)

A

AHP 80% duration (ms)

A

Although it is hard to record intracellularly from small afferent fibers and terminals, it is possible to make such recordings from the soma. Because the properties of DRG neurons in vitro may be altered by axotomy and removal of the normal cellular and molecular environments, this section is limited (unless stated otherwise) to intracellular recordings from DRG neurons in vivo.

100

10

NOC LTM NOC LTM NOC LTM Aδ Aα/β C ns ns ns ns ns ns

*

* *

* * *

1 NOC LTM NOC LTM NOC LTM Aδ Aα/β C

FIGURE 8–2 A, Examples of action potential configurations from intracellular recordings of guinea pig DRG neurons, with identified sensory properties and conduction velocities (indicated for each trace). HTM  high threshold mechanoreceptor; LTM  low threshold mechanoreceptor; Unresp  unresponsive; G  G hair receptor; MS  muscle spindle afferent; T hair  Tylotrich hair. B, The relationship of the median action potential (AP) duration and afterhyperpolarization (AHP) duration to 80% recovery. NOC  nociceptors; ns  median not significantly different; *  P  0.05; ***  P  .001 Mann Whitney U Test. (Data from Lawson, S. N.: Phenotype and function of somatic primary afferent nociceptive neurons with C-, Adelta or Aalpha/beta-fibres. J. Exp. Physiol. 87:239, 2002.)

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons

potential overshoots,75 and in the C- and A-fiber ranges they have significantly longer action potential durations with both longer rise times and fall times71,156,263 (for a summary, see Lawson166 and Fig. 8–2). As well as being related to sensory properties, action potential duration is also inversely related to the conduction velocity of the fiber (see Fig. 8–2B).106,122,166,330 Extracellular recordings of afferent C fibers in rat and pig97 also show a trend of longer action potential durations in nociceptive compared with LTM fibers. These slower membrane kinetics of nociceptive neurons in both soma and fiber, plus their higher activation thresholds, may have an important role in normally limiting the amount of information reaching the CNS via the nociceptive primary afferent pathway.

Non-nociceptive Neurons There are differences in intracellularly recorded membrane properties between different subtypes of LTM neurons. In guinea pig71 and rat (X. Fang et al., unpublished data) DRGs, the action potential duration of muscle spindle Afiber afferents have faster kinetics (shorter durations) than those of cutaneous A-fiber LTM units (see Fig. 8–2A).

Changes in Electrophysiology in Chronic Pain Models The two main types of animal model of chronic pain states are the induction of inflammation in the tissues and damage to the peripheral nerve. During Chronic Peripheral Inflammation Inflammatory mediators cause rapid (minutes to hours) changes in sensitivity of peripheral terminals in inflamed tissues. However, clinically more important is chronic pain that may last for days, weeks, and months. Animal models of chronic inflammatory pain include intradermal injections of complete Freund’s adjuvant (CFA) or carrageenan in one hind limb, which induce both inflammation and enhanced withdrawal from noxious stimuli, both immediately and persisting for days/weeks. Changes in electrophysiologic properties of nociceptive but not LTM DRG neurons in vivo in guinea pig take more than 2 days to be complete after intradermal CFA injection.72 They include increased incidence (threefold) of C- and A-fiber nociceptors that fire spontaneously.72 These changes are accompanied by decreased withdrawal latencies to noxious heat. The slow time course of the changes is compatible with altered protein expression of ion channels or associated molecules. Indeed, Na channel protein overall is upregulated by the first day after CFA injection, with a longer lasting increase in small than large DRG neurons.108 Many inflammation-induced electrophysiologic changes in guinea pig including the increased spontaneous firing, are nerve growth factor (NGF) dependent,72 as is the inflammatory hyperalgesia.338

171

Following Peripheral Nerve Injury In humans, peripheral nerve injury leads to neuropathic pain involving hyperalgesia, tactile and cold allodynia, and spontaneous pain. A number of animal models of neuropathic pain involve peripheral nerve injury. These include chronic constriction injury (CCI)20 and tight ligation or cut of the L5 (or L5 and L6) spinal nerves (Chung model).153 It is not known which neuronal subtypes make the greatest contribution to neuropathic pain: whether it is those with C, A, or A fibers; those originally with nociceptive or LTM properties; those with intact or severed peripheral fibers; those with fibers affected by inflammation at, or distal to, the injury site; and/or those that are spontaneously active or not. Some evidence in human patients indicates a correlation between spontaneous fiber activity and spontaneous pain.110,225,346 In animal models, peripheral nerve injury (axotomy or CCI) leads to a brief burst of firing, and after 2 to 3 days a gradual increase in spontaneous activity.109,188,197 The activity may originate at the lesion site or in the DRG.137,302 Many afferents develop “an ectopic repetitive firing capability” in response to stimuli in the region of the injury,197,331 which may be blocked by Na channel blockers (tetrodotoxin [TTX], lidocaine) and enhanced by veratridine (prolongs open time of Na channels), implicating Na currents.197 One study found no spontaneous activity in damaged cutaneous afferents.209 Axotomy leads in vivo to broader action potentials in A-fiber DRG neurons than usual, coupled with increased spontaneous firing in A- but not C-fiber neurons.154,295 Intact neurons with C- (but not A-) fibers that run alongside axotomized fibers also show spontaneous firing which may prove to be important in generation of spontaneous pain.340 It may therefore be that a combination of spontaneous activity from axotomized A-fiber and adjacent intact C-fiber neurons both contribute to the generation of neuropathic pain.

CHEMICAL PHENOTYPE OF DRG NEURONS There is a vast literature covering the expression of peptides, proteins, enzymes, receptors, signaling proteins, and carbohydrate groups in DRG neurons (for previous reviews, see Lawson165 and Millan211). Some aspects of the chemical phenotype are expressed transiently during development, and/ or show altered expression following damage to the nerves or other tissues, including inflammation. The focus here is on molecules that are selectively localized in clear subgroups (e.g., certain sizes) of DRG neurons, because this may be indicative of a link with selective functions of a neuronal type. (For a detailed review of co-localization as known in 1992, see Lawson.165) Percentages of DRG neurons reported to be positive for a particular marker are dependent on differences in in situ probes, antibodies, immunocytochemical techniques, expression at different ages, rostro-caudal level of DRGs, species or strains within

SMALL

MEDIUM

LARGE

C FIBER / NF POOR Aδ-FIBER NF-RICH Aδ /β-FIBER NF-RICH NOCICEPTORS LOW THRESHOLD MECHANORECEPTORS Peptide Peptide Peptide Peptide REC REC REC REC REC ION CH ION CH ION CH ION CH ION CH / REC Cytokine Cytokine Enzyme CBH / LECTIN Peptide Peptide REC REC ION CH / REC Peptide REC REC REC REC ION CH ION CH / REC ENZYME REC ION CH ION CH ION CH / REC REC REC REC REC REC REC ENZ ENZ ION CH ION CH ION CH ION CH Cytokine CBH CBH ION CH / REC REC ION CH

ET1 Galanin Nociceptin Somatostatin GLUR1,GLUR5,AMPA,Kainate B1, B2, H1 EGF, FGF2 NK1, SSTR2a P2X3 BK, KCa current Ca++ current L-type, N-type Kv1.4 Nav1.9, Naβ3 TRPA1,TPRM8,TRPV1,TRPV3 TGFα, TNFα LIF retrograde transport COX1 IB4 binding / PNA ET1, PACAP, VIP Substance P / NKA GAL2, ETA receptor GFRα3 TRPV4 CGRP GFRα1,GFRα2, RET MOR/DOR/KOR ORL1 TrkA, p75 Nav1.7, Nav1.8 P2X2/3 nNOS TrkB Ca++ current T-Type Kir TRPV2 CB1 cannabinoid receptor GAL1 receptor GLUR2/3 GM1 receptor / Cholera toxin B binding P2Y1 TrkC / NT3 uptake Calbindin Carbonic anhydrase α3 Na+ pump subunit Naβ1.1,Naβ2.1 Kv1.1, 1.2, Kvβ2.1 H current, HCN1, HCN2 IL1β GSAII / LCA / PSA lectin binding SSEA3/4 BNC1/ASIC2 GluR2/3, gp130, CNTF-R, IL6-R, KCNQ2, KCNQ3, KCNQ5

FIGURE 8–3 A visual indication of the approximate reported sizes of neurons labeled by different markers or expressing different currents (details in text). The length of line relates to neuronal relative sizes (small, medium, and large) as used in Figure 8–1, from published size distributions or from terminology used by authors. Thus, for this figure, the word “small” is translated into a line the width of the SD/C-fiber/NF-poor population. This figure does not provide any indication about proportions of neurons expressing these phenotypes. The order used relates to (a) cell size, (b) the type of molecule (see column 1), and (c) within these groups is alphabetical. Abbreviations are found within the text, but those in column 1 are as follows: REC  receptor; ION CH  ion channel; ION CH/REC  ion channel gated chemically, thermally, or mechanically; CBH/lectin  lectin binding to carbohydrate groups or use of antibodies against particular oligosaccharide groups (see text).

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons

species, and, importantly, the level of staining required for a neuron to be judged positive, which is rarely explained. These factors together largely explain differences in reported percentages. Although much extremely useful knowledge has been acquired from DRG neurons in tissue culture, cultured neurons may show altered phenotypes compared with their in vivo counterparts as a result of physical damage such as axotomy close to the soma145,274 and loss of in vivo influences such as trophic factors. Percentage values for human DRGs labeled by immunocytochemistry or in situ hybridization may be influenced by long postmortem times, resulting in degradation of molecules under examination. Because of the complexity and amount of information, a number of figures and tables summarize the detail provided in the text, as follows. An indication of the sizes of neurons reported to have different chemical phenotypes is provided in Figure 8–3, and examples of staining for NF, isolectin B4 (IB4), calcitonin gene–related peptide (CGRP), and tyrosine kinase A (TrkA) are shown in Figure 8–4. A cartoon of sources of endogenous substances that may influence DRG neuronal function is given in Figure 8–5A and a summary of altered expression of substances by DRG neurons or non-

173

neuronal cells after axotomy and inflammation is provided in Figure 8–5B. Tables relating chemical phenotype to aspects of afferent neuron function are included in subsequent sections for trophic factor receptors (Table 8–2); properties of neurons projecting to skin, muscle, and viscera (Table 8–3); and putative molecular transduction candidates (Table 8–4).

Neurotrophic Factors and Their Receptors Neurotrophic factors are factors that support growth, repair, or survival of neurons. These may be target derived, that is, produced by the tissues to which the neuron projects, or produced by cells, such as Schwann/satellite cells within the nerve or DRG. These factors tend to bind to specific membrane receptors, and may be internalized and transported to the neuronal soma, where they can affect intracellular events including gene expression. Their direct influence is therefore most powerful in neurons that express appropriate high-affinity receptors, although they may also bind with lower affinity to other receptors or have indirect effects via influence on other cell types. Thus understanding of (1) the effects of these factors and (2) which neurons express the receptors for

FIGURE 8–4 Interference contrast images of 8 m sections of adult rat L5 DRG showing immunostaining using ABC immunocytochemistry (Vector) with a monoclonal antibody RT97 (courtesy of John Wood. Available from Chemicon) raised against the 200-kDa neurofilament subunit (NF200); and with polyclonal antibodies raised against IB4 (Vector Laboratories), CGRP (Peninsula Laboratories) and TrkA (courtesy of L. Reichardt). Note that the NF200 staining pattern appears reciprocal to that of IB4 binding in terms of neuronal size.

174

Function of the Peripheral Nervous System

RELEASE OF SUBSTANCES FROM CELLS OTHER THAN DRG NEURONS Glut SP Somatostatin Noradrenalin Opioids BDNF FGF2 5HT NT4 GDNF CNTF IL1β PGs

ET1 IL1β TGFα EGF

NGF TNFα GDNF CNTF LIF ET1

TNFα ET1

NGF NT3 NT4 GDNF Neurturin FGFs

SKIN

BDNF Glut SP CGRP Somatostatin TNFα ?Nociceptin/orphanin

NT3 GDNF FGF2 ATP Lactate

SKEL MUSC

BDNF Glutamate SP CGRP Somatostatin

RELEASE OF SUBSTANCES FROM NORMAL DRG NEURONS A FIGURE 8–5 Endogenous substance that influence DRG neurons. A, Some of the endogenous chemicals that are known to have a direct influence (usually via specific membrane receptors) or an indirect influence on DRG neurons. Structures represented are spinal cord, a DRG neuron, satellite cells, Schwann cells, leukocytes and mast cells, skin, and skeletal muscle. Top, Chemicals secreted and released by cells other than DRG neurons, which may influence the DRG neuron at the central or peripheral terminals or at the cell body in the DRG. Bottom, Substances released by DRG neurons that may influence the same or other DRG neurons again through the central or peripheral terminals or at the cell body in the DRG. (For details, see text.) See Color Plate Figure continued on opposite page

these factors is essential for a full understanding of their roles. Additional complexity is added by altered expression or transport of these factors and/or of their receptors on the DRG neurons during development and following tissue damage such as inflammation and peripheral nerve injury (see Fig. 8–5B). The resulting plasticity in the properties of DRG neurons is extensive and makes this a very complex field. Trk Receptors for Neurotrophins The neurotrophin family includes NGF, brain-derived neurotrophic factor (BDNF), and neurotrophin-3 (NT-3) and neurotrophin-4/5 (NT-4/5). High-affinity receptors for neurotrophins are members of the tyrosine kinase (Trk) family of neurotrophin receptors and include TrkA, TrkB, and TrkC; p75 is a low-affinity receptor for these neurotrophins. TrkA, TrkB, TrkC, and p75 are all expressed on subpopulations of DRG neurons, and their activation has important effects on sensory neuron function. High-affinity binding occurs between TrkA and NGF, between TrkB and both BDNF and NT-4/5, and between TrkC and NT-3. There is also some cross-reactivity between these factors and

the other receptors12; for example, NT-3 can also interact with TrkA and TrkB on DRG neurons. The selectivity of neurotrophin receptors for high-affinity binding with their ligands is enhanced by p75 (see later section on p75). Neurotrophins and their receptors are essential for the survival of sensory neurons (and other neurons outside the remit of this chapter) and in the development and maintenance of their phenotype in the adult. Studies on transgenic mice have found losses of DRG neurons of 70%, about 30%, and 55% to 80% for NGF, BDNF, and NT-3 knockouts, respectively, indicating that not only do the majority of DRG neurons depend on these neurotrophic factors for survival at some time during development, but that many require more than one.204 About 75% of DRG neurons in adult rats express at least one of the Trk receptors (TrkA, TrkB, and/or TrkC), which are expressed mainly by small-, medium-, and large-sized DRG neurons, respectively.217 Many neurons express more than one Trk and about 4% express all three.140,143 The percentages of adult rat DRG neurons that express TrkA, TrkB, and TrkC are about 40%, 5% to 10%, and 20%, respectively.147

175

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons

INFLAMMATION / NGF ↑SP ↑CGRP ↑Nociceptin ↑BDNF ↑B2 ↑GAL2 ↑MOR ↑Nav1.7 ↑Nav1.8

↓↑GAL ↓GAL1 ↓DOR ↓KOR

↑VIP ↑PACAP ↑GAL nNOS ↑EGF-R ↑GRFα3 ↑LIF-R ↑B2 ↑α2δ (Ca++) ↑Nav1.3 ↑Naβ3 ↑KCNQ2,3,5

↓BDNF ↓SP ↓CGRP ↓SOM ↓trkA ↓IB4 ↓P2X3 ↓GAL2 ↓GFRα2 ↓MOR ↓Nav1.9 ↓Nav1.8 ↓IKIR ↓Kv1.4

↑CGRP ↓GAL1

LARGE

SMALL

↑IL1 ↑TNFα

↑IL1 ↑TNFα ↑LIF ↑PGs ↑5HT

↑GLUT ↑NGF ↑NT4 ↑H+ ↑BK ↑ATP ↑Hist ↑K+ ↑ET1 ↑5HT ↑IL1

↑GLUT ↑SP ↑CGRP

SKIN ↑BDNF ↑NPY ↑EGF-R ↑GFRα1 ↑LIFR ↑α2A (adren) ↑TNFα ↑IL6 ↑Ih ↑KCNQ2,3,5

↓CGRP ↓GAL1 ↓fast IA ↓IKIR ↓T Ca2+ ↓Kv1.1, 1.2 ↓Kvβ2.1

↑NGF ↑GDNF ↑NT3 ↑TGFα ↑TNFα ↑p75 ↑TNFR1

↑NGF ↑BDNF ↑NT4 ↑GDNF ↑FGF2 ↑p75 ↑GFRα ↑FGFR1-3 ↑LIF ↑IL1β ↑IL6 ↑TNFα

↑FGF2 ↑FGFR1-3 ↑IL1α,β ↑5HT ↑TNFα,

SKEL MUSC

↑IL6 ↑LIF

↓↑CNTF ? NEURONAL SIZE ↑TNFR1,2 ↑nAChR7

PERIPHERAL AXOTOMY

Color code : PEPTIDES and GLUT (glutamate); RECEPTORS, ION CHANNELS//CURRENTS; TROPHIC FACTORS ; OTHER MEDIATORS RELEASED BY METABOLISM, TISSUE DAMAGE OR INFLAMMATION; OTHER ;

= DRG NEURON

B FIGURE 8–5 Continued Alterations in expression after axotomy or during inflammation. B, Alterations, gleaned from the literature, and detailed in the text, in expression of a variety of molecules in small and large DRG neurons and in non-neuronal cells closely associated with DRG neurons, that may influence DRG neuron function directly or indirectly. Neuronal cells shown are small and large DRG neurons. Non-neuronal cells represented are, from left to right, satellite cells, Schwann cells, leukocytes and mast cells, skin, and skeletal muscle; to the extreme right is a sensory nerve terminal. Top, Changes in expression following inflammation or application of exogenous NGF. Bottom, Changes in expression that occur in these cells following peripheral nerve axotomy. Because most nerve injury or neuropathic states have associated inflammation, it is likely that changes characteristic of both inflammation and axotomy may contribute to the resulting neuropathic pain. This representation is not intended to be complete, but nonetheless serves to illustrate the complexity of the altered environment that might contribute to altered DRG neuron function. The non-neuronal contribution is contained within the box. See Color Plate

Possible relationships between expression of trophic factor receptors and neurons with different sensory receptive properties are shown in Table 8–2, and the background to this table is given in the text below. The expression of these receptors in relation to the peripheral projection sites of the neurons is given in Table 8–3. TrkA. At birth 70% to 80% of rat lumbar DRG neurons express TrkA, and this declines to about 40% of adult rat DRG neurons that express TrkA strongly.15,215 In the adult rat about 20% of the positive neurons are NF rich and thus

are likely173 to have myelinated fibers. The appearance of TrkA-positive neurons can be seen in Figure 8–4. TrkA expression in adult DRG neurons appears to be partially dependent on target-derived NGF.183 NGF binds to TrkA in membranes of the soma and fibers of TrkA-expressing DRG neurons. Part of the signaling mechanism involves endocytosis and transport of the NGF-TrkA complex to the neuronal soma, and can thus chronically influence both the phenotype and function of TrkA-positive DRG neurons,181 whereas local intracellular signaling pathways39 enable acute local responses such as sensitization to heat (see later).

176

Function of the Peripheral Nervous System

Table 8–2. Trophic Factor Receptors and Sensory Receptive Properties Receptor

Ligand

Size (S/M/L)

Expressed by Possible Afferent Type

CV Range

TrkA

NGF  (NT-3)

SML

C  A  A

TrkB

M

TrkC

BDNF NT-4 NT-3

GFR-1 GFR-3 FGF2

GDNF Artemin FGF2

SML SM S

Mainly nociceptor or silent, some cutaneous LTMs (weak), Visceral, Merkel cell neurites*, whisker follicle afferents* LTM SA & D hair?, Ruffini endings*, whisker follicle afferents* D hair development LTM SA, Proprioceptive, Merkel cell neurites* D hair development Must include some C-fiber nociceptors (IB4) ? Mainly nociceptors Possibly muscle nociceptors

SML

A A ? C

This table summarizes information in the text relating to possible expression/dependencies of different trophic factor receptors, and the possible sensory receptive properties of neurons that express them, or show dependencies on the ligand or receptor in knockout mouse studies of ligand or receptor. Apart from TrkA, their expression has not been directly determined in individual identified neurons in vivo. *Results of studies on growth factor dependencies of large-diameter cutaneous afferent fibers projecting to the whisker complex studied using a variety of neurotrophin and neurotrophin receptor knockout mice. These studies show impairments of development of certain types of nerve fibers and cutaneous structures. The interpretation of these studies was that Ruffini endings were dependent on BDNF and TrkB; Merkel cell neurite complexes were dependent on NT-3, TrkC, NGF, TrkA, and p75; and whisker follicle afferents were dependent on NGF, TrkA, and some also on BDNF and Trk.94 Thus there appear to be complex codependencies on different factors.

Axotomy results in a 25% reduction in the percentage of rat DRG neurons that express TrkA after 1 week and 35% after 2 weeks,282 and 2 weeks after inflammation (adjuvantinduced arthritis), an increase in TrkA protein was reported,247 although other studies showed little or no change in TrkA messenger RNA (mRNA). NGF. NGF is a target-derived growth factor that binds with high affinity to TrkA on sensory neurons. Target tissue cells that secrete NGF include epithelial cells (an important source), smooth muscle cells, and fibroblasts and also, especially during development or after nerve injury, Schwann cells.10,123,341 About 70% of DRG neurons are NGF dependent during embryonic development, and some of

these (mainly the IB4-binding small neurons) switch to glial cell line–derived neurotrophic factor (GDNF) dependence in early postnatal life.216 In the adult, target-derived NGF is normally important in maintaining expression of a variety of molecules in TrkA-expressing DRG neurons, many of which are upregulated as a result of increased NGF expression in the tissues during inflammation.268 These include substance P, CGRP,326 possibly TrkA,183 (although reports on this vary), BDNF,207 growth-associated protein-43 (GAP-43),178 some channel subunits (including voltage-gated Na [Nav] channels Nav1.7,107 Nav1.8, and Nav1.986), bradykinin B2 receptors,177 and auxiliary channel proteins such as P11230 (see Fig. 8–5B). Other peptides are downregulated by NGF, and consequently upregulated after axotomy. These include

Table 8–3. Phenotypic Properties of Rat DRG Neurons That Project to Skin, Muscle, and Viscera Neurons† Marker IR/mRNA*

Cutan.

Skel. Muscle

Visc. Pelvic

Visc. Splanchnic‡

References

TrkA IR or mRNA TrkB mRNA TrkC mRNA None of TrkA, B, or C IB4 binding IR Carbonic anhydrase IR CGRP IR Substance P IR Somatostatin IR NF rich 3 Na pump IR

ⴙⴙ/ⴙⴙ   ⴙⴙ ⴙⴙ  ⴙⴙⴙ ⴙ  ⴙⴙ

ⴙ/() ⴙⴙ ⴙⴙⴙ

ⴙⴙⴙⴙ ⴙⴙⴙⴙ () () ⴙ

ⴙⴙⴙ

17,189,201 201 201 201 17,189 244 17,244 244 244 244 77

ⴙ ⴙⴙⴙ ⴙⴙ () ⴙⴙ In fibers

()/

ⴙⴙⴙⴙⴙ/ⴙⴙⴙ ⴙⴙⴙⴙ

ⴙⴙ

*IR  immunoreactivity; messenger RNA (mRNA) shown by in situ hybridization. † Approximate percentages of neurons that were retrogradely labeled via nerves that project to muscle, skin or viscera, that also expressed the marker indicated in the first column. The exception was the 3 Na pump, which is based on fiber staining. ‡ Splanchnic includes retrograde labeling via the splanchnic nerve and from organs innervated by the splanchnic innervation. Note that, of DRG neurons innervating epidermis, 14% were TrkA positive and 70% showed IB4 binding.189 Percentages indicated as follows: ⴙⴙⴙⴙⴙ  100%; ⴙⴙ  50%; ⴙ  20%;   10%; ()  5%;  none.

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons

vasoactive intestinal polypeptide (VIP), cholecystokinin (CCK), neuropeptide Y (NPY), and galanin326 (Fig. 8–5B). NGF and Nociception. NGF is important for the development and maintenance of the normal phenotype of nociceptive neurons by controlling expression of many different molecules.204 When NGF levels were lowered over a critical period (postnatal days 4 to 11 in the rat), fewer A nociceptors and more LTM D hair units were reported, perhaps indicating that A neurons may develop into D hair units rather than nociceptors in the absence of NGF.204 Raised NGF is thought to have acute local effects on peripheral terminals of DRG neurons. For example, NGF causes enhanced responses of cultured DRG neurons to capsaicin within 10 minutes, and in vivo it causes heat hyperalgesia that is dependent at 1 hour on mast cells and at 5 hours on neutrophils.19,285 The mast cell involvement is thought to be due to NGF-triggered NGF release presumably via activation of mast cell TrkA receptors.285 Direct analysis of TrkA immunoreactivity in neurons with identified sensory properties in vivo found intense TrkA-LI expression only in nociceptive neurons with C, A, or A fibers; both the percentage of positive neurons and the intensity of the expression were as high in A-fiber as C-fiber nociceptors.86 In addition, a few D hair units were weakly positive.86 NGF and Inflammation. NGF is upregulated in inflamed tissues. For instance, in inflamed skin it is upregulated secondary to increased levels of cytokines such as interleukins and tumor necrosis factor- (TNF-),336 and this increased NGF is “necessary and sufficient” for the production of inflammatory hyperalgesia.204,336 Its transport to DRG neuronal somata is increased, causing altered expression of many molecules important in the function of nociceptive neurons (see earlier), and leading to altered properties of nociceptive neurons. These altered properties probably contribute to hyperalgesia. They include, in adult rats or guinea pigs, lowered threshold to heat but not mechanical stimuli, increased spontaneous firing, and increased fiber firing rate; these changes and the hyperalgesia are prevented if NGF is sequestered in the inflamed tissues.75,159,337 Some of the electrophysiological changes take more than 1 day to develop—enough time for retrograde transport of NGF, altered gene expression (see earlier), and transport of products of gene expression.75 Apart from its effects on primary afferent neurons, elevated NGF has many effects on local inflammatory and immune responses. For instance, it acts as a chemotactic factor for leukocytes (including neutrophils) and is important in wound healing.163 NGF and Axotomy. Axotomy interrupts the supply of peripheral target-derived NGF to the DRG neuronal soma, and some DRG neurons die. After sciatic nerve lesion, NGF is upregulated in non-neuronal cells in the sciatic nerve, triggered, at least partially, by interleukin-1 (IL-1) released from invading macrophages. However, this NGF

177

increase does not fully replace the normal supply from the periphery.123,186 NGF plays a trophic role after nerve injury, including protection against neuron death.314 Regrowth of damaged fibers within the nerve is said not to be dependent on NGF and may even be encouraged by reduced NGF, although increased NGF encourages collateral sprouting of cutaneous afferent fibers within the skin.68,202,314 TrkB. TrkB, the high-affinity receptor for BDNF and NT-4/5,138 is normally expressed in about half the neurons in rat DRGs, in small- to medium-sized neurons, with about 30% having NF-rich somata. TrkB protein was also found in about 30% of human DRG neurons.136 After nerve injury the number of DRG neurons expressing it increases by 15% to 20%.89 BDNF. BDNF is a survival factor for some sensory neurons during development.47 In the adult, BDNF is expressed in several sites, including CNS DRG neurons and nerve bundles in skin.27 It is expressed in 30% to 50% of DRG neurons; positive neurons are small- to medium-sized, mainly TrkA-expressing neurons; it is also expressed in some TrkA-negative/IB4-positive neurons and in some TrkCpositive but in few TrkB-positive neurons.149,207,246 BDNF can be upregulated by NGF in the BDNF-expressing, TrkApositive neurons. It is downregulated by NT-3139 both in TrkC- and non–TrkC-expressing neurons. BDNF is transported both peripherally and centrally in DRG neurons; centrally it is found in dense-cored and clear vesicles in the superficial dorsal horn (laminae I and II), where it is colocalized in many nerve terminals with CGRP and p75.190,207,317 It is released in the dorsal horn in response to bursts of firing in afferent C fibers and has been proposed to act as an excitatory neurotransmitter/neuromodulator, acting on TrkB receptors on spinal dorsal horn neurons.246 As regards its effects on DRG neurons, evidence from transgenic mice implicates it in the maintenance of the sensory properties of SA mechanoreceptive neurons.47 It can also cause acute heat hyperalgesia.285 BDNF and Inflammation. During peripheral inflammation there is an NGF-dependent upregulation of BDNF in TrkA-expressing DRG neurons and in nerve terminals in the dorsal horn; there is evidence to suggest a role for BDNF in inflammation-induced allodynia/hyperalgesia, perhaps by contributing to central sensitization.56,57,195,317 BDNF and Axotomy. Peripheral nerve injury results in complex changes in BDNF expression in affected DRG neurons. Axotomy of peripheral sensory fibers results in long-lasting (several weeks) changes in BDNF expression of two kinds. There is upregulation of BDNF in somata of medium- to large-sized, non–TrkA-expressing DRG neurons, including TrkB- and TrkC-expressing neurons and in their central projections in deep dorsal horn and gracile nucleus. At the same time, a decreased expression of BDNF is seen in small-diameter neurons and in the

178

Function of the Peripheral Nervous System

superficial dorsal horn.55,144,208 Crush,55 CCI,227 and sciatic nerve section139 are complex nerve injuries that result in a mixture of intact and damaged/axotomized fibers in the nerve, or neurons in the DRG. In such injuries increased BDNF expression occurs in intact small- to medium-sized DRG neurons and in superficial layers of the spinal cord, as well as in large DRG neurons. After sciatic nerve section there was a substantial increase in expression in medium to large neurons for at least 3 weeks.139 Possible reasons for the complex pattern after complex nerve injury are as follows: Axotomy results in loss of NT-3 from the periphery (NT-3 normally downregulates BDNF expression in medium to large neurons) and a loss of NGF (NGF normally maintains, i.e., upregulates, BDNF expression in the smaller TrkA-positive neurons). This would result in BDNF downregulation in small neurons and upregulation in large neurons in a purely axotomized population of neurons. After L5 spinal nerve injury, there is increased expression of BDNF in small- to medium-sized neurons in the adjacent “intact” L4 DRG.93,115 These intact DRG neurons, whose fibers run adjacent to axotomized degenerating fibers, may (1) be subject to inflammatory mediators/NGF as a result of degeneration of the axotomized fibers and (2) have more NGF availability at their terminals, both of which may contribute to the BDNF upregulation in smaller DRG neurons. It has been suggested that increased release of BDNF in the dorsal horn after nerve injury might contribute to central sensitization, although the relative roles of the increased expression in axotomized large primary afferent neurons, or the NGF-related increase in intact small- to medium-sized, probably mainly nociceptive neurons is as yet unclear. NT-4. NT-4 is secreted by a variety of cell types, with high levels of expression embryonically, reducing around birth and increasing in some tissues postnatally. In the adult rat and human it is expressed in skin (including hair follicles), mainly in keratinocytes,27,113 and it is upregulated in human inflamed skin.113 Like BDNF, it can cause acute heat hyperalgesia, an effect dependent on functional mast cells.285 NT-4 is also widely expressed in the CNS.179 The loss of about 75% of TrkB-positive DRG neurons in NT-4 knockout mice299 shows NT-4 to be essential for survival of TrkB expressing neurons during development. Studies on knockout mice299 indicate that survival of DRG neurons with properties of D hair receptors is dependent on NT-3 early in postnatal development and on NT-4 later in the mature animal. TrkC. TrkC is the high-affinity receptor for NT-3. It is expressed by a subpopulation of medium to large DRG neurons, mainly those that project to skeletal muscle (see Table 8–3). TrkC knockout mice lack Ia muscle afferent projections to spinal motor neurons, have fewer large myelinated axons in the dorsal root and dorsal columns, and show movement disorders.155 Studies on knockout mice indicate that many DRG neurons depend on TrkC activation

in early embryonic development and that a substantial number of these switch their dependence to TrkA activation between 11.5 and 13.5 embryonic days334; they also provide evidence that survival of D hair LTMs in early postnatal development is dependent on NT-3.299 NT-3. NT-3 is critical for the survival and function of a population of large DRG neurons (in rat, about 20% of all DRG neurons) that express carbonic anhydrase and parvalbumin; these neurons are probably proprioceptors and cutaneous mechanoreceptors (slowly adapting A units or D hair receptors).82,349 NT-3 is necessary for the development, survival, and/or function of a subpopulation of muscle spindles, Golgi tendon organs, Merkel cells, and D hair receptors.82,349 In adult rats NT-3 is detectable in large neurons most (94%) of which express TrkC, suggesting that the NT-3 is derived from target tissues.54,314 In adults NT-3 is expressed in skeletal muscle and in skin, in skin muscles (canniculus parnosus and arrectores pilorum) and hair follicles.27 NT-3 signaling appears to be necessary for the development of the proprioceptive phenotype.226 Furthermore, there is emerging evidence that, although NT-3 does not affect TrkC expression in adult DRG neurons, it can downregulate TrkA expression especially in neurons with high TrkA expression but not TrkC expression, possibly acting via TrkA receptors.111 NT-3 and Axotomy. Application of NT-3 to the axotomized nerve enhances nerve regeneration and reverses the upregulation of NPY in axotomized neurons.296 p75. p75 is a member of the TNF receptor superfamily. It binds to all the neurotrophins (NGF, BDNF, NT-3, and NT4/5) with a similar low affinity (see Chapter 17). It has important roles in both apoptosis (having a death domain similar to that of many of the TNF receptor family) and in cell survival (for review, see Casaccia-Bonnefil and colleagues48), and these roles are modulated by NGF/TrkA signaling. For example, its absence in p75 knockout mice results in a 50% loss of DRG neurons.300 In adult DRG neurons it is expressed in TrkA-, TrkB-, and TrkCexpressing DRG neurons.140,143,339 Relationships (either ligand passing or binding between receptors) between p75 and Trk receptors can influence the function of these receptors. For example, the presence of p75 enhances ligand specificity for the Trk receptors. Thus, although BDNF, NT-3, and NT-4/5 may each bind to TrkB, TrkB activation by NT-3 and NT-4/5 is reduced if p75 is present, making TrkB more selective for BDNF.266 Furthermore, although NGF and NT-3 can both bind to and activate TrkA of p75, TrkA shows a greater affinity for NGF in the presence and the activation of TrkA by NT-3 is reduced.266 Thus the selectivities of TrkB for BDNF and of TrkA for NGF are both strongly enhanced by the presence of p75. There is evidence that the pro-apoptotic effects of p75 are normally suppressed by NGF/TrkA signaling but that after

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons

axotomy, with the loss of target-derived NGF, p75 activation does indeed become pro-apoptotic (see Chapter 17). Receptors for the GDNF Family of Trophic Factors Besides the neurotrophins, a number of other neurotrophic factors are thought to influence DRG neurons and may play roles in the regeneration of peripheral nerves. An important family is the GDNF family, which includes GDNF, neurturin, artemin, and persephin; apart from persephin (which is not further discussed) these all support survival of peripheral neurons in vitro. In the adult, GDNF is found in the CNS, ovary, prostate, adrenal gland, and skin; it is also found in skeletal muscle (human, not mouse) and in axons and Schwann cells in peripheral nerve.105,305 Peripheral sources of GDNF for DRG neurons normally include basal keratinocytes and Schwann cells, with upregulation in Schwann cells after nerve injury.7 Sources of artemin are not yet established. Neurturin is expressed in adult mice in certain visceral organs (gastrointestinal tract, bladder, reproductive organs) and skin.105 Many DRG neurons that are NGF dependent in embryonic life switch dependency to GDNF, neurturin, and/or artemin postnatally.13 The GDNF family of growth factors acts through a receptor complex with two components. This includes a common signal-transducing domain receptor tyrosine kinase RET (or c-Ret), and one of four ligand-binding domains (GDNF family receptor  [GFR]-1 through -4). GDNF acts preferentially through GFR-1, neurturin through GFR-2, artemin through GFR-3, and persephin through GFR-4 (see Chapter 17). GFR-1 through -3 receptors are expressed by mammalian DRG neurons. The growth factors are bound and transported centrally in DRG neurons that express the appropriate ligand-binding domains. Overall about 60% of DRG neurons express RET; these neurons are mainly small to medium with some large neurons, and express one or more GFR subunits147 and many show IB4 binding (see later). Reported percentages of adult rat DRG neurons expressing mRNA for RET, GFR-1, GFR-2, and GFR-3 are 60% to 64%, 40% to 50%, 20% to 33%, and 20%, respectively.18,147 RET, GFR-1, and GFR-2 are in both small and large DRG neurons.18 A higher proportion of large neurons express GFR-1 than GFR-2, and 30% of GFR-1– positive neurons are NF rich, and these probably have myelinated fibers.18 About 20% of RET-positive neurons are TrkC positive147 and thus likely to be LTMs. mRNAs for GFR-2 and GFR-3 were shown to be highly co-localized but have low co-localization with GFR-1 mRNA.147 Few GFR-1 (20%) or GFR-2 (3%) positive neurons were reported to express TrkA,18 (see below for comment) but GFR-3 immunoreactivity was described as having substantial colocalization with TrkA.232 These GFR-3–immunoreactive neurons were 20% of mouse and rat lumbar DRG neurons, most of which (90%) were NF poor and positive for RET, TrkA, CGRP, and transient receptor potential protein V1

179

(TRPV1), and about 30% were IB4 positive,232 features generally indicative of slow conduction velocities and nociceptive properties. However, there is a discrepancy in the above literature between the high co-localization between GRF-1 and TrkA and the high co-localization of GFR-2 and GFR3, coupled with the very low co-localization of GRF-2 and TrkA mentioned previously. The reasons for this are not clear. Thus GFR-3 may be expressed in nociceptive neurons, and, unlike other receptor components in this group, is almost exclusively expressed in sensory neurons, and thus is of potential interest as a therapeutic target (see later). Table 8–2 presents a summary of possible sensory properties of neurons that express the different receptor components. Nerve Injury. Nerve injury leads to upregulation of GFR and GDNF in Schwann cells in the nerve distal to the injury; the level in the nerve is related to the extent of infiltration by macrophages and lymphocytes. In the somata of axotomized DRG neurons, upregulation of GFR-1 in larger neurons and GFR-3 in smaller neurons, decreased GFR-2 levels and unchanged RET expression are reported.16 GDNF can reverse the following postaxotomy changes in DRG neurons: reduced IB4 binding; decreased expression of both somatostatin and the P2X purinergic receptor P2X3; increased expression of activation transcription factor 3 (ATF3), galanin, and NPY in large DRG neurons; slowing of conduction velocity and behavioral changes including tactile hypersensitivity and thermal hyperalgesia.216,332 Systemic exogenous artemin can also reverse or prevent a number of changes associated with nerve injury, including behavioral measures of neuropathic pain.95 Comparison of IB4- and TrkA-Labeled Populations. It is frequently stated that IB4 binding and TrkA expression define separate subpopulations of small or nociceptive neurons. Most or all of the population of small neurons that does not express any TrkA, B, or C shows IB4 binding. Thus most small neurons express either TrkA or IB4 or both. It is therefore worth comparing these two populations. IB4 is a lectin from the plant Griffonia simplicifolia that binds to -D-galactose residues in glycoconjugates on the cell surface and Golgi apparatus of small, NF-poor DRG neurons in a variety of species.214 In Figure 8–4 it can be seen that the staining pattern for IB4 appears to be reciprocal to that for the 200-kDa neurofilament subunit (NF200). Percentages of rat lumbar DRG neurons that bind IB4 increase from 9% at birth to 40% at 14 days postnatal,15 as they switch their dependence from NGF in embryonic life to GDNF in postnatal life.216 Reports on adult rat DRGs indicate that most or all (95% to 100%) IB4-positive neurons express RET while 63% of RET-positive neurons express IB4; up to 80% of IB4-positive cells express GFR-1 and about 50% express GFR-2, whereas only about 30% of GFR-3–positive neurons express IB4.18,216,232 The RET/ GFR receptor components show a distribution similar to

180

Function of the Peripheral Nervous System

that of IB4 in the inner part of lamina II in the dorsal horn.216 While all RET-positive neurons show retrograde transport of 125 I-labeled GDNF, only 13% of TrkA-positive neurons show such transport.216 Coexpression with TrkA is low for GFR1– and GFR-2– (although see discrepancy noted earlier) and high for GFR-3–expressing neurons.18,232 Thus overall, although there is some overlap between the RET/GFR1/IB4 and TrkA populations, the former take up and transport GDNF, whereas the TrkA-positive population binds and transports NGF and expresses little GFR-1 or GFR-2 mRNA, and half of this population expresses GFR-3.232 That there is some co-localization between TrkA expression and IB4 binding8,18 was confirmed with intracellular recording and dye injection studies in rat DRGs. These showed (1) that a third of C-fiber neurons were positive for both TrkA and IB4, with a tendency for reciprocal staining intensities for these two markers; (2) that most nociceptors strongly expressed TrkA or IB4 binding sites; (3) that IB4 binding sites were present on C-fiber but not A-fiber nociceptive neurons, whereas TrkA expression was in both C- and Afiber nociceptors86; and (4) that some weak positive labeling for TrkA and IB4 was seen in some D hair units.86 TrkA-, and not IB4-, positive neurons express substance P and CGRP.18 Other differences between these neurons include projection of TrkA-positive neurons to laminae I and IIo and IB4-positive neurons mainly to IIi.214 Compared with IB4-negative small neurons, IB4-positive neurons have longer duration action potentials and a smaller noxious heat-activated current,301 and the TTX-resistant (TTXR) Na channel subunit Nav1.9 is preferentially expressed in IB4-positive cells.87 In summary, A-fiber nociceptors express TrkA but not IB4 binding sites, while most C-fiber nociceptors express one or the other, or both of these. Apart from the NGF and GDNF dependence of TrkA expressing and IB4 binding neurons respectively, the functional differences between these groups of nociceptors are poorly understood. The Fibroblast Growth Factor Family and Their Receptors Almost all tissues express fibroblast growth factors (FGFs). The FGF family can influence growth, development, differentiation, and fiber branching and appears to play important roles in the responses of DRG neurons to injury. These are part of a larger family of at least 22 different FGFs. There are four FGF receptors (FGFR-1 through -4) that are tyrosine kinase receptors. Some of these have several isoforms. mRNAs for at least seven FGFs (2, 5, 7, 9, 10, 13, and 14) and several receptor types are present normally in DRGs, in neurons and/or non-neuronal cells.182 A subpopulation of DRG neurons that is NGF dependent during embryonic life switches to dependence on FGF-2 later in development; this group of neurons expresses the receptor for FGF-2, expresses substance P, and is NF poor, and thus presumably (see earlier) has C fibers.1 Skeletal muscle as well as DRG and spinal cord are sources of FGF-2, and it has been suggested that these

FGF-2–dependent DRG neurons may be muscle C-fiber nociceptive neurons.1 Axotomy. In DRGs following peripheral axotomy, some FGFs are upregulated, as are their receptors (e.g., FGF-1, FGF-2, FGFR-1 through -3), and some are downregulated (e.g., FGF-13).135,182 FGF-2 and FGFR-1 through -3 are upregulated in Schwann cells and macrophages at the injury site, and FGF-2 stimulates nerve regeneration and Schwann cell proliferation after axotomy.114 Neuropoietic Cytokines and Their Receptors These cytokines, expressed by Schwann cells, fibroblasts, and sensory neurons, as well as by macrophages, mast cells, and lymphocytes have local influences on neurons including sensory neurons, and on glial cells. Examples are members of the interleukin-6 (IL-6) family: ciliary neurotrophic factor (CNTF), leukemia inhibitory factor (LIF), IL-6, IL-1, TNF-, and transforming growth factor- and - (TGF- and TGF-). Many of these are upregulated in tissue damage and/or inflammation and may also be known as inflammatory mediators. (See Fig. 8–5A.) The IL-6 Cytokine Family. This family includes the structurally related molecules CNTF, LIF, IL-6, and IL-1. CNTF and LIF both support the survival of sensory neurons after axotomy and in culture.315 They have ligand-specific  receptors (CNTF-R, LIFR, IL-6R, IL-1R). They also have signal-transducing components one of which, gp130, is common to signaling by IL-6, CNTF, and LIF.315 gp130 is expressed in most DRG neurons and its expression is unchanged after axotomy.96 The soluble ligand-specific receptors can be released from cells and act as receptors on other cells.315 CNTF. CNTF is a neurotrophic cytokine expressed by glial cells in the peripheral nervous system and CNS. The CNTF receptor (CNTF receptor ) is expressed in all DRG neurons.200 Following nerve injury, the expression of CNTF in Schwann cells is reduced peripheral to the axotomy but increases as nerve regeneration begins, perhaps because axonal contact with glial cells is required for its expression.314 It enhances neurite outgrowth of DRG neurons and is upregulated by IL-6.287 IL-6. IL-6 is not normally detected in adult DRG neurons but is synthesized in a third of DRG neurons (medium and large sized) and in non-neuronal cells such as Schwann cells after nerve injury.219 After axotomy, endogenous IL-6 is pro-nociceptive and promotes DRG neuron survival; these effects may be in part due to its upregulation of BDNF in medium- to large-sized DRG neurons.219 IL-6 also enhances neurite outgrowth in axotomized DRG neuron fibers, possibly by upregulating CNTF and gp130 expression.287 IL-6 is upregulated in skeletal muscle after denervation.161 IL-6R. Almost all DRG neurons express IL-6R, although IL-6 supports the survival of only about 30% of embryonic

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons

DRG neurons; this is thought to be because IL-6 acts by induction of, for example, BDNF expression in DRG neurons.219 IL-6 also supports survival of newborn rat DRG neurons, but only if exogenous IL-6R is also provided.315 IL-1. IL-1 is expressed in most medium to large DRG neurons and some satellite cells.61 It enhances neurite regeneration, causes some afferent fibers in skin to fire spontaneously, can cause substance P release from DRG neurons, and induces cyclooxygenase-2 (COX-2) expression.129 IL-1 is upregulated in inflamed tissues, and an antagonist to IL-1 reduces inflammation-induced hyperalgesia.268 LIF. About 25% of rat DRG neurons retrogradely transport LIF in vivo. These are small neurons, mainly the CGRP/ TrkA-positive population, but also some IB4-binding neurons.318 LIF immunoreactivity was found in Schwann cells and nerve fibers. In damaged peripheral nerve, LIF upregulation is predominantly in Schwann cells. LIF promotes nerve regrowth and normal regeneration of nerve fibers after injury.37 In human nerves following injury, LIF expression is upregulated within hours and can stay high for years where a neuroma is present.79 During an inflammatory response, it is upregulated in neutrophils, macrophages, mast cells, and blood vessel walls.79 After denervation LIF is also upregulated in skeletal muscle cells.161 Although LIF can be pro-inflammatory, it has anti-inflammatory effects following CFA injection to the tissues.303 LIFR. LIFR is normally expressed in the nucleus of most DRG neurons, but in the cytoplasm of only a few; however, following peripheral axotomy, it is expressed for at least 2 weeks in the cytoplasm of a high proportion of small, medium, and large DRG neurons.96 TNF-. TNF- is a pro-inflammatory cytokine, as well as a “potent activator of Schwann cell and other glial cell cytokines,” and plays an important role in the early degenerative changes after nerve injury.286 TNF- is also strongly implicated in both inflammatory and neuropathic pain. It causes nociceptive fibers to fire spontaneously, and intraneural injection of TNF- induces neuropathic pain– related behavior in rats through the activation of its receptor TNF-R1.292 In a neuropathic pain model, it markedly enhances neuropathic pain behavior, including spontaneous foot lifting.270 Normally it is found in small-diameter DRG neurons and is transported peripherally in muscle but not cutaneous afferent DRG neurons.270 Intraneural TNF- is derived from Schwann cells and macrophages, although some may be expressed by DRG neurons themselves.270 Nerve Injury. After nerve injury, TNF- is upregulated in Schwann and satellite cells.228,270 It is transported centrally in DRG neurons via the DRG to the spinal cord, where it is released and taken up by spinal cord neurons (dorsal horn and motoneurons) and glia.286 At this time increased numbers of medium to large DRG neurons were reported to contain TNF- in both ligated L5 (injured) and nonligated adjacent L4 DRG neurons, although to what

181

extent this is derived from glial cells or DRG neurons is not clear.270 The two TNF- receptors TNFR1 (R55) and TNFR2 (R75) are both upregulated in damaged and in adjacent, presumed intact, DRG neurons after spinal nerve injury, peaking within 24 hours.228,271 In addition, TNFR1 is upregulated in satellite cells after nerve injury.228 TGF-. TGF- is a member of the epidermal growth factor (EGF) family, and is important in the control of glial and Schwann cell proliferation and survival of differentiated neurons.294 Within the DRG, TGF- and its receptor (EGF) are expressed mainly in small DRG neurons and in satellite cells surrounding some large- or medium-sized neurons.350 Nerve injury results in upregulation of TGF- in satellite cells, and of its receptor in all DRG neurons from 1 to more than 14 days.350 Its role in the DRG after nerve injury is not fully understood, although it may stimulate proliferation of satellite cells. (See Fig. 8–5B.) TGF-. The TGF- family is a superfamily of cytokines secreted by a variety of immune cells and with a variety of functions at least some of which are anti-inflammatory. During development they are necessary for the survival enhancement offered to DRG neurons by some of the neurotrophins. In the adult, they are involved in Schwann cell proliferation, secretion of extracellular matrix and neurotrophins, and control of expression of cell adhesion molecules. TGF- and TGF receptor 2 are found in satellite cells. TGF-2 is present in about half the DRG neurons and in fibroblasts in the DRG.294 It is expressed in mechanoreceptor-related structures in skin, being present in pacinian corpuscles (TGF-1 and -2) and in Meissner corpuscles (TGF-2), and TGF- receptors (I and II) were found in Merkel cells.294

Inflammatory Mediators and Tissue Damage There is substantial evidence of the importance of inflammatory mediators and the immune system in chronic pain states, with half the clinical cases of human neuropathic pain being associated with infection or inflammation of a peripheral nerve (see review by Watkins and Maier333). Pathogens and other types of tissue damage can activate inflammatory and immune responses. Pro-inflammatory mediators released locally from damaged tissues include bradykinin, ATP, K, H, and 5-hydroxytryptamine (5-HT or serotonin), and activation of the arachidonic acid pathway in these tissues results in production of prostaglandins, prostacyclin, leukotrienes, and thromboxanes that tend to be proinflammatory, acting as chemoattractants to encourage infiltration of immune cells. Such cells, as well as other local cells within peripheral nerves, including macrophages, T lymphocytes, mast cells, fibroblasts, endothelial cells, and Schwann cells in the nerve or satellite cells in the DRG, secrete further mediators such as cytokines and growth

182

Function of the Peripheral Nervous System

factors, including those described earlier. Many of these (e.g., NGF) are also pro-inflammatory.333 For example, peripheral nerve injury involves proliferation and activation of satellite cells; these release pro-inflammatory cytokines such as TNF-,228 as well as neurotrophic factors such as NGF, BDNF, and NT-3, and they also express p75.351 The ensuing Wallerian degeneration involves upregulation of TNF-, IL-1, and IL-1 in Schwann cells and in recruited macrophages in peripheral nerve280; the Schwann cells produce neurotrophic factors (such as NGF, GDNF, NT-4, and p75) (see reviews by Terenghi314 and Watkins and Maier333). Mechanical damage as well as pathogens may result in local tissue damage, including exposure of proteins (such as the P0 and P2 peripheral nerve proteins).333 These are normally hidden from the immune system but, once exposed, can trigger inflammatory and immune responses.333 In addition, mast cells and macrophages, when activated, produce destructive molecules such as digestive enzymes that are not only effective in destroying pathogens, but may also damage nerve tissue,333 thus triggering further involvement of inflammatory and immune processes in a positive feedback loop. Positive feedback loops involving cascades of mediators involved in inflammatory and immune responses may have the effect of developing and sustaining these responses. Primary afferent fibers are intricately involved in these processes. They may be activated or sensitized by a number of pro-inflammatory mediators, leading to greater intrinsic excitability. Sensitization of afferent terminals may be a local acute effect or may result from alteration of the neuronal synthetic activity, leading to altered expression or properties of molecules such as ion channels, receptors and/or neuropeptides, or insertion of novel channels and/or receptors into the membrane of the whole neuron. For example, activation or sensitization of nociceptive fibers by proinflammatory mediators can cause release from both peripheral and central afferent terminals of the pro-inflammatory peptides substance P and CGRP. These further enhance inflammatory and immune responses and may excite afferent terminals peripherally, and enhance transmission in nociceptive pathways centrally. Thus cascades of inflammatory mediators and sensitization of primary afferent neurons can cause positive feedback loops that are likely to be very important for developing and sustaining inflammatory or neuropathic pain.333 Some of the molecules that are up- or downregulated within DRG neurons or that might, via receptors on DRG neurons, influence DRG neurons are illustrated in Figure 8–5B for peripheral inflammation (top) or nerve injury (bottom).

Further Inflammation-related Receptors and Mediators Apart from the neurotrophic factors and cytokines, DRG neurons are influenced by a number of other substances released by damaged tissues, by local cells such as leukocytes,

fibroblasts, epithelial cells, and Schwann or satellite cells or even by the neurons themselves. Many of these substances could be classed as pro-inflammatory, or occasionally as anti-inflammatory, mediators. These are introduced below. Prostaglandins, Prostaglandin Receptors, and COX Enzymes. Prostaglandins are important inflammatory mediators. Apart from widespread pro-inflammatory effects, prostaglandins such as prostaglandin E2 (PGE2) have a number of effects on primary afferent neurons, including sensitization of nociceptors to bradykinin and other mediators as well as to natural stimuli.152 One mechanism contributing to nociceptor sensitization is that PGE2 lowers the activation threshold of the TTXR (tetrodotoxin resistant) inward Na current that is found in nociceptive DRG neurons.102 PGE2 is also implicated in mechanotransduction in the renal pelvis; here stretch causes the release of PGE2, which causes the release of substance P, which in turn activates primary afferent fibers.160 PGE2 binds to prostaglandin E (EP) receptors. Of the four EP receptor isoforms, EP1, EP3, and EP4 have been found in DRG neurons.11 Synthesis of prostaglandins requires COX (cyclooxygenase) enzymes, including the constitutive COX-1 and the inducible COX-2 isoforms. COX-2 is induced in peripheral tissues by cytokines, growth factors, and other inflammatory stimuli,152 resulting in increased prostaglandin release. DRG neurons express COX-1 in many of the SD neuron (C-fiber) population, both CGRP-positive and IB4-positive neurons, but its expression is unchanged by peripheral inflammation.58 COX-2 is not expressed in DRG neurons either normally or during inflammation.58 ATP Receptors. A discussion of purinergic receptors (P2X and P2Y receptors) is presented in the later sections on mechanotransduction and chemotransduction. Histamine1 Receptors. The predominant sensation when histamine is applied to the receptive field of human afferent C fibers is itch.272 Histamine sources include mast cells in rat. The histamine1 (H1) receptors are expressed on a subpopulation of small DRG neurons141,142 with unmyelinated fibers that bind IB4. After nerve injury these IB4 binding neurons downregulate H1, although it is upregulated in other mainly small neurons that express substance P– and CGRP.141 Bradykinin B1 and B2 Receptors. Bradykinin is algesic and pro-inflammatory. It is released as a result of damage to the tissues, and contributes to the inflammatory cascade. It causes pain, sensitization, and hyperalgesia in humans, and in a rat skin-nerve preparation it activates greater than 50% of C-polymodal nociceptors; all bradykinin-sensitive neurons are also capsaicin sensitive.162 Bradykinin acts via B1 and B2 receptors on DRG neurons and sensitizes TRPV1 receptors to heat.304 Both receptors are expressed in some of the small

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons

DRG neurons; B2 was strongly upregulated (in the same group of neurons that normally express it) by activation of TrkA receptors by NGF and also to a lesser extent by GDNF,177 and a B2 receptor antagonist (Bradyzide) reduces inflammatory hyperalgesia in animal models.152 Two weeks after CCI, B2 and B1 were upregulated in rat DRGs,180 although B1 expression in DRG neurons was decreased 2 weeks after axotomy (sciatic nerve section).90 Endothelin Receptors. There are three endothelins that bind to two receptors, ETA and ETB, with distinct effects. Endothelin 1 (ET1) can induce pain-related behavior through the ETA receptor.249 ETA receptors, when activated by ET1, activate nociceptive C and A fibers in rats in vivo,101 possibly by causing a lowered activation threshold for TTXR Na channels.352 ET1 is expressed by endothelial cells, macrophages, and Schwann/satellite cells as well as by small- to medium-sized DRG neurons. Skin injury causes its release. ETA receptors are found in small- and some medium-sized DRG neurons, but ETB receptors are mainly found on satellite cells or Schwann cells around unmyelinated fibers.249 Peptides. Several peptides released by DRG neurons can influence the inflammatory and immune systems. Substance P and CGRP are vasoactive, pro-inflammatory and proimmune, and somatostatin is anti-inflammatory and antihyperalgesic. For more detail see the appropriate section later under Neuropeptides. Serotonin (5-Hydroxytryptamine). 5-HT is released following tissue injury from mast cells, can cause pain or hyperalgesia,272,310 and is vasoactive. One of its effects is to influence the properties of the TTXR inward Na current typical of nociceptive neurons, lowering its threshold and making the neurons more excitable.41,102 Of the 15 receptor subtypes known, mRNAs for 6 have been described in DRG neurons: 5-HT1B, 5-HT1D, 5-HT2A, 5-HT2B, 5-HT3B, and 5-HT4 receptors.222 Nitric Oxide. Nitric oxide (NO) is synthesized by the enzyme nitric oxide synthase (NOS). Normally in DRGs endothelial cells express eNOS (the endothelial form), and a small population of small- to medium-sized peptide (substance P/CGRP)–containing DRG neurons express nNOS (the neuronal form), the expression of which is downregulated in these neurons by target-derived NGF.316 Axotomy. After peripheral nerve exotomy, nNOS mRNA and protein are upregulated in small- to medium-sized peptide-containing DRG neurons, perhaps as a result of the loss of target-derived NGF.316 It has been suggested (see review by Thippeswamy and Morris316) that NO release from axotomized neurons in neuroprotective, by (1) upregulation of cyclic GMP in satellite cells, which may initiate

183

expression of neuroprotective molecules and/or inhibit apoptosis, and (2) induction of NGF expression in satellite cells.

Other Receptors Adrenoreceptors. A small number of DRG neurons of all sizes normally express the 2A adrenoreceptor; this number was increased sixfold after peripheral axotomy, with the greatest increase being in medium to large neurons, but was unchanged after inflammation of the tissues.23 In contrast, the relatively low numbers of neurons normally expressing the 2C receptor remained unchanged after axotomy or inflammation.23 This increase in 2A adrenoreceptor expression may be important in sympathetically associated neuropathies. Cholinergic Receptors. Cholinergic agonists can excite peripheral sensory fibers in skin as well as excite isolated DRG neurons; in addition, cholinergic fibers make presynaptic connections with primary afferent fibers in the dorsal horn (for references, see Xiao et al.342) Nicotinic acetylcholine receptor  subunits (nAChR) 2 through 7 have been reported in DRG neurons,98 and nAChR7 is upregulated in DRGs after axotomy.342 Cannabinoid Receptors. Cannabinoids have an antihyperalgesic effect in the periphery, thought to be mediated by their actions on CB1 receptors on DRG neurons and on CB2 receptors on non-neuronal cells involved in generation of inflammation.151 CB1 receptors are expressed mainly by medium to large NF-rich (presumably A-fiber) DRG neurons, but there is also some expression in smaller, possibly C-fiber neurons.30,151 GM1 Receptors. The ganglioside GM1 receptor, which has a high affinity for the cholera toxin B subunit, is normally expressed in nearly all NF-rich LL DRG neurons.265 However, after peripheral axotomy, this ceases to be a selective marker for LL neurons because SD neurons also express it.321

Oligosaccharides Expressed by DRG Neurons Cell surface glycoconjugates decorate the surface of several populations of DRG neurons (for a review, see Lawson165). Other surface markers of large DRG neurons include the antibodies to stage-specific embryonic antigen-3 and -4 (SSEA3 and SSEA4), and the lectins GSAII (Griffonia simplicifolia agglutinin II), PSA (Pisum sativum antigen), and LCA (Lens culinaris antigen), while markers of small cells include many antibodies to glycoconjugates, including lactoseries carbohydrates and lectins IB4 and PNA (peanut agglutinin).78,165,298 (For information about neurons to which the lectin IB4 binds, see earlier.)

184

Function of the Peripheral Nervous System

TRANSMITTERS, NEUROMODULATORS, PEPTIDES, AND THEIR RECEPTORS Some of the possible transmitter or modulator substances expressed by DRG neurons are mentioned briefly here. It is not the remit of this chapter to discuss the postsynaptic receptors for these transmitters. However, brief mention of receptors is made where these are expressed on DRG neurons because this may indicate a function on presynaptic terminals in the dorsal horn and/or on fibers in the periphery or even on the DRG soma. Possible transmitters or modulators known to be expressed by DRG neurons that have so far been identified include the excitatory amino acids glutamate and aspartate, the peptides substance P and CGRP, peptides upregulated after nerve injury (galanin, NPY, and VIP), and BDNF (see earlier under Neurotrophins and Trk Receptors), all of which are released from terminals of DRG neurons and may act as transmitters or neuromodulators. This is not a complete list, nor is it likely that the complete list is yet known. Complex cross-interactions between transmitters and receptors for different transmitters/modulators (e.g., tachykinins on N-methyl-D-aspartate [NMDA] receptors) are not covered here. See summary Figures 8–3 and 8–5.

Excitatory Amino Acid Transmitters Glutamate immunoreactivity is in about 70% of DRG neurons, and is also seen in primary afferent fibers in the dorsal horn, in some cases co-localized with substance P.66 Glutamate and to a lesser extent aspartate immunoreactivity was seen in primary afferent terminals in the superficial dorsal horn, but only glutamate was also in laminae III and IV.324 These excitatory amino acids are thought to act as transmitters from the primary afferent terminals, with the widespread expression of glutamate making it a candidate transmitter for most primary afferent synapses. In addition, DRG neurons express a number of receptors for glutamate (GluR), which they transport both centrally and peripherally along their fibers. Peripherally, glutamate receptors may have a role in sensory transduction at the periphery. Glutamate receptors fall into several classes: the NMDA, amino-5-methyl-4-isoxazole-pro acid (AMPA), and kainate receptors. To date, limited immunocytochemical or in situ hybridization studies have suggested the following. Most DRG neurons express the NMDA receptor subunit 1 (NMDAR1).269a AMPA receptor subunits are also expressed: GluR1 by small DRG neurons and GluR2/3 by both small and large neurons;269a the GluR5 subunit of the kainate receptor is expressed strongly in many small DRG neurons (see review by Ruscheweyh and Sandkuhler267). NMDA receptors are found on central terminals of putative nociceptive neurons; NMDA, AMPA, and kainate receptors are also found on both myelinated and unmyelinated fibers in skin.42 Behavioral responses to glutamate, NMDA, AMPA, or kainate injected into skin or intra-articularly are indicative of allodynia and hyperalgesia, and these behaviors

can be blocked by appropriate antagonists or enhanced by local substance P injection.42 For further information on glutamate receptor function, see reference 211. Nerve Injury. Nerve injury results in increased immunoreactivity for the GluR2/3 AMPA receptors in primary afferents in the dorsal horn.250 Inflammation. Glutamate levels increase in inflamed tissues and in synovial fluid of human arthritic joints. Sources of glutamate may include afferent (possibly nociceptive) fibers, dermal/epidermal cells, macrophages, and Schwann cells.42 Peripherally applied glutamate receptor antagonists can attenuate the behavioral responses to inflammation of the tissues, and 48 hours after inflammation induced by CFA, increased numbers of fibers in the tissues show expression of NMDA, AMPA, or kainate receptors; although a simultaneous decrease in NMDAR1 in DRG neuronal somata is seen, possibly because the receptor is transported out of the somata to the fibers/terminals.42

Neuropeptides The term peptide-containing is commonly used to refer to the population of DRG neurons that normally expresses CGRP/substance P. However, other peptides are expressed in DRG neurons, including somatostatin, VIP, galanin, dynorphin, and enkephalin/leu-enkephalin. Substance P and CGRP. Substance P and neurokinin A are tachykinins that are expressed by DRG neurons.162 About 20% of rat and 15% to 40% of human DRG neurons express substance P, and 40% to 60% of rat or human DRG neurons express CGRP148,165 (S. Border and S. N. Lawson, unpublished data). The most intense expression of both is in small neurons, with clear expression in medium (both peptides) and large (CGRP) neurons. Figure 8–4 illustrates the appearance of rat DRG neurons that express CGRP. Substance P–expressing DRG neurons are small,165 and in rat almost all also express CGRP.291 Most (84%) DRG neurons that express CGRP in rat also express TrkA mRNA, and CGRP expression is upregulated by NGF; in contrast, few express TrkB or TrkC mRNA.8,148 Functions. In guinea pig, substance P was in half the cutaneous nociceptive DRG neurons examined, mainly in those with C fibers, but also in some with A and a few with A fibers.170,172 CGRP-positive neurons were mainly nociceptive with C, A, and A fibers, although a few LTM units with A fibers were also weakly positive.169 Noxious stimuli cause both substance P and CGRP to be released from both peripheral and central terminals of DRG neurons262; both can cause depolarization of dorsal horn neurons, and there is considerable evidence that substance P acts as an important neurotransmitter or neuromodulator in the spinal cord.291 CGRP may act to prolong the central effects of substance P released from primary afferent neurons.291 Peripherally released substance P and CGRP cause vasodilation, substance P causes extravasation of plasma

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons

from small blood vessels,126 and both substance P and CGRP are pro-inflammatory and promote wound healing.28 Substance P/CGRP– containing nociceptive afferent fibers are therefore important for integrity of the tissues, in influencing the local blood supply, and in local inflammatory and immune responses. The expression of both peptides is upregulated by NGF; thus in nerve injury both substance P and CGRP are downregulated in DRG neurons, whereas inflammation causes their upregulation.38,125 RECEPTORS. Substance P is thought to act via the neurokinin-1 (NK-1) receptor291; all three neurokinin receptors (NK-1 through -3) are expressed on DRG neurons.29 Substance P injected into rat skin was reported to cause mechanical hyperalgesia and allodynia; this effect may have been via activation of the NK-1 receptors that are seen on 30% of cutaneous unmyelinated fibers.45 CGRP receptor component protein was shown immunocytochemically in small- to medium-sized DRG neurons.192 Inflammation and Nerve Injury. A 50% increase in the proportion of unmyelinated fibers in rat skin that express NK-1 receptors was found 48 hours after CFA-induced inflammation.43 The expression of CGRP receptor component protein was increased by inflammation and decreased by nerve injury.192 Somatostatin. Between 5% and 15% of DRG neurons in rat express somatostatin-LI, with positive neurons being exclusively small, mainly IB4 positive, and NF poor, thus probably with C fibers.146,165 Somatostatin expression is upregulated by GDNF,18 and GDNF can reverse its downregulation after axotomy.18 Unlike substance P and CGRP, somatostatin-expressing neurons do not express TrkA, TrkB, or TrkC mRNA, and thus are unlikely to be regulated by NGF, BDNF, or NT-3/4.148 Possible Functions. Somatostatin is released centrally in the spinal dorsal horn from primary afferents during heat- but not mechano-nociception, and it is also released peripherally from DRG neurons during tissue injury/ inflammation.194 Somatostatin can inhibit firing in both dorsal horn and DRG neurons, and peripherally it attenuates the inflammatory/immune response194 and is thus both antinociceptive and anti-inflammatory (see, e.g., Malcangio et al.194). GDNF acutely promotes activity-induced, but not basal, somatostatin release from sensory nerve terminals in the dorsal horn.194 The afferent receptive properties of somatostatin-expressing neurons are not yet established. Somatostatin Receptors. There are several somatostatin receptor (SSTR) subtypes: SSTR1 through -5, with two forms of SSTR2, a and b. Several are expressed in the dorsal horn, but only SSTR2a has been found in DRG neurons, in a small group of small- to medium-sized DRG neurons that do not express the peptides substance P or CGRP.44,276

185

Galanin. Galanin is detectable in few rat DRG neurons normally, mainly in small substance P/CGRP–expressing neurons, although there is evidence of its synthesis in a higher proportion.187 After nerve injury it is strongly upregulated, and after inflammation it is initially downregulated followed by later upregulation “suggesting transition from an inflammatory to a nerve injury state.”38,187 Galanin Receptors. There are three galanin (GAL) receptors: GAL1, GAL2, and GAL3. GAL1 mRNA is present in half the DRG neurons, in large neurons, whereas GAL2 mRNA is expressed in 80% of DRG neurons, but these are mainly small DRG neurons; in addition, many lamina II neurons express GAL2.187 In DRGs, after axotomy GAL1 is strongly reduced and GAL2 moderately reduced.187 After inflammation, DRGs show decreased GAL1 mRNA and a transient increase in GAL2.187 It has been suggested187 that pronociceptive effects of galanin are due to activation of GAL2 receptors on primary afferent terminals and that high levels of galanin protect against nerve injury–induced pain by acting on postsynaptic GAL1 receptors. VIP and Pituitary Adenyl Cyclase–Activating Protein. VIP and pituitary adenyl cyclase–activating protein (PACAP) are members of the same glucagon/ secretin superfamily and are widely expressed in the CNS. They are normally present in primary afferent fibers in laminae I and II of the dorsal horn and are both expressed at low levels in small- to medium-sized DRG neurons.70 There is evidence that PACAP may promote differentiation of DRG neurons in rat, via the PAC1 receptor.223 Because PACAP and VIP are markedly upregulated (PACAP over 2 days and VIP over 2 weeks) in small DRG neurons after nerve injury, they may play different roles in neuropathic pain development.70 Neuropeptide Y. NPY is not normally expressed in DRG neurons but is upregulated in medium to large DRG neurons after axotomy, in the same neurons that upregulate BDNF after axotomy (see later).144,184 This increase in NPY remains until reinnervation occurs and is partially reversed in large DRG neurons by administration of NT-3, perhaps because NT-3 enhances regeneration, or because NT-3 causes a downregulation of NPY, or both.296 Centrally NPY inhibits activity-evoked substance P release.81 Opioid Receptors. Endogenous enkephalins and opiates affect transmission of nociceptive information by activating postsynaptic K channels in the CNS and inhibiting presynaptic Ca2 channels on DRG neurons.14 They act via opioid receptors on central (CNS) or peripheral terminals (e.g., in skin) of afferent fibers. A combination of in situ hybridization and functional studies indicate the following. The , , and opioid receptors (MORs, DORs, and KORs) are expressed by adult DRG neurons, mainly small neurons, with MORs but not DORs being

186

Function of the Peripheral Nervous System

co-localized with substance P/preprotachykinin A.134,213 Activation of opioid receptors inhibits whole-cell Ca2 currents in small but not large DRG neurons, and activation of DORs on dissociated cultured DRG neurons from young rats reduced N-, L-, P-, and Q- but not R-type Ca2 currents.2 Peripheral activation of MORs can inhibit glutamate-induced activity in (possibly nociceptive) primary afferent fibers in skin.59 Interestingly, MORs are downregulated after axotomy and upregulated after peripheral inflammation, whereas DORs and KORs are downregulated in DRG neurons after inflammation.134 Opioid-like Receptors. Nociceptin, the endogenous ligand for opioid-like receptor 1 (ORL1) receptors, is algesic, inducing thermal hyperalgesia–type behavior in rats. Prepronociceptin in RNA is expressed in a few small DRG neurons that do not themselves express substance P and CGRP but are adjacent to neurons that do express these peptides.210 In contrast, ORL1 is abundantly expressed in small- to medium-sized neurons many of which express substance P and CGRP, suggesting a possible presynaptic action of nociceptin.210 Peripheral inflammation causes rapid (30 minutes, recovering by 6 hours) upregulation of prepronociceptin mRNA in a subpopulation of TRPV1expressing small- to medium-sized DRG neurons.132

CYTOCHEMISTRY OF CUTANEOUS, VISCERAL, AND MUSCLE AFFERENTS Retrograde labeling studies on primary afferent neurons that project to different types of tissue show patterns in the cytochemical properties of afferents projecting to skin, muscle, and viscera. These patterns may reflect differences in afferent receptor type, differences in trophic factor availability in the tissues, or perhaps differences in conduction velocity or modality of neurons projecting to the different types of tissue. Such studies often involve either (1) cutting the nerve projecting selectively to one of these types of tissue and applying dye to the cut end of the nerve, or (2) injecting the dye into the intact tissues. Technical problems accompanying such studies include phenotypic changes of neurons with cut or damaged fibers, damage to the nerve terminals that take up the dye in the tissues, and spread of dye into other tissues (see Fox and Powley92 and Perry and Lawson244). A summary of some of the differences in chemical phenotype of neurons projecting to skin, muscle, and viscera is shown in Table 8–3. Note that the percentage of visceral afferents labeled by a particular marker may differ according to the type of innervation (splanchnic compared with vagal afferents). Overall, splanchnic visceral afferents have the highest expression of TrkA, TrkB, substance P, and CGRP; muscle afferents have the highest expression of TrkC, carbonic anhydrase, and the 3 sodium pump subunit; and cutaneous

afferents have the highest expression of binding for several lectins, including IB4, with particularly high IB4 expression in afferents projecting to epidermis (see Table 8–3). Several markers are selectively associated with a subpopulation of DRG neurons with a generally large size; these may be related to LTMs, including cutaneous or proprioceptive afferents. These include cytochrome oxidase, carbonic anhydrase, parvalbumin, and calbindin D28k (see Carr and Nagy46 and Lawson165). It was suggested that carbonic anhydrase was localized mainly in muscle afferents, and in rat, DRG neurons with carbonic anhydrase activity were those with short-duration action potentials and short afterhyperpolarizations.46,251 More recently, immunocytochemical evidence indicates that the 3 subunit of the sodium pump (Na,K-ATPase) is expressed, perhaps selectively, in afferent and efferent fibers that innervate intrafusal muscle fibers of muscle spindles.77

MEMBRANE RECEPTORS AND SENSORY TRANSDUCTION In recent years a number of molecules have been cloned that are thought to be responsible for transduction of different types of sensory stimuli. From the transduction properties of molecules expressed in subpopulations of sensory neurons, it is possible to make more or less informed deductions about the possible or likely sensory receptive properties of the DRG neurons that express them. These are summarized in Table 8–4 and a brief introduction to them is provided below. Also see summary Figures 8–3 and 8–5A and B. Many studies of the transduction mechanisms have been made in heterologous expression systems, or from patch-clamp studies of dissociated DRG neurons in vitro. Other information is derived from defects in behavioral responses to selected stimuli in transgenic (knockout) mice. Thus the certainty with which predictions of the possible physiologic functions that different receptors might have in sensory neurons in vivo is very variable. This is a fast-moving and changing field, and the table is aimed at providing starter references and illustrating some of the possibilities as seen at the time of writing.

Thermoreceptive Molecules Heat Transduction Several candidate molecules for detection of different thermal stimuli have recently been identified and cloned. These include several members of the TRP protein family. TRPV1 (VR1) was cloned in 1997; it is the capsaicin receptor and responds to noxious heat (42° C).50 It has a thermal threshold similar to that of C-polymodal nociceptors and is expressed in small DRG neurons. Because capsaicin appears to selectively activate C-polymodal nociceptors,150,309 an (unconfirmed) possibility is that TRPV1 may be responsible for noxious heat transduction in these

187

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons

Table 8–4. Putative Transduction Molecules Stimulus Type* Molecule

Mechanical

TRPV1 TRPV2 TRPV3 TRPV4 TRPM8 TRPA1/ANKTM1 ASIC/ASIC1 DRASIC/ASIC3 BNC1/ASIC2 Stomatin MEC-2 P2X3 P2Y1

Chemical

Thermal

Size (S/M/L)

? Afferent Type†

References

Low pH, capn

42° C

S

50

52° C 39° C 32° C

SML SML SM NF-rich

C PM or heat nociceptors A-MN Warm A-MN

22° C 17° C

S S

C-cooling C-cold?

(S)ML SML S L

C PM; A-MN, LTMs LTM RA Mech (in C. elegans) Nociceptor LTM

199,238 297 6 6,253 6,252 196 220 220

Noxious Mech, stretch Menthol

Mech? LTM? Mech? Mech? LTM?

DRG Neurons

Low pH Low pH Low pH

Heat ?

ATP ATP

49 6,33,239,290 306,307

A  A fiber; C  C fiber; capn  capsaicin; C PM  C-fiber polymodal nociceptor; L  large; LTM  low-threshold mechanoreeptor; M  medium; Mech  mechanoreceptive; MN  mechanoreceptor; NF  neurofilament; RA  rapidly adapting; S  small; ?  speculative. *Stimulus type refers to the type of stimulus that can activate the molecule, usually in a heterologous expression system. Note that, under “Thermal,” thermal thresholds given are for channels in such expression systems and thresholds may differ in intact DRG neurons, as appears likely for the menthol-sensitive response to cooling (presumably TRPM8) in intact DRG neurons, which has a threshold of about 30° C. An added caveat is that a measured thermal threshold may depend on the previous holding temperature. † Refers to afferent properties of neurons (suggested in the references cited in the last column) that might express these molecules, although in no case has this been directly determined.

neurons and/or in specific heat nociceptors. TRPV2 (VRL-1) is present in medium-sized and large neurons and responds to temperatures greater than 52° C; it was suggested that this may be responsible for the higher thermal threshold of A-fiber mechanoheat nociceptors.49 More recently, two receptors that respond to warm, rather than noxious, heat have been cloned. These are again TRP molecules and are called TRPV3239,290,343 and TRPV4.307 Interestingly these are not only expressed in DRG neurons and fibers but are also reported to be in epidermal keratinocytes, raising the possibility that keratinocytes might contribute in some way to activation of these receptors. TRPV3 shows co-localization with TRPV1 in DRG neurons, with which it can form heteromeric vanilloid channels.290 TRPV4 is expressed in about 10% of lumbar DRG neurons; many (80%) of these are NF rich, and thus presumably myelinated (see earlier), although interestingly most were small.307 Cold Transduction Two cooling/cold receptors have recently been cloned. These are the TRPM8 (CMR1) cold and menthol-sensitive channel and the TRPA1 (ANKTM1) cold-sensitive and menthol-insensitive channel. TRPM8 is present in small DRG neurons,199,238,257 and has a threshold of 22° to 24° C in heterologous expression systems, although in native DRG neurons a channel with properties similar to TRPM8 activates around 30° C (see Reid and colleagues257,259), a threshold consistent with that of low-threshold cooling C

fibers in the human.40 The TRPA1 receptor has a colder threshold for activation (17° C) and is present in a separate population of small DRG neurons.297 DRG neurons that respond to cooling or cold in vivo include specific C-cooling or cold units, C-mechanocold nociceptors, some C-polymodal nociceptors, and several groups of LTMs including C-mechanoreceptors, D hair units, and some A-fiber LTM units (e.g., SA II units). The small sizes of DRG neurons that express TRPM8 or TRPA1 may be indicative of these channels being expressed in some of the above-mentioned C-fiber units, but deduction of the likely sensory receptive properties that such neurons might possess is premature, although these channels seem unlikely to account for cold responses in the larger LTM units. Overall the proportion of neurons activated by cold (see, e.g., Simone and Kajander288) exceeds the proportions expressing either TRPM8 (5% to 10%238) or TRPA1 (4%297). Thus these two channels seem unlikely to account for all cooling or cold transduction in DRG neurons. In addition to the above-noted thermal receptor molecules, inhibition of a background K conductance by cooling was shown to contribute to cold transduction in rat DRG neurons in vitro.258,328 One candidate is TREK1, a two-pore domain background K channel present in DRG neurons (as well as many CNS neurons), that is closed by cooling.193 In addition, the hyperpolarization-activated nonselective cation current Ih328 (see later) may be inhibited by cooling and this could possibly moderate (reduce) neuronal responses to cooling;

188

Function of the Peripheral Nervous System

interestingly this Ih is present in some medium- to largediameter DRG neurons.53

Mechanoreceptive Molecules Several receptors/channels have been implicated in mechanotransduction to date, but this area is less well understood than that of thermal transduction, and transduction mechanisms in different types of neurons are not fully established. There is a group of degenerin/epithelial Na channel (DEG/ENaC) voltage-insensitive sodium channels (mammalian counterparts are the acid-sensing ion channels [ASICs]) some of which (e.g., MEC4, MEC10, and the stomatin-related MEC2) have a role in mechanotransduction in Caenorhabditis elegans. A model of a mechanically gated ion channel that is linked to the cytoskeleton via MEC2 as well as to the extracellular matrix has been proposed in which movement of the membrane causes the channel to open.196 The MEC2 homologue stomatin is widely expressed in DRG neurons, consistent with a possible role in mechanotransduction in many neurons.196 One of the DEG/ENaC family known variously as BNC1, ASIC2, MDEG, or BNaC1 is implicated in mechanotransduction in mammals and is expressed more strongly in large than small DRG neurons: transgenic BNC1 / mice showed reduced sensitivity especially of RA LTM units, consistent with the localization of this protein in lanceolate endings of hair follicle afferents in skin.252 Another of the DEG/ENaC channel group, DRASIC/ASIC3, is thought to be expressed only in primary afferent neurons, and responds to low pH; DRASIC knockout mice show defects in behavioral responses to noxious mechanical stimuli, with decreased responses in mechanical nociceptors and increased responses in LTMs and decreased responses of nociceptors to acid and noxious heat,253 raising the possibility that it may contribute to mechanotransduction mechanisms. Purinergic receptors may also play a role in mechanical transduction. ATP, released through low-level mechanical distortion, can activate heterologously expressed P2Y1 receptors in oocytes, whereas a greater distortion may release enough ATP to activate heterologously expressed P2X3 receptors. P2Y1 receptors are present on large DRG neurons and P2X3 receptors are present on small DRG neurons220; it was therefore suggested that these receptors may contribute to mechanotransduction in LTM neurons (P2Y1) or in nociceptors (P2X3).220 The voltage-gated calcium channel Cav3.2 and the associated T-type calcium current have been suggested to play a role in mechanotransduction in D hair units, because their expression is reduced in mice that are D hair deficient as a result of knocking out the NT-4 gene.284 TRPV4 may have a role in mechanical transduction: there are TRPV4-positive fibers in the skin, and TRPV4

knockout mice show impairment of responses to noxious mechanical stimuli.306 The description of TRPV4 expression in about 10% of DRG neurons, mainly in small, NFrich neurons,307 is consistent with expression in some A-fiber mechanical nociceptors. In addition, BDNF has been implicated in the regulation of mechanosensitivity of low-threshold, SA mechanoreceptors.47

Chemoreceptive Molecules Although it is clear what is meant by chemoreceptor in the context of olfaction and taste, it is much less clear what the limitations of this term should be in the context of the somatosensory system. It is likely that there are fibers that detect, for example, glucose, oxygen, CO2, and other metabolite levels in the tissues, and thus act as homeostatic afferents. In addition, it is known that certain afferent fibers can respond to inflammatory mediators, chemicals released by damaged tissues (such as ATP and bradykinin), and low pH. Information from such afferents could contribute to the detection of inflammatory processes and monitoring of the state of tissue integrity. Many C-fiber nociceptors also respond to capsaicin by activating the capsaicin/noxious heat receptor TRPV1. P2X purinergic receptors respond to ATP,36 and several are expressed in DRG neurons. P2X3 is found only in sensory neurons, where it can form heteromultimers (P2X2/3) with P2X2. Homomeric P2X3 receptors are expressed more highly in smaller diameter DRG neurons, whereas P2X2/3 heteromeric receptors are found more in medium-sized neurons.36 Heterologously expressed P2X3 responds to levels of ATP released by high-intensity (noxious) mechanical distortion, and thus (as mentioned earlier) may play a role in transduction of noxious mechanical stimuli.220 P2Y1 purinergic receptors are expressed in NF-rich large-diameter DRG neurons.220 As mentioned earlier, when expressed in chick oocytes they may respond to ATP released as a result of mild mechanical distortion, and thus may modulate responses to low-intensity mechanical stimulation in large DRG neurons.220

Acid-Sensing Molecules Acid pH in the tissues can be detected by primary afferent neurons resulting in pain. Acid pH may result from lactic acid accumulation in muscle, or from damage to the tissues. Importantly, TRPV1 is activated by low pH.320 The ASICs (also known as DEG/ENaC voltage-insensitive sodium channels, mentioned earlier under Mechanoreceptive Molecules) are activated by extracellular protons. ASIC/ASIC1, BNC1/ASIC2, and DRASIC/ASIC3 are all expressed in DRG neurons—both large and small,6,252,253 but BNC1/ASIC2 is more strongly expressed in large neurons; heteromultimers between these channels give rise to different H-gated channels.21 Other channels

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons

may also contribute to transduction of low pH/high proton levels. TREK1 is one such channel. TREK1 is a polymodal background K channel that is closed by stretch, PGE2, cyclic AMP, and intracellular acidosis and is activated by heat.193 For further information on the molecular identities of proton-sensitive channels, see Wood.335

ION CURRENTS AND CHANNELS Na⫹ Currents and Channels Voltage-gated sodium channels are essential for generation and conduction of action potentials. The currents are subdivided into TTX-sensitive (TTXS) and TTX-resistant (TTXR) groups. The  subunits Nav1.1, 1.2, 1.3, 1.6, and 1.7 are TTXS, and Nav1.5, 1.8, and 1.9 are TTXR. Three of the  subunit proteins, Nav1.8 and Nav1.9 (TTXR), and Nav1.7 (TTXS), are expressed more in small than large DRG neurons. Nav1.8 (SNS/PN3). In rats, Nav1.8 is expressed in small- to medium-sized DRG neurons,4 most of which are nociceptive.74 Nav1.8 gives rise to a TTXR Na current that is thought to contribute the majority of the inward Na current in action potentials of small DRG neurons and is thought to be responsible for the TTXR current at nociceptive receptor terminals.5,31 Its properties probably contribute to properties of nociceptive neurons, including its high activation threshold (about 35 mV), its slow kinetics giving rise to long-duration action potentials, its contribution to the large action potential overshoot, and its rapid repriming, which may enable firing even in depolarized fibers.4,5,74,261 Inflammation and Nerve Injury. Several inflammatory mediators can acutely (minutes to hours) decrease the activation threshold, and/or increase the kinetics or magnitude of the TTXR Na current (presumed Nav1.8 related)102 and may contribute to acute hypersensitivity at nerve terminals. In the longer term, Nav1.8 mRNA and TTXR Na current in small DRG neurons and in cutaneous fibers are upregulated during most studies of inflammation.229,312 In some models of neuropathic pain, TTXR current density and Nav1.8 mRNA and protein (see, e.g., Sleeper et al.289) are decreased. Nav1.9 (NaN/SNS2). Nav1.9 also gives rise to a TTXR current. Nav1.9-LI is exclusively in small DRG neurons,69,313 preferentially in those with IB4 binding87 and along C fibers and at nodes of Ranvier of thinly myelinated fibers.88 It gives rise to a persistent, depolarizing TTXR Na current in small DRG neurons as a result of its low activation threshold and ultra-slow inactivation,62 thus it is likely to influence membrane excitability. GDNF, but not NGF, upregulates Nav1.9 mRNA.87 Inflammation and Nerve Injury. Nav1.9 mRNA in DRG neurons is increased after inflammation,313 and after exogenous GDNF administration.87 Axotomy induces a decrease

189

in Nav1.9 mRNA and protein in the DRG87,289,313 that is reversed in IB4-positive neurons by exogenous GDNF.87 Nav1.7 (PN1). Nav1.7 is more highly expressed in small than large DRG neurons76 despite the mRNA being in neurons of all sizes.25,76 It is thought to carry much of the TTXS inward current in action potentials in small neurons.63 Nav1.7 protein is in fibers and terminals of cultured DRG neurons.319 Its slow inactivation, combined with its low activation threshold,63 may be important in generation of receptor potentials and contributing to generation of action potentials. NGF causes a long-lasting (weeks) increase in Nav1.7 protein in DRG neurons in vivo.107 Other ␣ Subunits. mRNAs of several other TTXS  subunits are more abundant in medium and large neurons. These include Nav1.1, 1.2, and 1.6 and Nav2.2 (NaG).25,269 Following axotomy, the appearance of a more rapidly repriming TTXS current thought to contribute to hyperexcitability in axotomized small DRG neurons in vitro24,64 has been ascribed to increased expression of Nav1.3 (also known as III or brain type III), but roles for other TTXS channels have not been ruled out. Afferent/Sensory Receptive Properties of Neurons Expressing Nav1.7, Nav1.8, and Nav1.9. After intracellular recording, identification of sensory properties, dye injection, and immunocytochemistry, Nav1.9 was found to be expressed in nociceptive but not LTM neurons, and Nav1.8 was expressed mainly in nociceptive neurons, although some LTMs were also weakly positive.74 Despite being expressed mainly in small- to medium-sized neurons, Nav1.7 was expressed in both nociceptive and LTM neurons, with a slightly higher mean intensity in nociceptive neurons; its expression was negatively correlated with conduction velocity.76 Naⴙ Channel  Subunits. Na channel  subunits may interact with cytoskeleton or extracellular matrix and play roles in Na channel trafficking within cells and insertion into membrane. When coexpressed with  subunits,  subunits can alter the kinetics, peak current, and/or voltage dependencies of  subunits.131 1 (Na1.1) mRNA is found in medium to large DRG neurons, and a splice variant (1A) is also in DRG neurons.131 2 (Na2.1) mRNA is found in medium to large DRG neurons.25 3 (Na3) mRNA is normally found in small DRG neurons and increases in these neurons in the CCI model of neuropathic pain.279

K⫹ Currents and Channels K channels are central to the control of membrane potential, afterhyperpolarization, and firing frequency, and they influence adaptation. They tend to increase membrane

190

Function of the Peripheral Nervous System

potential stability, and channels that contribute to longduration afterhyperpolarizations may prove to be more highly expressed in nociceptors. Voltage-Gated K⫹ Currents and Kv Channels The two main groups of calcium-insensitive voltage-gated K currents are the depolarization-activated delayed rectifier (IKv) and fast transient (IA) currents. The protein  subunits of the channels that underlie these currents are the Kv subunits. Delayed Rectifier Currents. Delayed rectifier currents (IDR; also called IKv) serve to terminate the action potential rapidly in the soma and inhibit repetitive firing in myelinated axons.9 They are especially prominent in some large cutaneous afferent neurons,84 but are also present in small DRG neurons.103 Information on dendrodotoxin sensitivity is presented in the next section. Fast Transient (A-type) K ⴙ Currents. Fast transient K (IA) currents tend to clamp resting potential at hyperpolarized voltages until they inactivate, thus prolonging the afterhyperpolarization and slowing repetitive discharges.60 IA currents are present in both large cutaneous afferents84 and small DRG neurons103 but are more prominent in slowly conducting afferents.329 IA can be subdivided into rapidly (fast IA) and slowly (slow IA) inactivating types.345 Slow IA is particularly prominent in small DRG neurons with TTXR action potentials (see Yang et al.345), and may therefore contribute to the broad afterhyperpolarizations of nociceptive neurons. In DRG neurons IKv and slow IA are sensitive to dendrotoxin, but fast IA is not.345 Kv1.1 and 1.2 are associated with the delayed rectifier, whereas Kv1.4 is associated with fast IA,255,293 and it has been suggested that Kv1.1/1.2 associated with Kv1.4 (dendrotoxin insensitive) may give rise to the slow IA in DRG neurons.255,345 Kv1.1, 1.2, 1.3, 1.4, 1.5, and 1.6 and Kv2.1 are all expressed in DRG neurons; of these, Kv1.1 and 1.2 are more abundant and more highly expressed in medium- to large-sized neurons, and Kv2.1 is also more highly expressed in medium to large neurons, whereas Kv1.4 is more highly expressed in small neurons.130,255,345 M Currents. M currents are voltage and time dependent, noninactivating, and activated at negative voltages (beginning at approximately 70 mV). Their inhibition by acetylcholine or other agents leads to increased neuronal excitability.32 M currents are associated with KCNQ2/3 and -5 subunits,236,264 members of the 6TM superfamily and are distantly related to Kv channels. M currents as well as KCNQ2, -3, and -5 have been detected in both small and large DRG neurons.236 Block of M current with linopirdine causes increased firing in response to current injection in small DRG neurons,236 indicating that the M current may normally act as a brake to firing in these neurons.

Background K ⴙ Channels. The greater part of the timeindependent resting conductance in a variety of neurons is contributed by background K channels. They are twopore domain, homo- or heterodimeric channels. Some can be inhibited by neurotransmitters212,311 or physical stimuli such as extracellular acidity (the TASK family, which may be involved in acid sensing by nociceptors256); activated by alkalinity (the TALK family), by mechanical stress (TRAAK, the TREK family), or by heat (TREK-1); or be targets for volatile anesthetics (TASK-1, TASK-2, TREK-1, THIK-1).237 Background K currents in DRG neurons play a role in cold transduction (see earlier).258,328 Some members of this family are highly expressed in DRGs in preference to brain (TWIK-2, TASK-1, TASK-2, TREK-1, and TRAAK),203 but at the time of writing nothing is known about expression of these channels in sensory neurons with different modalities, nor whether their expression changes after nerve injury or in inflammation in mammals. Inwardly Rectifying K ⴙ Channels. Inwardly rectifying K channel (KIR) current contributes to the resting membrane potential in a number of cell types, including DRG neurons, where KIR current is mostly in medium-size neurons.278 Little is known of KIR-related channel subunits in DRGs. Ca2ⴙ-Activated K ⴙ Currents. Ca2-activated K currents (IKCa) are of three types, related to BK (slow), IK, and SK channels (big, intermediate, and small conductance channels, respectively), all activated by raised intracellular Ca2; BK is also voltage dependent. Functionally BK is associated with afterhyperpolarizations that develop rapidly and decay in 10 to 100 ms, and SK with slower afterhyperpolarizations that may last for seconds and limit firing frequency (spike frequency adaptation).158,325 Either or both of these could therefore contribute to the long afterhyperpolarization in nociceptors, although this remains to be established. BK and SK currents are found in DRG neurons.3,104,275 BK currents were in two thirds of small DRG neurons. Immunoreactivity for SK1 and IK1 channels was found in many human and rat DRG neurons.26 A related channel, SLACK (sequence like a Ca2-activated K), is expressed in rat DRG neurons.22 SLACK has a conductance similar to that of BK, with which it can interact to generate an intermediate-conductance KCa channel (different from IK1). Its role in DRG neurons is not yet understood. Axotomy Changes in DRG neurons after axotomy suggest decreased K channel expression or activation. Decreases in IKIR and fast IA occur after axotomy in large cutaneous afferent neurons83 and in IKIR in small neurons.345 Reductions in expression of Kv1.1, 1.2, and 1.4 and Kv2.1 have been reported.130,255,345 Reverse transcription– polymerase chain reaction (RTPCR) showed increased KCNQ2, -3, and -5 in both large

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons

and small DRG neurons, and enhancement of M currents with retigabine resulted in decreased electrophysiologic and behavioral changes in a model of neuropathic pain.236 Inflammation After acute (2- to 3-hour) inflammation, activation of KCNQ/M currents with retigabine resulted in animals putting increased weight on the inflamed foot.236

Hyperpolarization-Activated Currents and Channels The H current (Ih) is a hyperpolarization-activated, timeand voltage-dependent, nonselective cation current that when activated causes depolarization of the membrane, tending to reduce afterhyperpolarization duration, increase firing frequency, and decrease adaptation.233 An H current is prominent in most or all large DRG neurons but in fewer small neurons.278 The channels that give rise to Ih are made up of HCN protein subunits, four isoforms (HCN1 through -4) of which have been cloned (see Chaplan et al.53). In DRG neurons there is strong expression of HCN1 mRNA in all large- to medium-diameter and most small-diameter DRG neurons, lower expression of HCN2 mRNA in approximately 80% of large and approximately 60% of small neurons, and low or undetectable levels of HCN3 and -4. HCN1 through -3 proteins are concentrated at the membrane especially of large neurons.53 Nerve injury causes increased Ih in large-diameter neurons dissociated in vitro, and ZD 7288 (a specific Ih blocker) blocks ectopic discharge in axotomized A-afferent fibers.53

Ca2ⴙ Currents and Channels The inflection on the falling phase of some of the broader action potentials in DRG neurons is partly due to inward Ca2 current. Ca2 has roles as a second messenger, in transmitter release, and in inhibiting firing by activation of IKCa. Several voltage-gated Ca2 currents have been described in DRG neurons, including L (high voltage activated), T (low voltage activated), and N (intermediate properties)91 (for more on properties of these currents, see review by Catterall51). Their amplitudes differ in neurons of different sizes, with relatively large L-type and N-type and smaller T-type currents in small cells, larger T- but little L- and N-type currents in medium-sized neurons, and little T-type current in large neurons.277 T-type Ca2 channels (Cav3.2) were suggested to be necessary for the normal mechanosensitivity of A-fiber D hair LTMs because expression of this channel was reduced in NT-4 knockout mice that were devoid of D hairs.284 Additionally, L- and N- but not T-type currents cause substance P release from isolated DRG neurons.119 Ca2 currents can be modulated by a variety of agonists. For instance activation of DORs on cultured young rat DRG neurons reduced N-, L-, P-, and Q- but not R-type currents,2 and 5-HT inhibits Ca2

191

currents in small DRG neurons probably via 5-HT1A receptors.67 Limited reports of expression include the following channel subunits demonstrated both immunocytochemically and by in situ hybridization (current type associated with the subunit in parentheses): Cav2.1 (P/Q), Cav2.2 (N), Cav1.2 and Cav1.3 (L), and Cav2.3 (R).51,218,348 NERVE INJURY. The T-type current in medium-sized neurons, as well as total Ca2 currents, were decreased 10 days after CCI of the sciatic nerve.198 In addition, the 2-1 subunit is upregulated (mRNA and protein)191,221 after different types of nerve injury. This may be an important site of action of the analgesic gabapentin.

CHEMICAL PHENOTYPE OF NOCICEPTIVE AND LTM NEURONS Although there is much indirect evidence to connect a number of molecular types to neurons with particular sensory properties, only for a few molecules has this relationship been directly demonstrated, by intracellular recording and determination of sensory receptive properties in skin and subcutaneous tissues followed by intracellular dye injection that enables subsequent immunocytochemistry. To summarize, in rat or guinea pig DRGs, substance P170 and Nav1.985 were found only in nociceptive neurons. CGRP124,169 and Nav1.874 showed intense labeling only in nociceptors, with weak labeling in some cutaneous LTM neurons. Strong TrkA and IB486 immunoreactivity was seen only in nociceptive neurons, and weak labeling of both was seen in a few D hair units.86 C-fiber and A-fiber nociceptors were both labeled for TrkA and IB4. Only C-fiber nociceptors showed IB4 binding,86 although in another study99 3 of 10 A-fiber nociceptors were also labeled by IB4.

SUMMARY It is beginning to be possible to interpret some aspects of the chemical phenotype of primary afferent neurons in terms of their afferent/sensory functions. However, this is a complex field, and the dynamically changing expression of a multitude of factors both by the DRG neurons themselves and by other cells of the tissues in which they reside makes this a complex and fascinating field. A large variety and number of metabolites, cytokines, inflammatory mediators, trophic factors, neuromodulators and transmitters may be released into the close environment of these neurons. That these endogenous chemicals may profoundly influence the properties and expression patterns of these neurons, highlights the advantages of studying them in vivo, both normally and, perhaps more particularly, in their dynamically changing environment following nerve or tissue damage.

192

Function of the Peripheral Nervous System

REFERENCES 1. Acosta, C. G., Fabrega, A. R., Masco, D. H., and Lopez, H. S.: A sensory neuron subpopulation with unique sequential survival dependence on nerve growth factor and basic fibroblast growth factor during development. J. Neurosci. 21:8873, 2001. 2. Acosta, C. G., and Lopez, H. S.: Delta opioid receptor modulation of several voltage-dependent Ca(2) currents in rat sensory neurons. J. Neurosci. 19:8337, 1999. 3. Akins, P. T., and McCleskey, E. W.: Characterization of potassium currents in adult rat sensory neurons and modulation by opioids and cyclic AMP. Neuroscience 56:759, 1993. 4. Akopian, A. N., Sivilotti, L., and Wood, J. N.: A tetrodotoxin-resistant voltage-gated sodium channel expressed by sensory neurons. Nature 379:257, 1996. 5. Akopian A. N., Souslova, V., England, S., et al.: The tetrodotoxin-resistant sodium channel SNS has a specialized function in pain pathways. Nat. Neurosci. 2:541, 1999. 6. Alvarez de la Rosa, D., Zhang, P., Shao, D., et al.: Functional implications of the localization and activity of acid-sensitive channels in rat peripheral nervous system. Proc. Natl. Acad. Sci. U. S. A. 99:2326, 2002. 7. Anand, P.: Neurotrophic factors and their receptors in human sensory neuropathies. Prog. Brain Res. 146:477, 2004. 8. Averill, S., McMahon, S. B., Clary, D. O., et al.: Immunocytochemical localization of trkA receptors in chemically identified subgroups of adult rat sensory neurons. Eur. J. Neurosci. 7:1484, 1995. 9. Baker, M., Bostock, H., Grafe, P., and Martius, P.: Function and distribution of three types of rectifying channel in rat spinal root myelinated axons. J. Physiol. (Lond.) 383:45, 1987. 10. Bandtlow, C. E., Heumann, R., Schwab, M. E., and Thoenen, H.: Cellular localization of nerve growth factor synthesis by in situ hybridization. EMBO J. 6:891, 1987. 11. Bar, K. J., Natura, G., Telleria-Diaz, A., et al.: Changes in the effect of spinal prostaglandin E2 during inflammation: prostaglandin E (EP1-EP4) receptors in spinal nociceptive processing of input from the normal or inflamed knee joint. J. Neurosci. 24:642, 2004. 12. Barbacid, M.: The Trk family of neurotrophin receptors. J. Neurobiol. 25:1386, 1994. 13. Baudet, C., Mikaels, A., Westphal, H., et al.: Positive and negative interactions of GDNF, NTN and ART in developing sensory neuron subpopulations, and their collaboration with neurotrophins. Development 127:4335, 2000. 14. Beedle, A. M., McRory, J. E., Poirot, O., et al.: Agonistindependent modulation of N-type calcium channels by ORL1 receptors. Nat. Neurosci. 7:118, 2004. 15. Bennett, D. L., Averill, S., Clary, D. O., et al.: Postnatal changes in the expression of the trkA high-affinity NGF receptor in primary sensory neurons. Eur. J. Neurosci. 8:2204, 1996. 16. Bennett, D. L., Boucher, T. J., Armanini, M. P., et al.: The glial cell line-derived neurotrophic factor family receptor components are differentially regulated within sensory neurons after nerve injury. J. Neurosci. 20:427, 2000. 17. Bennett, D. L., Dmietrieva, N., Priestley, J. V., et al.: trkA, CGRP and IB4 expression in retrogradely labelled cuta-

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32. 33.

neous and visceral primary sensory neurons in the rat. Neurosci. Lett. 206:33, 1996. Bennett, D. L., Michael, G. J., Ramachandran, N., et al.: A distinct subgroup of small DRG cells express GDNF receptor components and GDNF is protective for these neurons after nerve injury. J. Neurosci. 18:3059, 1998. Bennett, G., al Rashed, S., Hoult, J. R., and Brain, S. D.: Nerve growth factor induced hyperalgesia in the rat hind paw is dependent on circulating neutrophils. Pain 77:315, 1998. Bennett, G. J., and Xie, Y. K.: A peripheral mononeuropathy in rat that produces disorders of pain sensation like those seen in man. Pain 33:87, 1988. Benson, C. J., Xie, J., Wemmie, J. A., et al.: Heteromultimers of DEG/ENaC subunits form H-gated channels in mouse sensory neurons. Proc. Natl. Acad. Sci. U. S. A. 99:2338, 2002. Bhattacharjee, A., Gan, L., and Kaczmarek, L. K.: Localization of the Slack potassium channel in the rat central nervous system. J. Comp. Neurol. 454:241, 2002. Birder, L. A., and Perl, E. R.: Expression of alpha2adrenergic receptors in rat primary afferent neurons after peripheral nerve injury or inflammation. J. Physiol.(Lond.) 515(Pt. 2):533, 1999. Black, J. A., Cummins, T. R., Plumpton, C., et al.: Upregulation of a silent sodium channel after peripheral, but not central, nerve injury in DRG neurons. J. Neurophysiol. 82:2776, 1999. Black, J. A., Dib-Hajj, S., McNabola, K., et al.: Spinal sensory neurons express multiple sodium channel alpha-subunit mRNAs. Brain Res. Mol. Brain Res. 43:117, 1996. Boettger, M. K., Till, S., Chen, M. X., et al.: Calcium-activated potassium channel SK1- and IK1-like immunoreactivity in injured human sensory neurons and its regulation by neurotrophic factors. Brain 125:252, 2002. Botchkarev, V. A., Metz, M., Botchkareva, N. V., et al.: Brain-derived neurotrophic factor, neurotrophin-3, and neurotrophin-4 act as “epitheliotrophins” in murine skin. Lab. Invest. 79:557, 1999. Brain, S. D.: Sensory neuropeptides: their role in inflammation and wound healing. Immunopharmacology 37:133, 1997. Brechenmacher, C., Larmet, Y., Feltz, P., and Rodeau, J. L.: Cultured rat sensory neurons express functional tachykinin receptor subtypes 1, 2 and 3. Neurosci. Lett. 241:159, 1998. Bridges, D., Rice, A. S., Egertova, M., et al.: Localisation of cannabinoid receptor 1 in rat dorsal root ganglion using in situ hybridisation and immunohistochemistry. Neuroscience 119:803, 2003. Brock, J. A., McLachlan, E. M., and Belmonte, C.: Tetrodotoxin-resistant impulses in single nociceptor nerve terminals in guinea-pig cornea. J Physiol (Lond) 512:211, 1998. Brown, D. A.: M currents. In Narahashi, T. (ed.): Ion Channels. New York, Plenum Press, p. 55, 1988. Burgess, P. R., and Perl, E. R.: Myelinated afferent fibres responding specifically to noxious stimulation of the skin. J. Physiol. (Lond.) 190:541, 1967.

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons 34. Burgess, P. R., and Perl, E. R.: Cutaneous mechanoreceptors and nociceptors. In Iggo, A. (ed.): Handbook of Sensory Physiology, Vol. 2. Somatosensory System. Berlin, SpringerVerlag, p. 29, 1973. 35. Burgess, P. R., Petit, D., and Warren, R. M.: Receptor types in cat hairy skin supplied by myelinated fibers. J. Neurophysiol. 31:833, 1968. 36. Burnstock, G.: P2X receptors in sensory neurons. Br. J. Anaesth. 84:476, 2000. 37. Cafferty, W. B., Gardiner, N. J., Gavazzi, I., et al.: Leukemia inhibitory factor determines the growth status of injured adult sensory neurons. J. Neurosci. 21:7161, 2001. 38. Calza, L., Pozza, M., Arletti, R., et al.: Long-lasting regulation of galanin, opioid, and other peptides in dorsal root ganglia and spinal cord during experimental polyarthritis. Exp. Neurol. 164:333, 2000. 39. Campenot, R. B., and MacInnis, B. L.: Retrograde transport of neurotrophins: fact and function. J. Neurobiol. 58:217, 2004. 40. Campero, M., Serra, J., Bostock, H., and Ochoa, J. L.: Slowly conducting afferents activated by innocuous low temperature in human skin. J. Physiol. (Lond.) 535:855, 2001. 41. Cardenas, L. M., Cardenas, C. G., and Scroggs, R. S.: 5HT increases excitability of nociceptor-like rat dorsal root ganglion neurons via cAMP-coupled TTX-resistant Na() channels. J. Neurophysiol. 86:241, 2001. 42. Carlton, S. M.: Peripheral excitatory amino acids. Curr. Opin. Pharmacol. 1:52–56, 2001. 43. Carlton, S. M., and Coggeshall, R. E.: Inflammation-induced up-regulation of neurokinin 1 receptors in rat glabrous skin. Neurosci. Lett. 326:29, 2002. 44. Carlton, S. M., Du, J., Davidson, E., et al.: Somatostatin receptors on peripheral primary afferent terminals: inhibition of sensitized nociceptors. Pain 90:233, 2001. 45. Carlton, S. M., Zhou, S., and Coggeshall, R. E.: Localization and activation of substance P receptors in unmyelinated axons of rat glabrous skin. Brain Res. 734:103, 1996. 46. Carr, P. A., and Nagy, J. I.: Emerging relationships between cytochemical properties and sensory modality transmission in primary sensory neurons. Brain Res. Bull. 30:209, 1993. 47. Carroll, P., Lewin, G. R., Koltzenburg, M., et al.: A role for BDNF in mechanosensation. Nat. Neurosci. 1:42, 1998. 48. Casaccia-Bonnefil, P., Gu, C., Khursigara, G., and Chao, M. V.: P75 neurotrophin receptor as a modulator of survival and death decisions. Microsc. Res. Tech. 45:217, 1999. 49. Caterina, M. J., Rosen, T. A., Tominaga, M., et al.: A capsaicin-receptor homologue with a high threshold for noxious heat. Nature 398:436, 1999. 50. Caterina, M. J., Schumacher, M. A., Tominaga, M., et al.: The capsaicin receptor: a heat-activated ion channel in the pain pathway. Nature 389:816, 1997. 51. Catterall, W. A.: Structure and regulation of voltage-gated Ca2 channels. Annu. Rev. Cell Dev. Biol. 16:521, 2000. 52. Cervero, F.: Sensory innervation of the viscera: peripheral basis of visceral pain. Physiol. Rev. 74:95, 1994. 53. Chaplan, S. R., Guo, H. Q., Lee, D. H., et al.: Neuronal hyperpolarization-activated pacemaker channels drive neuropathic pain. J. Neurosci. 23:1169, 2003.

193

54. Chen, C., Zhou, X. F., and Rush, R. A.: Neurotrophin-3 and trkC-immunoreactive neurons in rat dorsal root ganglia correlate by distribution and morphology. Neurochem. Res. 21:809, 1996. 55. Cho, H. J., Kim, J. K., Park, H. C., et al.: Changes in brainderived neurotrophic factor immunoreactivity in rat dorsal root ganglia, spinal cord, and gracile nuclei following cut or crush injuries. Exp. Neurol. 154:224, 1998. 56. Cho, H. J., Kim, J. K., Zhou, X. F., and Rush, R. A.: Increased brain-derived neurotrophic factor immunoreactivity in rat dorsal root ganglia and spinal cord following peripheral inflammation. Brain Res. 764:269, 1997. 57. Cho, H. J., Kim, S. Y., Park, M. J., et al.: Expression of mRNA for brain-derived neurotrophic factor in the dorsal root ganglion following peripheral inflammation. Brain Res. 749:358, 1997. 58. Chopra, B., Giblett, S., Little, J. G., et al.: Cyclooxygenase-1 is a marker for a subpopulation of putative nociceptive neurons in rat dorsal root ganglia. Eur. J. Neurosci. 12:911, 2000. 59. Coggeshall, R. E., Zhou, S., and Carlton, S. M.: Opioid receptors on peripheral sensory axons. Brain Res. 764:126, 1997. 60. Connor, J. A., and Stevens, C. F.: Voltage clamp studies of a transient outward membrane current in gastropod neural somata. J. Physiol. (Lond.) 213:21, 1971. 61. Copray, J. C., Mantingh, I., Brouwer, N., et al.: Expression of interleukin-1 beta in rat dorsal root ganglia. J. Neuroimmunol. 118:203, 2001. 62. Cummins, T. R., Dib-Hajj, S. D., Black, J. A., et al.: A novel persistent tetrodotoxin-resistant sodium current In SNS-null and wild-type small primary sensory neurons. J. Neurosci. 19:RC43, 1999. 63. Cummins, T. R., Howe, J. R., and Waxman, S. G.: Slow closed-state inactivation: a novel mechanism underlying ramp currents in cells expressing the hNE/PN1 sodium channel. J. Neurosci. 18:9607, 1998. 64. Cummins, T. R., and Waxman, S. G.: Downregulation of tetrodotoxin-resistant sodium currents and upregulation of a rapidly repriming tetrodotoxin-sensitive sodium current in small spinal sensory neurons after nerve injury. J. Neurosci. 17:3503, 1997. 65. Davis, K. D., and Pope, G. E.: Noxious cold evokes multiple sensations with distinct time courses. Pain 98:179, 2002. 66. De Biasi, S., and Rustioni, A.: Glutamate and substance P coexist in primary afferent terminals in the superficial laminae of spinal cord. Proc. Natl. Acad. Sci. U. S. A. 85:7820, 1988. 67. Del Mar, L. P., Cardenas, C. G., and Scroggs, R. S.: Serotonin inhibits high-threshold Ca2 channel currents in capsaicin-sensitive acutely isolated adult rat DRG neurons. J. Neurophysiol. 72:2551, 1994. 68. Diamond, J., Foerster, A., Holmes, M., and Coughlin, M.: Sensory nerves in adult rats regenerate and restore sensory function to the skin independently of endogenous NGF. J. Neurosci. 12:1467, 1992. 69. Dib-Hajj, S. D., Tyrrell, L., Black, J. A., and Waxman, S. G.: NaN, a novel voltage-gated Na channel, is expressed preferentially in peripheral sensory neurons and downregulated after axotomy. Proc. Natl. Acad. Sci. U. S. A. 95:8963, 1998.

194

Function of the Peripheral Nervous System

70. Dickinson, T., and Fleetwood-Walker, S. M.: VIP and PACAP: very important in pain? Trends Pharmacol. Sci. 20:324, 1999. 71. Djouhri, L., Bleazard, L., and Lawson, S. N.: Association of somatic action potential shape with sensory receptive properties in guinea pig dorsal root ganglion neurons. J. Physiol. (Lond.) 513:857, 1998. 72. Djouhri, L., Dawbarn, D., Robertson, A., et al.: Time course and nerve growth factor dependence of inflammationinduced alterations in electrophysiological membrane properties in nociceptive primary afferent neurons. J. Neurosci. 21:8722, 2001. 73. Djouhri, L., and Lawson, S. N.: Abeta-fiber nociceptive primary afferent neurons: a review of incidence and properties in relation to other afferent A-fiber neurons in mammals. Brain Res. Brain Res. Rev. 46:131, 2004. 74. Djouhri, L., Fang, X., Okuse, K., et al.: The TTX-resistant sodium channel Nav1.8 (SNS/PN3): expression and correlation with membrane properties in rat nociceptive primary afferent neurons. J. Physiol. (Lond.) 550:739, 2003. 75. Djouhri, L., and Lawson, S. N.: Differences in the size of the somatic action potential overshoot between nociceptive and non-nociceptive dorsal root ganglion neurons in the guinea-pig. Neuroscience 108:479, 2001. 76. Djouhri, L., Newton, R., Levinson, S. R., et al.: Sensory and electrophysiological properties of guinea-pig sensory neurons expressing Na(v)1.7 (PN1) Na() channel alpha subunit protein. J. Physiol. (Lond.) 546:565, 2003. 77. Dobretsov, M., Hastings, S. L., Sims, T. J., et al.: Stretch receptor-associated expression of alpha(3) isoform of the Na(),K()-ATPase in rat peripheral nervous system. Neuroscience 116:1069, 2003. 78. Dodd, J., and Jessell, T. M.: Lactoseries carbohydrates specify subsets of dorsal root ganglion neurons projecting to the superficial dorsal horn of rat spinal cord. J. Neurosci. 12:3278, 1985. 79. Dowsing, B. J., Romeo, R., and Morrison, W. A.: Expression of leukemia inhibitory factor in human nerve following injury. J. Neurotrauma 18:1279, 2001. 80. Duce, I., and Keen, P.: An ultrastructural classification of the neuronal cell bodies of the rat dorsal root ganglion using zinc iodide-osmium impregnation. Cell Tiss. Res. 185:263, 1977. 81. Duggan, A. W., Hope, P. J., and Lang, C. W.: Microinjection of neuropeptide Y into the superficial dorsal horn reduces stimulus-evoked release of immunoreactive substance P in the anaesthetized cat. Neuroscience 44:733, 1991. 82. Ernfors, P., Lee, K. F., Kucera, J., and Jaenisch, R.: Lack of neurotrophin-3 leads to deficiencies in the peripheral nervous system and loss of limb proprioceptive afferents. Cell 77:503, 1994. 83. Everill, B., and Kocsis, J. D.: Reduction in potassium currents in identified cutaneous afferent dorsal root ganglion neurons after axotomy. J. Neurophysiol. 82:700, 1999. 84. Everill, B., Rizzo, M. A., and Kocsis, J. D.: Morphologically identified cutaneous afferent DRG neurons express three different potassium currents in varying proportions. J. Neurophysiol. 79:1814, 1998. 85. Fang, X., Djouhri, L., Black, J. A., et al.: The presence and role of the tetrodotoxin-resistant sodium channel Na(v)1.9 (NaN) in nociceptive primary afferent neurons. J. Neurosci. 22:7425, 2002.

86. Fang, X., Djouhri, L., and Lawson, S. N.: TrkA expression and IB4 binding in functionally identified dorsal root ganglion (DRG) nociceptive neurons in rats in vivo. J. Physiol. (Lond.) 536P:36P, 2001. 87. Fjell, J., Cummins, T. R., Dib-Hajj, S. D., et al.: Differential role of GDNF and NGF in the maintenance of two TTXresistant sodium channels in adult DRG neurons. Brain Res. Mol. Brain Res. 67:267, 1999. 88. Fjell, J., Hjelmstrom, P., Hormuzdiar, W., et al.: Localization of the tetrodotoxin-resistant sodium channel NaN in nociceptors. Neuroreport 11:199, 2000. 89. Foster, E., Robertson, B., and Fried, K.: trkB-like immunoreactivity in rat dorsal root ganglia following sciatic nerve injury. Brain Res. 659:267, 1994. 90. Fox, A., Wotherspoon, G., McNair, K., et al.: Regulation and function of spinal and peripheral neuronal B1 bradykinin receptors in inflammatory mechanical hyperalgesia. Pain 104:683, 2003. 91. Fox, A. P., Nowycky, M. C., and Tsien, R. W.: Kinetic and pharmacological properties distinguishing three types of calcium currents in chick sensory neurons. J. Physiol. (Lond.) 394:149, 1987. 92. Fox, E. A., and Powley, T. L.: False-positive artifacts of tracer strategies distort autonomic connectivity maps. Brain Res. Rev. 14:53, 1989. 93. Fukuoka, T., Kondo, E., Dai, Y., et al.: Brain-derived neurotrophic factor increases in the uninjured dorsal root ganglion neurons in selective spinal nerve ligation model. J. Neurosci. 21:4891, 2001. 94. Fundin, B. T., Silos-Santiago, I., Ernfors, P., et al.: Differential dependency of cutaneous mechanoreceptors on neurotrophins, trk receptors, and P75 LNGFR. Dev. Biol. 190:94, 1997. 95. Gardell, L. R., Wang, R., Ehrenfels, C., et al.: Multiple actions of systemic artemin in experimental neuropathy. Nat. Med. 9:1383, 2003. 96. Gardiner, N. J., Cafferty, W. B., Slack, S. E., and Thompson, S. W.: Expression of gp130 and leukaemia inhibitory factor receptor subunits in adult rat sensory neurons: regulation by nerve injury. J. Neurochem. 83:100, 2002. 97. Gee, M. D., Lynn, B., Basile, S., et al.: The relationship between axonal spike shape and functional modality in cutaneous C-fibres in the pig and rat. Neuroscience 90:509, 1999. 98. Genzen, J. R., Van Cleve, W., and McGehee, D. S.: Dorsal root ganglion neurons express multiple nicotinic acetylcholine receptor subtypes. J. Neurophysiol. 86:1773, 2001. 99. Gerke, M. B., and Plenderleith, M. B.: Binding sites for the plant lectin Bandeiraea simplicifolia I-isolectin B(4) are expressed by nociceptive primary sensory neurons. Brain Res. 911:101, 2001. 100. Gerke, M. B., and Plenderleith, M. B.: Analysis of the distribution of binding sites for the plant lectin Bandeiraea simplicifolia I-isolectin B4 on primary sensory neurons in seven mammalian species. Anat. Rec. 268:105, 2002. 101. Gokin, A. P., Fareed, M. U., Pan, H. L., et al.: Local injection of endothelin-1 produces pain-like behavior and excitation of nociceptors in rats. J. Neurosci. 21:5358, 2001. 102. Gold, M. S., Reichling, D. B., Shuster, M. J., and Levine, J. D.: Hyperalgesic agents increase a tetrodotoxin-resistant

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons

103.

104.

105a.

106.

107.

108.

109.

110.

111.

112.

113.

114.

115.

116. 117.

118.

119.

Na current in nociceptors. Proc. Natl. Acad. Sci. U. S. A. 93:1108, 1996. Gold, M. S., Shuster, M. J., and Levine, J. D.: Characterization of six voltage-gated K currents in adult rat sensory neurons. J. Neurophysiol. 75:2629, 1996. Gold, M. S., Shuster, M. J., and Levine, J. D.: Role of a Ca(2)-dependent slow afterhyperpolarization in prostaglandin E2-induced sensitization of cultured rat sensory neurons. Neurosci. Lett. 205:161, 1996. Golden, J. P., DeMaro, J. A., Osborne, P. A., et al.: Expression of neurturin, GDNF, and GDNF family-receptor mRNA in the developing and mature mouse. Exp. Neurol. 158:504, 1999. Görke, K., and Pierau, F. K.: Spike potentials and membrane properties of dorsal root ganglion cells in pigeons. Pflügers Arch. Eur. J. Physiol. 386:21, 1980. Gould, H. J. III, Gould, T. N., England, J. D., et al.: A possible role for nerve growth factor in the augmentation of sodium channels in models of chronic pain. Brain Res. 854:19, 2000. Gould, H. J. III, Gould, T. N., Paul, D., et al.: Development of inflammatory hypersensitivity and augmentation of sodium channels in rat dorsal root ganglia. Brain Res. 824:296, 1999. Govrin-Lippmann, R., and Devor, M.: Ongoing activity in severed nerves: source and variation with time. Brain Res. 159:406, 1978. Gracely, R. H., Lynch, S. A., and Bennett, G. J.: Painful neuropathy: altered central processing maintained dynamically by peripheral input. Pain 51:175, 1992. Gratto, K. A., and Verge, V. M.: Neurotrophin-3 downregulates trkA mRNA, NGF high-affinity binding sites, and associated phenotype in adult DRG neurons. Eur. J. Neurosci. 18:1535, 2003. Gray, J. A. B., and Matthew, P. B. C.: A comparison of the adaptation of the pacinian corpuscle with the accommodation of its own axon. J. Physiol. (Lond.) 114:454, 1951. Grewe, M., Vogelsang, K., Ruzicka, T., et al.: Neurotrophin-4 production by human epidermal keratinocytes: increased expression in atopic dermatitis. J. Invest. Dermatol. 114:1108, 2000. Grothe, C., and Nikkhah, G.: The role of basic fibroblast growth factor in peripheral nerve regeneration. Anat. Embryol. (Berl.) 204:171, 2001. Ha, S. O., Kim, J. K., Hong, H. S., et al.: Expression of brain-derived neurotrophic factor in rat dorsal root ganglia, spinal cord and gracile nuclei in experimental models of neuropathic pain. Neuroscience 107:301, 2001. Halata, Z.: Sensory innervation of the hairy skin: light- and electronmicroscopic study. J. Invest. Dermatol. 101:75S, 1993. Halata, Z., Grim, M., and Bauman, K. I.: Friedrich Sigmund Merkel and his “Merkel cell,” morphology, development, and physiology: review and new results. Anat. Rec. 271A:225, 2003. Hamburger, V., and Levi-Montalcini, R.: Proliferation, differentiation and degeneration in the spinal ganglia of the chick embryo under normal and experimental conditions. J. Exp. Zool. 111:457, 1949. Harding, L. M., Beadle, D. J., and Bermudez, I.: Voltagedependent calcium channel subtypes controlling somatic

120.

121.

122.

123.

124.

125.

126. 127.

128.

129.

130.

131.

132.

133.

134.

135.

136.

195

substance P release in the peripheral nervous system. Prog. Neuropsychopharmacol. Biol. Psychiatry 23:1103, 1999. Harper, A. A.: Similarities between some properties of the soma and sensory receptors of primary afferent neurons. Exp. Physiol. 76:369, 1991. Harper, A. A., and Lawson, S. N.: Conduction velocity is related to morphological cell type in rat dorsal root ganglia. J. Physiol. (Lond.) 359:31, 1985. Harper, A. A., and Lawson, S. N.: Electrical properties of rat dorsal root ganglion neurons with different peripheral conduction velocities. J. Physiol. (Lond.) 359:47, 1985. Heumann, R., Lindholm, D., Bandtlow, C., et al.: Differential regulation of mRNA encoding nerve growth factor and its receptor in rat sciatic nerve during development, degeneration, and regeneration: role of macrophages. Proc. Natl. Acad. Sci. U. S. A. 84:8735, 1987. Hoheisel, U., Mense, S., and Scherotzke, R.: Calcitonin gene-related peptide-immunoreactivity in functionally identified primary afferent neurons in the rat. Anat. Embryol. (Berl.) 189:41, 1994. Hokfelt, T., Zhang, X., and Wiesenfeld-Hallin, Z.: Messenger plasticity in primary sensory neurons following axotomy and its functional implications. Trends Neurosci. 17:22, 1994. Holzer, P.: Neurogenic vasodilatation and plasma leakage in the skin. Gen. Pharmacol. 30:5, 1998. Horch, K. W., Tuckett, R. P., and Burgess, P. R.: A key to the classification of cutaneous mechanoreceptors. J. Invest. Dermatol. 69:75, 1977. Hu, P., and McLachlan, E. M.: Selective reactions of cutaneous and muscle afferent neurons to peripheral nerve transection in rats. J. Neurosci. 23:10559, 2003. Inoue, A., Ikoma, K., Morioka, N., et al.: Interleukin-1beta induces substance P release from primary afferent neurons through the cyclooxygenase-2 system. J. Neurochem. 73:2206, 1999. Ishikawa, K., Tanaka, M., Black, J. A., and Waxman, S. G.: Changes in expression of voltage-gated potassium channels in dorsal root ganglion neurons following axotomy. Muscle Nerve 22:502, 1999. Isom, L. L.: I. Cellular and molecular biology of sodium channel beta-subunits: therapeutic implications for pain? Am. J. Physiol. Gastrointest. Liver Physiol. 278:G349, 2000. Itoh, M., Takasaki, I., Andoh, T., et al.: Induction by carrageenan inflammation of prepronociceptin mRNA in VR1-immunoreactive neurons in rat dorsal root ganglia. Neurosci. Res. 40:227, 2001. Janig, W.: Neurobiology of visceral afferent neurons: neuroanatomy, functions, organ regulations and sensations. Biol. Psychol. 42:29, 1996. Ji, R. R., Zhang, Q., Law, P. Y., et al.: Expression of mu-, delta-, and kappa-opioid receptor-like immunoreactivities in rat dorsal root ganglia after carrageenan-induced inflammation. J. Neurosci. 15:8156, 1995. Ji, R. R., Zhang, Q., Zhang, X., et al.: Prominent expression of bFGF in dorsal root ganglia after axotomy. Eur. J. Neurosci. 7:2458, 1995. Josephson, A., Widenfalk, J., Trifunovski, A., et al.: GDNF and NGF family members and receptors in human fetal and adult spinal cord and dorsal root ganglia. J. Comp. Neurol. 440:204, 2001.

196

Function of the Peripheral Nervous System

137. Kajander, K. C., and Bennett, G. J.: Onset of a painful peripheral neuropathy in rat: a partial and differential deafferentation and spontaneous discharge in A beta and A delta primary afferent neurons. J. Neurophysiol. 68:734, 1992. 138. Kaplan, D. R., and Stephens, R. M.: Neurotrophin signal transduction by the Trk receptor. J. Neurobiol. 25:1404, 1994. 139. Karchewski, L. A., Gratto, K. A., Wetmore, C., and Verge, V. M.: Dynamic patterns of BDNF expression in injured sensory neurons: differential modulation by NGF and NT-3. Eur. J. Neurosci. 16:1449, 2002. 140. Karchewski, L. A., Kim, F. A., Johnston, J., et al: Anatomical evidence supporting the potential for modulation by multiple neurotrophins in the majority of adult lumbar sensory neurons. J. Comp. Neurol. 413:327, 1999. 141. Kashiba, H., Fukui, H., Morikawa, Y., and Senba, E.: Gene expression of histamine H1 receptor in guinea pig primary sensory neurons: a relationship between H1 receptor mRNA-expressing neurons and peptidergic neurons. Brain Res. Mol. Brain Res. 66:24, 1999. 142. Kashiba, H., Fukui, H., and Senba, E.: Histamine H1 receptor mRNA is expressed in capsaicin-insensitive sensory neurons with neuropeptide Y-immunoreactivity in guinea pigs. Brain Res. 901:85, 2001. 143. Kashiba, H., Noguchi, K., Ueda, Y., and Senba, E.: Coexpression of trk family members and low-affinity neurotrophin receptors in rat dorsal root ganglion neurons. Brain Res. Mol. Brain Res. 30:158, 1995. 144. Kashiba, H., and Senba, E.: Up- and down-regulation of BDNF mRNA in distinct subgroups of rat sensory neurons after axotomy. Neuroreport 10:3561, 1999. 145. Kashiba, H., Senba, E., Kawai, Y., et al.: Axonal blockade induces the expression of vasoactive intestinal polypeptide and galanin in rat dorsal root ganglion neurons. Brain Res. 577:19, 1992. 146. Kashiba, H., Uchida, Y., and Senba, E.: Difference in binding by isolectin B4 to trkA and c-ret mRNA-expressing neurons in rat sensory ganglia. Brain Res. Mol. Brain Res. 95:18, 2001. 147. Kashiba, H., Uchida, Y., and Senba, E.: Distribution and colocalization of NGF and GDNF family ligand receptor mRNAs in dorsal root and nodose ganglion neurons of adult rats. Brain Res. Mol. Brain Res. 110:52, 2003. 148. Kashiba, H., Ueda, Y., and Senba, E.: Coexpression of preprotachykinin-A, alpha-calcitonin gene-related peptide, somatostatin, and neurotrophin receptor family messenger RNAs in rat dorsal root ganglion neurons. Neuroscience 70:179, 1996. 149. Kashiba, H., Ueda, Y., and Senba, E.: Systemic capsaicin in the adult rat differentially affects gene expression for neuropeptides and neurotrophin receptors in primary sensory neurons. Neuroscience 76:299, 1997. 150. Kenins, P.: Responses of single nerve fibres to capsaicin applied to the skin. Neurosci. Lett. 29:83, 1982. 151. Khasabova, I. A., Harding-Rose, C., Simone, D. A., and Seybold, V. S.: Differential effects of CB1 and opioid agonists on two populations of adult rat dorsal root ganglion neurons. J. Neurosci. 24:1744, 2004. 152. Kidd, B. L., and Urban, L. A.: Mechanisms of inflammatory pain. Br. J. Anaesth. 87:3, 2001.

153. Kim, S. H., and Chung, J. M.: An experimental model for peripheral neuropathy produced by segmental spinal nerve ligation in the rat. Pain 50:355, 1992. 154. Kim, Y. I., Na, H. S., Kim, S. H., et al.: Cell type-specific changes of the membrane properties of peripherallyaxotomized dorsal root ganglion neurons in a rat model of neuropathic pain. Neuroscience 86:301, 1998. 155. Klein, R., Silos-Santiago, I., Smeyne, R. J., et al.: Disruption of the neurotrophin-3 receptor gene trkC eliminates Ia muscle afferents and results in abnormal movements. Nature 368:249, 1994. 156. Koerber, H. R., Druzinsky, R. E., and Mendell, L. M.: Properties of somata of spinal dorsal root ganglion cells differ according to peripheral receptor innervated. J. Neurophysiol. 60:1584, 1988. 157. Koerber, H. R., and Mendell, L. M.: Functional heterogeneity of dorsal root ganglion cells. In Scott, S. A. (ed.): Sensory Neurons: Diversity, Development and Plasticity. New York, Oxford University Press, p. 77, 1992. 158. Kohler, M., Hirschberg, B., Bond, C. T., et al.: Smallconductance, calcium-activated potassium channels from mammalian brain. Science 273:1709, 1996. 159. Koltzenburg, M., Bennett, D. L., Shelton, D. L., and McMahon, S. B.: Neutralization of endogenous NGF prevents the sensitization of nociceptors supplying inflamed skin. Eur. J. Neurosci. 11:1698, 1999. 160. Kopp, U. C., Cicha, M. Z., Smith, L. A., et al.: Cyclooxygenase-2 involved in stimulation of renal mechanosensitive neurons. Hypertension 35:373, 2000. 161. Kurek, J. B., Austin, L., Cheema S. S., et al.: Up-regulation of leukaemia inhibitory factor and interleukin-6 in transected sciatic nerve and muscle following denervation. Neuromuscul. Disord. 6:105, 1996. 162. Lang, E., Novak, A., Reeh, P. W., and Handwerker, H. O.: Chemosensitivity of fine afferents from rat skin in vitro. J. Neurophysiol. 63:887, 1990. 163. Lawman, M. J., Boyle, M. D., Gee, A. P., and Young, M.: Nerve growth factor accelerates the early cellular events associated with wound healing. Exp. Mol. Pathol. 43:274, 1985. 164. Lawson, S. N.: The postnatal development of large light and small dark neurons in mouse dorsal root ganglia: a statistical analysis of cell numbers and size. J. Neurocytol. 8:275, 1979. 165. Lawson, S. N.: Morphological and biochemical cell types of sensory neurons. In Scott, S. A. (ed.): Sensory Neurons: Diversity, Development and Plasticity. New York, Oxford University Press, p. 27, 1992. 166. Lawson, S. N.: Phenotype and function of somatic primary afferent nociceptive neurons with C-, Adelta or Aalpha/beta-fibres. J. Exp. Physiol. 87:239, 2002. 167. Lawson, S. N., and Biscoe, T. J.: Development of mouse dorsal root ganglia: an autoradiographic and quantitative study. J. Neurocytol. 8:265, 1979. 168. Lawson, S. N., Caddy, K. W. T., and Biscoe, T. J.: Development of rat dorsal root ganglion neurons: studies of cell birthdays and changes in mean cell diameter. Cell Tiss. Res. 153:399, 1974. 169. Lawson, S. N., Crepps, B., and Perl, E. R.: Calcitonin generelated peptide immunoreactivity and afferent receptive properties of dorsal root ganglion neurons in guinea-pigs. J. Physiol. (Lond.) 540:989, 2002.

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons 170. Lawson, S. N., Crepps, B. A., and Perl, E. R.: Relationship of substance P to afferent characteristics of dorsal root ganglion neurons in guinea-pig. J. Physiol. (Lond.) 505(Pt. 1):177, 1997. 171. Lawson, S. N., Harper, A. A., Harper, E. I., et al.: A monoclonal antibody against neurofilament protein specifically labels a subpopulation of rat sensory neurons. J. Comp. Neurol. 228:263, 1984. 172. Lawson, S. N., McCarthy, P. W., and Prabhakar, E.: Electrophysiological properties of neurons with CGRP-like immunoreactivity in rat dorsal root ganglia. J. Comp. Neurol. 365:355, 1996. 173. Lawson, S. N., and Waddell, P. J.: Soma neurofilament immunoreactivity is related to cell size and fibre conduction velocity in rat primary sensory neurons. J. Physiol. (Lond.) 435:41, 1991. 174. Le Douarin, N. M., Kalcheim, C., and Teillet, M. A.: The cellular and molecular basis of early sensory ganglion development. In Scott, S. A. (ed.): Sensory Neurons: Diversity, Development and Plasticity. New York, Oxford University Press, p. 143, 1992. 175. Lee, K. H., Chung, K., Chung, J. M., and Coggeshall, R. E.: Correlation of cell body size, axon size and signal conduction velocity for individually labelled dorsal root ganglion cells in the cat. J. Comp. Neurol. 243:335, 1986. 176. Lee, L. Y., and Pisarri, T. E.: Afferent properties and reflex functions of bronchopulmonary C-fibers. Respir. Physiol. 125:47, 2001. 177. Lee, Y. J., Zachrisson, O., Tonge, D. A., and McNaughton, P. A.: Upregulation of bradykinin B2 receptor expression by neurotrophic factors and nerve injury in mouse sensory neurons. Mol. Cell. Neurosci. 19:186, 2002. 178. Leslie, T. A., Emson, P. C., Dowd, P. M., and Woolf, C. J.: Nerve growth factor contributes to the up-regulation of growth-associated protein 43 and preprotachykinin A messenger RNAs in primary sensory neurons following peripheral inflammation. Neuroscience 67:753, 1995. 179. Lessmann, V., Gottmann, K., and Malcangio, M.: Neurotrophin secretion: current facts and future prospects. Prog. Neurobiol. 69:341, 2003. 180. Levy, D., and Zochodne, D. W.: Increased mRNA expression of the B1 and B2 bradykinin receptors and antinociceptive effects of their antagonists in an animal model of neuropathic pain. Pain 86:265, 2000. 181. Lewin, G. R., and Mendell, L. M.: Nerve growth factor and nociception. Trends Neurosci. 16:353, 1993. 182. Li, G. D., Wo, Y., Zhong, M. F., et al.: Expression of fibroblast growth factors in rat dorsal root ganglion neurons and regulation after peripheral nerve injury. Neuroreport 13:1903, 2002. 183. Li, L., Deng, Y. S., and Zhou, X. F.: Downregulation of TrkA expression in primary sensory neurons after unilateral lumbar spinal nerve transection and some rescuing effects of nerve growth factor infusion. Neurosci. Res. 38:183, 2000. 184. Li, W. P., Xian, C., Rush, R. A., and Zhou, X. F.: Upregulation of brain-derived neurotrophic factor and neuropeptide Y in the dorsal ascending sensory pathway following sciatic nerve injury in rat. Neurosci. Lett. 260:49, 1999. 185. Light, A. R., and Perl, E. R.: Peripheral sensory systems. In Dyck, P. J., Thomas, P. K., Griffin, J. W., et al. (eds.):

186.

187.

188.

189.

190.

191.

192.

193.

194.

195.

196.

197.

198.

199.

200.

201.

202.

203.

197

Peripheral Neuropathy, 3rd ed. Philadelphia, W. B. Saunders, p. 149, 1993. Lindholm, D., Heumann, R., Meyer, M., and Thoenen, H.: Interleukin-1 regulates synthesis of nerve growth factor in non-neuronal cells of rat sciatic nerve. Nature 330:658, 1987. Liu, H. X., and Hokfelt, T.: The participation of galanin in pain processing at the spinal level. Trends Pharmacol. Sci. 23:468, 2002. Liu, X., Eschenfelder, S., Blenk, K. H., et al.: Spontaneous activity of axotomized afferent neurons after L5 spinal nerve injury in rats. Pain 84:309, 2000. Lu, J., Zhou, X. F., and Rush, R. A.: Small primary sensory neurons innervating epidermis and viscera display differential phenotype in the adult rat. Neurosci. Res. 41:355, 2001. Luo, X. G., Rush, R. A., and Zhou, X. F.: Ultrastructural localization of brain-derived neurotrophic factor in rat primary sensory neurons. Neurosci. Res. 39:377, 2001. Luo, Z. D., Chaplan, S. R., Higuera, E. S., et al.: Upregulation of dorsal root ganglion (alpha)2(delta) calcium channel subunit and its correlation with allodynia in spinal nerve-injured rats. J. Neurosci. 21:1868, 2001. Ma, W., Chabot, J. G., Powell, K. J., et al.: Localization and modulation of calcitonin gene-related peptide-receptor component protein-immunoreactive cells in the rat central and peripheral nervous systems. Neuroscience 120:677, 2003. Maingret, F., Lauritzen, I., Patel, A. J., et al.: TREK-1 is a heat-activated background K() channel. EMBO J. 19:2483, 2000. Malcangio, M., Getting, S. J., Grist, J., et al.: A novel control mechanism based on GDNF modulation of somatostatin release from sensory neurons. FASEB J. 16:730, 2002. Mannion, R. J., Costigan, M., Decosterd, I., et al.: Neurotrophins: peripherally and centrally acting modulators of tactile stimulus-induced inflammatory pain hypersensitivity. Proc. Natl. Acad. Sci. U. S. A. 96:9385, 1999. Mannsfeldt, A. G., Carroll, P., Stucky, C. L., and Lewin, G. R.: Stomatin, a MEC-2 like protein, is expressed by mammalian sensory neurons. Mol. Cell. Neurosci. 13:391, 1999. Matzner, O., and Devor, M.: Hyperexcitability at sites of nerve injury depends on voltage-sensitive Na channels. J. Neurophysiol. 72:349, 1994. McCallum, J. B., Kwok, W. M., Mynlieff, M., et al.: Loss of T-type calcium current in sensory neurons of rats with neuropathic pain. Anesthesiology 98:209, 2003. McKemy, D. D., Neuhausser, W. M., and Julius, D.: Identification of a cold receptor reveals a general role for TRP channels in thermosensation. Nature 416:52, 2002. McLennan, A. J., Vinson, E. N., Marks, L., et al.: Immunohistochemical localization of ciliary neurotrophic factor receptor a expression in the rat nervous system. J. Neurosci. 16:621, 1996. McMahon, S. B., Armanini, M. P., Ling, L. H., and Phillips, H. S.: Expression and coexpression of Trk receptors in subpopulations of adult primary sensory neurons projecting to identified peripheral targets. Neuron 12:1161, 1994. Mearow, K. M.: The effects of NGF and sensory nerve stimulation on collateral sprouting and gene expression in adult sensory neurons. Exp. Neurol. 151:14, 1998. Medhurst, A. D., Rennie, G., Chapman, C. G., et al.: Distribution analysis of human two pore domain potassium

198

204.

205.

206.

207.

208.

209.

210.

211. 212.

213.

214.

215.

216.

217.

218.

Function of the Peripheral Nervous System channels in tissues of the central nervous system and periphery. Brain Res. Mol. Brain Res. 86:101, 2001. Mendell, L. M., Albers, K. M., and Davis, B. M.: Neurotrophins, nociceptors, and pain. Microsc. Res. Tech. 45:252, 1999. Mense, S., and Meyer, H.: Different types of slowly conducting afferent units in cat skeletal muscle and tendon. J. Physiol. (Lond.) 363:403, 1985. Meyer, R. A., Davis, K. D., Cohen, R. H., et al.: Mechanically insensitive afferents (MIAs) in cutaneous nerves of monkey. Brain Res. 561:252, 1991. Michael, G. J., Averill, S., Nitkunan, A., et al.: Nerve growth factor treatment increases brain-derived neurotrophic factor selectively in TrkA-expressing dorsal root ganglion cells and in their central terminations within the spinal cord. J. Neurosci. 17:8476, 1997. Michael, G. J., Averill, S., Shortland, P. J., et al.: Axotomy results in major changes in BDNF expression by dorsal root ganglion cells: BDNF expression in large trkB and trkC cells, in pericellular baskets, and in projections to deep dorsal horn and dorsal column nuclei. Eur. J. Neurosci. 11:3539, 1999. Michaelis, M., Liu, X., and Janig, W.: Axotomized and intact muscle afferents but no skin afferents develop ongoing discharges of dorsal root ganglion origin after peripheral nerve lesion. J. Neurosci. 20:2742, 2000. Mika, J., Li, Y., Weihe, E., and Schafer, M. K.: Relationship of pronociceptin/orphanin FQ and the nociceptin receptor ORL1 with substance P and calcitonin gene-related peptide expression in dorsal root ganglion of the rat. Neurosci. Lett. 348:190, 2003. Millan, M. J.: The induction of pain: an integrative review. Prog. Neurobiol. 57:1, 1999. Millar, J. A., Barratt, L., Southan, A. P., et al.: A functional role for the two-pore domain potassium channel TASK-1 in cerebellar granule neurons. Proc. Natl. Acad. Sci. U. S. A. 97:3614, 2000. Minami, M., Maekawa, K., Yabuuchi, K., and Satoh, M.: Double in situ hybridization study on coexistence of mu-, delta- and kappa-opioid receptor mRNAs with preprotachykinin A mRNA in the rat dorsal root ganglia. Brain Res. Mol. Brain. Res. 30:203, 1995. Molliver, D. C., Radeke, M. J., Feinstein, S. C., and Snider, W. D.: Presence or absence of TrkA protein distinguishes subsets of small sensory neurons with unique cytochemical characteristics and dorsal horn projections. J. Comp. Neurol. 361:404, 1995. Molliver, D. C., and Snider, W. D.: Nerve growth factor receptor TrkA is down-regulated during postnatal development by a subset of dorsal root ganglion neurons. J. Comp. Neurol. 381:428, 1997. Molliver, D. C., Wright, D. E., Leitner, M. L., et al.: IB4-binding DRG neurons switch from NGF to GDNF dependence in early postnatal life. Neuron 19:849, 1997. Mu, X., Silos-Santiago, I., Carroll, S. L., and Snider, W. D.: Neurotrophin receptor genes are expressed in distinct patterns in developing dorsal root ganglia. J. Neurosci. 13:4029, 1993. Murakami, M., Suzuki, T., Nakagawasai, O., et al.: Distribution of various calcium channel alpha(1) subunits in

219.

220.

221.

222.

223.

224.

225.

226.

227.

228.

229.

230.

231.

232.

233.

234.

235.

murine DRG neurons and antinociceptive effect of omegaconotoxin SVIB in mice. Brain Res. 903:231, 2001. Murphy, P. G., Borthwick, L. A., Altares, M., et al.: Reciprocal actions of interleukin-6 and brain-derived neurotrophic factor on rat and mouse primary sensory neurons. Eur. J. Neurosci. 12:1891, 2000. Nakamura, F., and Strittmatter, S. M.: P2Y1 purinergic receptors in sensory neurons: contribution to touch-induced impulse generation. Proc. Natl. Acad. Sci. U. S. A. 93:10465, 1996. Newton, R. A., Bingham, S., Case, P. C., et al.: Dorsal root ganglion neurons show increased expression of the calcium channel alpha2delta-1 subunit following partial sciatic nerve injury. Brain Res. Mol. Brain Res. 95:1, 2001. Nicholson, R., Small, J., Dixon, A. K., et al.: Serotonin receptor mRNA expression in rat dorsal root ganglion neurons. Neurosci. Lett. 337:119, 2003. Nielson, K. M., Chaverra, M., Hapner, S. J., et al.: PACAP promotes sensory neuron differentiation: blockade by neurotrophic factors. Mol. Cell. Neurosci. 25:629, 2004. Nishii, K., Oura, C., and Pallie, W.: Ultrastructure of the mature pacinian corpuscle in the mesentery of the cat. J. Anat. 106:208, 1970. Nystrom, B., and Hagbarth, K. E.: Microelectrode recordings from transected nerves in amputees with phantom limb pain. Neurosci. Lett. 27:211, 1981. Oakley, R. A., Lefcort, F. B., Clary, D. O., et al.: Neurotrophin-3 promotes the differentiation of muscle spindle afferents in the absence of peripheral targets. J. Neurosci. 17:4262, 1997. Obata, K., Yamanaka, H., Fukuoka, T., et al.: Contribution of injured and uninjured dorsal root ganglion neurons to pain behavior and the changes in gene expression following chronic constriction injury of the sciatic nerve in rats. Pain 101:65, 2003. Ohtori, S., Takahashi, K., Moriya, H., and Myers, R. R.: TNF-alpha and TNF-alpha receptor type 1 upregulation in glia and neurons after peripheral nerve injury: studies in murine DRG and spinal cord. Spine 29:1082, 2004. Okuse, K., Chaplan, S. R., McMahon, S. B., et al.: Regulation of expression of the sensory neuron-specific sodium channel SNS in inflammatory and neuropathic pain. Mol. Cell. Neurosci. 10:196, 1997. Okuse, K., Malik-Hall, M., Baker, M. D., et al.: Annexin II light chain regulates sensory neuron-specific sodium channel expression. Nature 417:653, 2002. Olausson, H., Lamarre, Y., Backlund, H., et al.: Unmyelinated tactile afferents signal touch and project to insular cortex. Nat. Neurosci. 5:900, 2002. Orozco, O. E., Walus, L., Sah, D. W., et al.: GFRalpha3 is expressed predominantly in nociceptive sensory neurons. Eur. J. Neurosci. 13:2177, 2001. Pape, H. C.: Queer current and pacemaker: the hyperpolarization-activated cation current in neurons. Annu. Rev. Physiol. 58:299, 1996. Pare, M., Behets, C., and Cornu, O.: Paucity of presumptive Ruffini corpuscles in the index finger pad of humans. J. Comp. Neurol. 456:260, 2003. Pare, M., Smith, A. M., and Rice, F. L.: Distribution and terminal arborizations of cutaneous mechanoreceptors in

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons

236.

237.

238.

239.

240.

241.

242.

243.

244.

245.

246.

247.

248.

249.

250.

251.

252.

253.

the glabrous finger pads of the monkey. J. Comp. Neurol. 445:347, 2002. Passmore, G. M., Selyanko, A. A., Mistry, M., et al.: KCNQ/M currents in sensory neurons: significance for pain therapy. J. Neurosci. 23:7227, 2003. Patel, A. J., and Honore, E.: Properties and modulation of mammalian 2P domain K channels. Trends Neurosci. 24:339, 2001. Peier, A. M., Moqrich, A., Hergarden, A. C., et al.: A TRP channel that senses cold stimuli and menthol. Cell 108:705, 2002. Peier, A. M., Reeve, A. J., Andersson, D. A., et al.: A heatsensitive TRP channel expressed in keratinocytes. Science 296:2046, 2002. Perl, E. R.: Myelinated afferent fibres innervating the primate skin and their response to noxious stimuli. J. Physiol. (Lond.) 197:593, 1968. Perl, E. R.: Function of dorsal root ganglion neurons: an overview. In Scott, S. A. (ed.): Sensory Neurons: Diversity, Development and Plasticity. New York, Oxford University Press, p. 3, 1992. Perl, E. R., and Burgess, P. R.: Classification of afferent dorsal root fibers: mammals. In Altman, P. L., and Dittmer, D. S. (eds.): Biology Data Book. Bethesda, MD, FASEB, p. 1141, 1973. Perry, M. J., and Lawson, S. N.: Neurofilament in feline primary afferent neurons: a quantitative immunocytochemical study. Brain Res. 607:307, 1993. Perry, M. J., and Lawson, S. N.: Differences in expression of oligosaccharides, neuropeptides, carbonic anhydrase and neurofilament in rat primary afferent neurons retrogradely labelled via skin, muscle or visceral nerves. Neuroscience 85:293, 1998. Perry, M. J., Lawson, S. N., and Robertson, J.: Neurofilament immunoreactivity in populations of rat primary afferent neurons: a quantitative study of phosphorylated and nonphosphorylated subunits. J. Neurocytol. 20:746, 1991. Pezet, S., Malcangio, M., and McMahon, S. B.: BDNF: a neuromodulator in nociceptive pathways? Brain Res. Brain Res. Rev. 40:240, 2002. Pezet, S., Onteniente, B., Jullien, J., et al.: Differential regulation of NGF receptors in primary sensory neurons by adjuvant-induced arthritis in the rat. Pain 90:113, 2001. Plenderleith, M. B., Cameron, A. A., Key, B., and Snow, P. J.: Soybean agglutinin binds to a subpopulation of primary afferent neurons in the cat. Neurosci. Lett. 86:257, 1988. Pomonis, J. D., Rogers, S. D., Peters, C. M., et al.: Expression and localization of endothelin receptors: implications for the involvement of peripheral glia in nociception. J. Neurosci. 21:999, 2001. Popratiloff, A., Weinberg, R. J., Rustioni, A.: AMPA receptors at primary afferent synapses in substantia gelatinosa after sciatic nerve section. Eur. J. Neurosci. 10:3220, 1998. Prabhakar, E., and Lawson, S. N.: The electrophysiological properties of rat primary afferent neurons with carbonic anhydrase activity. J. Physiol. (Lond.) 482:609, 1995. Price, M. P., Lewin, G. R., McIlwrath, S. L., et al.: The mammalian sodium channel BNC1 is required for normal touch sensation. Nature 407:1007, 2000. Price, M. P., McIlwrath, S. L., Xie, J., et al.: The DRASIC cation channel contributes to the detection of

254.

255.

256.

257.

258.

259.

260.

261.

262.

263.

264. 265.

266. 267. 268.

269.

269a.

270.

271.

199

cutaneous touch and acid stimuli in mice. Neuron 32:1071, 2001. Rambourg, A., Clermont, Y., and Beaudet, A.: Ultrastructural features of six types of neurons in rat dorsal root ganglia. J. Neurocytol. 12:47, 1983. Rasband, M. N., Park, E. W., Vanderah, T. W., et al.: Distinct potassium channels on pain-sensing neurons. Proc. Natl. Acad. Sci. U. S. A. 98:13373, 2001. Reeh, P. W., and Kress, M.: Molecular physiology of proton transduction in nociceptors. Curr. Opin. Pharmacol. 1:45, 2001. Reid, G., Babes, A., and Pluteanu, F.: A cold- and mentholactivated current in rat dorsal root ganglion neurons: properties and role in cold transduction. J. Physiol. (Lond.) 545:595, 2002. Reid, G., and Flonta, M.: Cold transduction by inhibition of a background potassium conductance in rat primary sensory neurons. Neurosci. Lett. 297:171, 2001. Reid, G., and Flonta, M. L.: Ion channels activated by cold and menthol in cultured rat dorsal root ganglion neurons. Neurosci. Lett. 324:164, 2002. Reinohl, J., Hoheisel, U., Unger, T., and Mense, S.: Adenosine triphosphate as a stimulant for nociceptive and non-nociceptive muscle group IV receptors in the rat. Neurosci. Lett. 338:25, 2003. Renganathan, M., Cummins, T. R., and Waxman, S. G.: Contribution of Na(v)1.8 sodium channels to action potential electrogenesis in DRG neurons. J. Neurophysiol. 86:629, 2001. Richardson, J. D., and Vasko, M. R.: Cellular mechanisms of neurogenic inflammation. J. Pharmacol. Exp. Ther. 302:839, 2002. Ritter, A. M., and Mendell, L. M.: Somal membrane properties of physiologically identified sensory neurons in the rat: effects of nerve growth factor. J. Neurophysiol. 68:2033, 1992. Robbins, J.: KCNQ potassium channels: physiology, pathophysiology, and pharmacology. Pharmacol. Ther. 90:1, 2001. Robertson, B., Perry, M., and Lawson, S. N.: Populations of rat spinal primary afferent neurons with choleragenoid binding compared with those labelled by markers for neurofilament and carbohydrate groups: a quantitative immunocytochemical study. J. Neurocytol. 20:387, 1991. Roux, P. P., and Barker, P. A.: Neurotrophin signalling through the p75 receptor. Prog. Neurobiol. 67:203, 2002. Ruscheweyh, R., and Sandkuhler, J.: Role of kainate receptors in nociception. Brain Res. Brain Res. Rev. 40:215, 2002. Safieh-Garabedian, B., Poole, S., Allchorne, A., et al.: Contribution of interleukin-1 beta to the inflammationinduced increase in nerve growth factor levels and inflammatory hyperalgesia. Br. J. Pharmacol. 115:1265, 1995. Sangameswaran, L., Fish, L. M., Koch, B. D., et al.: A novel tetrodotoxin-sensitive, voltage-gated sodium channel expressed in rat and human dorsal root ganglia. J. Biol. Chem. 272:14805, 1997. Sata, K., Kiyama, H., Park, H. T., et al: AMPA, KA and NMDA receptors are expressed in the rat DRG neurones. Neuroreport 4:1263–1265, 1993. Schafers, M., Geis, C., Brors, D., et al.: Anterograde transport of tumor necrosis factor-alpha in the intact and injured rat sciatic nerve. J. Neurosci. 22:536, 2002. Schafers, M., Forkin, L. S., Geis, C., and Shubayev, V. I.: Spinal nerve ligation induces transient upregulation of

200

272.

273. 274.

275.

276.

277.

278.

279.

280.

281.

282.

283.

284.

285. 286.

287.

Function of the Peripheral Nervous System tumor necrosis factor receptors 1 and 2 in injured and adjacent uninjured dorsal root ganglia in the rat. Neurosci. Lett. 347:179, 2003. Schmelz, M., Schmidt, R., Weidner, C., et al.: Chemical response pattern of different classes of C-nociceptors to pruritogens and algogens. J. Neurophysiol. 89:2441, 2003. Schmidt, R. F.: The articular polymodal nociceptor in health and disease. Prog. Brain Res. 113:53, 1996. Schoenen, J., Delree, P., Leprince, P., and Moonen, G.: Neurotransmitter phenotype plasticity in cultured dissociated adult rat dorsal root ganglia: an immunocytochemical study. J. Neurosci. Res. 22:473, 1989. Scholz, A., Gruss, M., and Vogel, W.: Properties and functions of calcium-activated K channels in small neurons of rat dorsal root ganglion studied in a thin slice preparation. J. Physiol. (Lond.) 513:55, 1998. Schulz, S., Schreff, M., Schmidt, H., et al.: Immunocytochemical localization of somatostatin receptor sst2A in the rat spinal cord and dorsal root ganglia. Eur. J. Neurosci. 10:3700, 1998. Scroggs, R. S., and Fox, A. P.: Multiple Ca2 currents elicited by action potential waveforms in acutely isolated adult rat dorsal root ganglion neurons. J. Neurosci. 12:1789, 1992. Scroggs, R. S., Todorovic, S. M., Anderson, E. G., and Fox, A. P.: Variation in I(H), I(IR), and I(LEAK) between acutely isolated adult rat dorsal root ganglion neurons of different size. J. Neurophysiol. 71:271, 1994. Shah, B. S., Stevens, E. B., Gonzalez, M. I., et al.: Beta3, a novel auxiliary subunit for the voltage-gated sodium channel, is expressed preferentially in sensory neurons and is upregulated in the chronic constriction injury model of neuropathic pain. Eur. J. Neurosci. 12:3985, 2000. Shamash, S., Reichert, F., and Rotshenker, S.: The cytokine network of wallerian degeneration: tumor necrosis factor-, interleukin-1, and interleukin-1. J. Neurosci. 22:3052, 2002. Sharp, G. A., Shaw, G., and Weber, K.: Immunoelectronmicroscopical localisation of the three neurofilament triplet proteins along neurofilaments of cultured dorsal root ganglion neurons. Exp. Cell Res. 137:403, 1982. Shen, H., Chung, J. M., Coggeshall, R. E., and Chung, K.: Changes in trkA expression in the dorsal root ganglion after peripheral nerve injury. Exp. Brain Res. 127:141, 1999. Shi, T. J., Tandrup, T., Bergman, E., et al.: Effect of peripheral nerve injury on dorsal root ganglion neurons in the C57 BL/6J mouse: marked changes both in cell numbers and neuropeptide expression. Neuroscience 105:249, 2001. Shin, J. B., Martinez-Salgado, C., Heppenstall, P. A., and Lewin, G. R.: A T-type calcium channel required for normal function of a mammalian mechanoreceptor. Nat. Neurosci. 6:724, 2003. Shu, X. Q., and Mendell, L. M.: Neurotrophins and hyperalgesia. Proc. Natl. Acad. Sci. U. S. A. 96:7693, 1999. Shubayev, V. I., and Myers, R. R.: Anterograde TNF alpha transport from rat dorsal root ganglion to spinal cord and injured sciatic nerve. Neurosci. Lett. 320:99, 2002. Shuto, T., Horie, H., Hikawa, N., et al.: IL-6 up-regulates CNTF mRNA expression and enhances neurite regeneration. Neuroreport 12:1081, 2001.

288. Simone, D. A., and Kajander, K. C.: Responses of cutaneous A-fiber nociceptors to noxious cold. J. Neurophysiol. 77:2049, 1997. 289. Sleeper, A. A., Cummins, T. R., Dib-Hajj, S. D., et al.: Changes in expression of two tetrodotoxin-resistant sodium channels and their currents in dorsal root ganglion neurons after sciatic nerve injury but not rhizotomy. J. Neurosci. 20:7279, 2000. 290. Smith, G. D., Gunthorpe, M. J., Kelsell, R. E., et al.: TRPV3 is a temperature-sensitive vanilloid receptor-like protein. Nature 418:186, 2002. 291. Snijdelaar, D. G., Dirksen, R., Slappendel, R., and Crul, B. J.: Substance P. Eur. J. Pain 4:121, 2000. 292. Sommer, C., Schmidt, C., and George, A.: Hyperalgesia in experimental neuropathy is dependent on the TNF receptor 1. Exp. Neurol. 151:138, 1998. 293. Song, W. J.: Genes responsible for native depolarizationactivated K currents in neurons. Neurosci. Res. 42:7, 2002. 294. Stark, B., Carlstedt, T., and Risling, M.: Distribution of TGF-beta, the TGF-beta type I receptor and the R-II receptor in peripheral nerves and mechanoreceptors: observations on changes after traumatic injury. Brain Res. 913:47, 2001. 295. Stebbing, M. J., Eschenfelder, S., Habler, H. J., et al.: Changes in the action potential in sensory neurons after peripheral axotomy in vivo. Neuroreport 10:201, 1999. 296. Sterne, G. D., Brown, R. A., Green, C. J., and Terenghi, G.: NT-3 modulates NPY expression in primary sensory neurons following peripheral nerve injury. J. Anat. 193(Pt. 2):273, 1998. 297. Story, G. M., Peier, A. M., Reeve, A. J., et al.: ANKTM1, a TRP-like channel expressed in nociceptive neurons, is activated by cold temperatures. Cell 112:819, 2003. 298. Streit, W. J., Schulte, B. A., Balentine, J. D., and Spicer, S. S.: Histochemical localization of galactose-containing glycoconjugates in sensory neurons and their processes in the central and peripheral nervous system of the rat. J. Histochem. Cytochem. 10:1042, 1985. 299. Stucky, C., Shin, J. B., and Lewin, G. R.: Neurotrophin-4: a survival factor for adult sensory neurons. Curr. Biol. 12:1401, 2002. 300. Stucky, C. L., and Koltzenburg, M.: The low-affinity neurotrophin receptor p75 regulates the function but not the selective survival of specific subpopulations of sensory neurons. J. Neurosci. 17:4398, 1997. 301. Stucky, C. L., and Lewin, G. R.: Isolectin B(4)-positive and -negative nociceptors are functionally distinct. J. Neurosci. 19:6497, 1999. 302. Study, R. E., and Kral, M. G.: Spontaneous action potential activity in isolated dorsal root ganglion neurons from rats with a painful neuropathy. Pain 65:235, 1996. 303. Sugiura, S., Lahav, R., Han, J., et al.: Leukaemia inhibitory factor is required for normal inflammatory responses to injury in the peripheral and central nervous systems in vivo and is chemotactic for macrophages in vitro. Eur. J. Neurosci. 12:457, 2000. 304. Sugiura, T., Tominaga, M., Katsuya, H., and Mizumura, K.: Bradykinin lowers the threshold temperature for heat activation of vanilloid receptor 1. J. Neurophysiol. 88:544, 2002.

The Peripheral Sensory Nervous System: Dorsal Root Ganglion Neurons 305. Suzuki, H., Hase, A., Miyata, Y., et al.: Prominent expression of glial cell line-derived neurotrophic factor in human skeletal muscle. J. Comp. Neurol. 402:303, 1998. 306. Suzuki, M., Mizuno, A., Kodaira, K., and Imai, M.: Impaired pressure sensation in mice lacking TRPV4. J. Biol. Chem. 278:22664, 2003. 307. Suzuki, M., Watanabe, Y., Oyama, Y., et al.: Localization of mechanosensitive channel TRPV4 in mouse skin. Neurosci. Lett. 353:189, 2003. 308. Szarijanni, N., and Rethelyi, M.: Differential distribution of small and large neurons in the sacrococcygeal dorsal root ganglion of the cat. Acta Morphol. Acad. Sci. Hung. 27:25, 1979. 309. Szolcsanyi, J., Anton, F., Reeh, P. W., and Handwerker, H. O.: Selective excitation by capsaicin of mechano-heat sensitive nociceptors in rat skin. Brain Res. 446:262, 1988. 310. Taiwo, Y. O., and Levine, J. D.: Serotonin is a directly-acting hyperalgesic agent in the rat. Neuroscience 48:485, 1992. 311. Talley, E. M., Lei, Q., Sirois, J. E., and Bayliss, D. A.: TASK-1, a two-pore domain K channel, is modulated by multiple neurotransmitters in motoneurons. Neuron 25:399, 2000. 312. Tanaka, M., Cummins, T. R., Ishikawa, K., et al.: SNS Na channel expression increases in dorsal root ganglion neurons in the carrageenan inflammatory pain model. Neuroreport 9:967, 1998. 313. Tate, S., Benn, S., Hick, C., et al.: Two sodium channels contribute to the TTX-R sodium current in primary sensory neurons. Nat. Neurosci. 1:653, 1998. 314. Terenghi, G.: Peripheral nerve regeneration and neurotrophic factors. J. Anat. 194(Pt. 1):1, 1999. 315. Their, M., Marz, P., Otten, U., et al.: Interleukin-6 (IL-6) and its soluble receptor support survival of sensory neurons. J. Neurosci. Res. 55:411, 1999. 316. Thippeswamy, T., and Morris, R.: The roles of nitric oxide in dorsal root ganglion neurons. Ann. N. Y. Acad. Sci. 962:103, 2002. 317. Thompson, S. W., Bennett, D. L., Kerr, B. J., et al.: Brainderived neurotrophic factor is an endogenous modulator of nociceptive responses in the spinal cord. Proc. Natl. Acad. Sci. U. S. A. 96:7714, 1999. 318. Thompson, S. W., Vernallis, A. B., Heath, J. K., and Priestley, J. V.: Leukaemia inhibitory factor is retrogradely transported by a distinct population of adult rat sensory neurons: co-localization with trkA and other neurochemical markers. Eur. J. Neurosci. 9:1244, 1997. 319. Toledo-Aral, J. J., Moss, B. L., He, Z. J., et al.: Identification of PN1, a predominant voltage-dependent sodium channel expressed principally in peripheral neurons. Proc. Natl. Acad. Sci. U. S. A. 94:1527, 1997. 320. Tominaga, M., Caterina, M. J., Malmberg, A. B., et al.: The cloned capsaicin receptor integrates multiple pain-producing stimuli. Neuron 21:531, 1998. 321. Tong, Y. G., Wang, H. F., Ju, G., et al.: Increased uptake and transport of cholera toxin B-subunit in dorsal root ganglion neurons after peripheral axotomy: possible implications for sensory sprouting. J. Comp. Neurol. 404:143, 1999. 322. Treede, R. D., Meyer, R. A., and Campbell, J. N.: Myelinated mechanically insensitive afferents from monkey hairy skin: heat-response properties. J. Neurophysiol. 80:1082, 1998.

201

323. Vallbo, A., Olausson, A., Wessberg, H., and Norrsell, U.: A system of unmyelinated afferents for innocuous mechanoreception in the human skin. Brain Res. 628:301, 1993. 324. Valtschanoff, J. G., Phend, K. D., Bernardi, P. S., et al.: Amino acid immunocytochemistry of primary afferent terminals in the rat dorsal horn. J. Comp. Neurol. 346:237, 1994. 325. Vergara, C., Latorre, R., Marrion, N. V., and Adelman, J. P.: Calcium-activated potassium channels. Curr. Opin. Neurobiol. 8:321, 1998. 326. Verge, V. M., Richardson, P. M., Wiesenfeld-Hallin, Z., and Hokfelt, T.: Differential influence of nerve growth factor on neuropeptide expression in vivo: a novel role in peptide suppression in adult sensory neurons. J. Neurosci. 15:2081, 1995. 327. Vestergaard, S., Tandrup, T., and Jakobsen, J.: Effect of permanent axotomy on number and volume of dorsal root ganglion cell bodies. J. Comp. Neurol. 388:307, 1997. 328. Viana, F., de la Pena, E., and Belmonte, C.: Specificity of cold thermotransduction is determined by differential ionic channel expression. Nat. Neurosci. 5:254, 2002. 329. Villiere, V., and McLachlan, E. M.: Electrophysiological properties of neurons in intact rat dorsal root ganglia classified by conduction velocity and action potential duration. J. Neurophysiol. 76:1924, 1996. 330. Waddell, P. J., and Lawson, S. N.: Electrophysiological properties of subpopulations of rat dorsal root ganglion neurons in vitro. Neuroscience 36:811, 1990. 331. Wall, P. D., and Gutnick, M.: Ongoing activity in peripheral nerves: the physiology and pharmacology of impulses originating from a neuroma. Exp. Neurol. 43:580, 1974. 332. Wang, R., Guo, W., Ossipov, M. H., et al.: Glial cell linederived neurotrophic factor normalizes neurochemical changes in injured dorsal root ganglion neurons and prevents the expression of experimental neuropathic pain. Neuroscience 121:815, 2003. 333. Watkins, L. R., and Maier, S. F.: Beyond neurons: evidence that immune and glial cells contribute to pathological pain states. Physiol. Rev. 82:981, 2002. 334. White, F. A., Silos-Santiago, I., Molliver, D. C., et al.: Synchronous onset of NGF and TrkA survival dependence in developing dorsal root ganglia. J. Neurosci. 16:4662, 1996. 335. Wood, J. N.: Recent advances in understanding molecular mechanisms of primary afferent activation. Gut 53(Suppl. 2):II9, 2004. 336. Woolf, C. J., Allchorne, A., Safieh-Garabedian, B., and Poole, S.: Cytokines, nerve growth factor and inflammatory hyperalgesia: the contribution of tumour necrosis factor alpha. Br. J. Pharmacol. 121:417, 1997. 337. Woolf, C. J., Ma, Q. P., Allchorne, A., and Poole, S.: Peripheral cell types contributing to the hyperalgesic action of nerve growth factor in inflammation. J. Neurosci. 16:2716, 1996. 338. Woolf, C. J., Safieh-Garabedian, B., Ma, Q. P., et al.: Nerve growth factor contributes to the generation of inflammatory sensory hypersensitivity. Neuroscience 62:327, 1994. 339. Wright, D. E., and Snider, W. D.: Neurotrophin receptor mRNA expression defines distinct populations of neurons in rat dorsal root ganglia. J. Comp. Neurol. 351:329, 1995. 340. Wu, G., Ringkamp, M., Hartke, T. V., et al.: Early onset of spontaneous activity in uninjured C-fiber nociceptors after injury to neighboring nerve fibers. J. Neurosci. 21:RC140, 2001.

202

Function of the Peripheral Nervous System

341. Wyatt, S., Shooter, E. M., and Davies, A. M.: Expression of the NGF receptor gene in sensory neurons and their cutaneous targets prior to and during innervation. Neuron 4:421, 1990. 342. Xiao, H. S., Huang, Q. H., Zhang, F. X., et al.: Identification of gene expression profile of dorsal root ganglion in the rat peripheral axotomy model of neuropathic pain. Proc. Natl. Acad. Sci. U. S. A. 99:8360, 2002. 343. Xu, H., Ramsey, I. S., Kotecha, S. A., et al.: TRPV3 is a calcium-permeable temperature-sensitive cation channel. Nature 418:181, 2002. 344. Yamadori, T.: A light and electron microscopic study on the postnatal development of spinal ganglia in rats. Acta Anat. Nippon 45:191, 1970. 345. Yang, E. K., Takimoto, K., Hayashi, Y., et al.: Altered expression of potassium channel subunit mRNA and alpha-dendrotoxin sensitivity of potassium currents in rat dorsal root ganglion neurons after axotomy. Neuroscience 123:867, 2004. 346. Yoon, Y. W., Na, H. S., and Chung, J. M.: Contributions of injured and intact afferents to neuropathic pain in an experimental rat model. Pain 64:27, 1996.

347. Yoshida, S., and Matsuda, Y.: Studies on sensory neurons of the mouse with intracellular recording and dye injection techniques. J. Neurophysiol. 42:1134, 1979. 348. Yusaf, S. P., Goodman, J., Pinnock, R. D., et al.: Expression of voltage-gated calcium channel subunits in rat dorsal root ganglion neurons. Neurosci. Lett. 311:137, 2001. 349. Zhou, X. F., Cameron, D., and Rush, R. A.: Endogenous neurotrophin-3 supports the survival of a subpopulation of sensory neurons in neonatal rat. Neuroscience 86:1155, 1998. 350. Zhou, X. F., Chie, E. T., Deng, Y. S., et al.: Injured primary sensory neurons switch phenotype for brain-derived neurotrophic factor in the rat. Neuroscience 92:841, 1999. 351. Zhou, S.-F., Deng, Y.-S., Chie, E., et al.: Satellite cell-derived nerve growth factor and neurotrophins are involved in noradrenergic sprouting in the dorsal root ganglia following peripheral nerve injury in the rat. Eur. J. Neurosci. 11:1711, 1999. 352. Zhou, Z., Davar, G., and Strichartz, G.: Endothelin-1 (ET-1) selectively enhances the activation gating of slowly inactivating tetrodotoxin-resistant sodium currents in rat sensory neurons: a mechanism for the pain-inducing actions of ET-1. J. Neurosci. 22:6325, 2002.

9 The Pupil JAMES J. CORBETT

Anatomy of Iris Innervation Sympathetic Innervation Parasympathetic Innervation to the Intraocular Muscles Horner’s Syndrome The Size of the Normal Pupil

Simple Anisocoria Pupillary Unrest and Hippus Pupil Cycle Time Iris Damage and Its Effect on the Light, Near, and Drug Responses

The iris can be seen easily, and it supports autonomically innervated constrictor and dilator muscles that, when properly examined, can be used as a rough indicator of damage to the peripheral nerves. The limited utility of such testing is related to the opposing nature of the iris muscles and to the fact that the iris is an extension of the uveal tract. The iris is involved in uveal inflammation that can alter its size, shape, and reactivity to light response, near reaction, and various drugs. The reaction of the iris to light, dark, and near and to adrenergic and cholinergic drugs, as well as supersensitivity reactions and the appearance on slit-lamp examination, have produced considerable literature with conflicting conclusions. This welter of information is best dealt with by separating the responses of the pupil to light, near, dark, and drugs and by examining the effects of classic isolated lesions on each system individually.

ANATOMY OF IRIS INNERVATION Sympathetic Innervation The fibers of the dilator muscle are in the most posterior layer of the iris and are saturated with the melanin that keeps the iris opaque to light (Fig. 9–1). The central cells of origin for sympathetic innervation of the iris dilator are in the posterolateral hypothalamus. The axons traverse the brainstem to spinal cord levels C7-T1, where they synapse in the ciliospinal center of Budge. Axons from the intermediolateral columns of the spinal cord provide the second step in the three-neuron arc. From the spinal cord

Efferent Pupillary Defects Defects of Parasympathetic Innervation The Pupil as a Measure of Autonomic Dysfunction

and beyond, the fibers form part of the peripheral nervous system. The fibers proceed across the apex of the lung, around the ansa subclavia, and pass through the stellate (inferior cervical) ganglion without synapse. They ascend to the superior cervical ganglion (C3-C4 level, angle of the mandible), where they synapse and travel with the internal carotid artery through the foramen lacerum into the cavernous sinus. The sympathetic fibers join the abducens nerve for about 4 mm and then enter the orbit with the ophthalmic division of the trigeminal nerve. They then follow the long posterior ciliary nerves to segmentally innervate the radially arranged fibers of the dilator muscles in the iris (Fig. 9–2).69 The entire three-neuron sympathetic outflow path must be intact to provide tonic output of norepinephrine at the effector cell. A lesion to any one of these three neuron/axon regions stops the tonic output of norepinephrine into the synaptic cleft at the receptors of the dilator muscle and causes a Horner’s syndrome (see below) (Fig. 9–3).

Parasympathetic Innervation to the Intraocular Muscles The iris sphincter and the ciliary muscle are controlled by the Edinger-Westphal nucleus, which is a dorsal-rostral part of the oculomotor nucleus. The third cranial nerve leaves the midbrain in the interpeduncular fossa. Near its epineurium are carried the small-caliber fibers that serve the intraocular muscles. The synapse in this typical two-neuron parasympathetic outflow path is in the ciliary 203

204

Function of the Peripheral Nervous System

FIGURE 9–1 The sphincter muscle (c) is innervated by axons from ciliary ganglion neurons. Slit-lamp examination of the iris suggests that there are between 20 and 30 motor units in the sphincter, but higher estimates (up to 70) have been made. The neuromuscular bundles are arranged in overlapping sectors around the edge of the iris. The dilator muscle (f) is arranged radially; the fibers insert by long extensions called spurs that blend into the sphincter muscle. a ⫽ Anterior bundle; b ⫽ pigment ruff; c ⫽ sphincter muscle; d ⫽ vascular arcades; e ⫽ groups of capillaries, nerves, and melanocytes; f ⫽ dilator muscle; g ⫽ anterior epithelium; h ⫽ Michel’s spur; i ⫽ Fuch’s spur; j ⫽ posterior epithelium. (From Hogan, M. J., Alvarodo, J. A., and Weddell, J. E.: Histology of the Human Eye. Philadelphia, W. B. Saunders, 1971, with permission.)

ganglion, which is located deep in the muscle cone of the orbit about 1 cm behind the globe and just lateral to the optic nerve and medial to the lateral rectus muscle. The postganglionic fibers follow the short ciliary nerves into the sclera to the anterior segment of the globe (see Fig. 9–2). The fibers of the ciliary nerves run between the choroid and the sclera and may be damaged by the retinal photocoagulation used to treat diabetic retinopathy.43,48,53

Horner’s Syndrome Horner’s syndrome is caused by interruption of the ipsilateral sympathetic outflow to the head and neck. It can result from lesions in the brainstem or spinal cord, damage at the lung apex and in the supraclavicular space, or damage to the

carotid plexus along the internal carotid artery all the way to the cavernous sinus. Since the same clinical picture is produced wherever this pathway is interrupted, the combination of miosis and ptosis by itself is not particularly helpful in localizing the lesion.65,73 Recent reports of isolated Horner’s syndrome as the only manifestation of syringomyelia emphasize the need to evaluate this syndrome thoroughly and to be sure that the ptosis and miosis are not due to pseudo-Horner’s syndrome.21,42,63 Clinical Signs Wherever Horner’s syndrome occurs, the lesion also affects neighboring anatomic structures, causing signs and symptoms that help localize the lesion. Unfortunately, efforts to evaluate sweating deficits add little, and the

The Pupil

205

FIGURE 9–2 The optic globe with the sclera peeled back, disclosing the segmental array of short and long ciliary nerves. These penetrate the posterior sclera in proximity to the optic nerve and lie between the choroids and the sclera. In this location, they are vulnerable to photocoagulation burns. At the limbus of the cornea, where they come into contact with the iris, the nerves may be torn loose by blows to the eye. (From Pernkoff, E.: Atlas of Topographical and Applied Human Anatomy. Berlin, Urban & Schwarzenberg, 1963.)

sweat “patterns” tend to be equivocal in their localizing specificity. Studies of sweating patterns in Horner’s syndrome using quinazirine and starch and iodine, with occasional erratic skin temperature results and galvanic skin responses, have been plagued by variability, most likely because of aberrant regeneration.36,49 Ptosis. Müller’s muscle, the sympathetically innervated retractor of the upper lid, is paralyzed in Horner’s syndrome, and this narrows the palpebral fissure. The ptosis in Horner’s syndrome, by itself, never covers the visual axis. With the patient’s eyes wide open and gazing upward, the ptosis almost disappears.65 The lower lid also has some sympathetically innervated retractor fibers. With sympathetic denervation the lower lid rises slightly (“upside-down ptosis”).37 This further adds to the narrowed palpebral fissure and gives the appearance of enophthalmos. There is no true enophthalmos.26,75 Miosis. The miosis of Horner’s syndrome is never intense. Tiny pupils (less than 3 mm in diameter) are due to some

other (or additional) cause, such as miotic drops, loss of supranuclear inhibition (as in coma), or long-standing immobility (as in Adie’s syndrome or Argyll Robertson pupil). Paralysis of the pupillodilator muscle causes only a moderate decrease in pupil size. In bright light, both pupils constrict normally and are nearly equal. Anisocoria in unilateral sympathetic denervation is more evident in dim light because the sphincter relaxes in both eyes; this leaves the normally functioning dilator muscle to dilate the normal pupil and a flaccid dilator in the affected eye, which eventually passively dilates to almost normal size. Dilation Lag. The slowness of the affected pupil to dilate is characteristic of Horner’s syndrome (Fig. 9–4). Since the problem is usually unilateral, this can be of great clinical help in making the diagnosis. Pupil dilation has two components: passive dilation of the iris as a result of inhibition of the sphincter, and active contraction of the sympathetically innervated radial iris muscles. When the lights are turned off, both pupils dilate. However, the

FIGURE 9–3 Parasympathetic and sympathetic pathways to the eye. In the lower part of the figure, the solid line indicates the pathway of the pupillodilator fibers, and the dashed lines show some of the other sympathetic pathways to the orbit and the face. 1, Edinger-Westphal nucleus; 2, oculomotor nerve; 3, branch to inferior oblique muscle; 4, motor root of ciliary ganglion; 5, short ciliary nerves; 6, ciliary muscle and iris sphincter; 7, superior cervical ganglion; 8, internal carotid artery; 9, external carotid artery; 10, sudomotor fibers to the face; 11, carotid plexus; 12, caroticotympanic nerve; 13, tympanic plexus; 14, deep petrosal nerve; 15, lesser superficial petrosal nerve; 16, sympathetic contribution to the vidian nerve; 17, ophthalmic division of trigeminal nerve; 18, nasociliary nerve; 19, long ciliary nerve; 20, ciliary muscle and iris dilator; 21, probable path of sympathetic contribution to retractor muscles of the lids; 22, vasomotor and some sudomotor fibers; 23, ophthalmic artery; 24, lacrimal gland; 25, short ciliary nerves; 26, sympathetic contribution to salivary glands; 27, greater superficial petrosal nerve. (Prepared by H. Stanley Thompson and drawn by Joan Esperson, University of California Medical School, San Francisco, California. From Walsh, F. B., and Hoyt, W. F.: Clinical Neuro-Ophthalmology, Vol. 1. Baltimore, Williams & Wilkins, 1969, with permission.)

The Pupil

207

FIGURE 9–4 Clinical example of “dilation lag” in Horner’s syndrome. Dilation of the pupil occurs when the sphincter stops contracting and the dilator fibers begin to fire. When the iris is sympathetically denervated, the iris dilators do not add speed to opening the pupil, whereas they do in the normal eye. A normal eye dilates almost fully in the dark within 5 seconds. Without sympathetics it takes much longer to accomplish the same phenomenon, hence the name of the phenomenon, dilation lag.

sympathetically denervated pupil dilates more slowly because no active contraction of the dilator muscle occurs; only sphincter inhibition permits the pupil to dilate.41 As the pupil slowly dilates, the anisocoria gradually decreases. Thus there is a lag in dilation when sympathetic denervation is present. This phenomenon can be seen by illuminating both eyes tangentially from below as the room lights are turned off. The anisocoria is much greater at 5 seconds after “lights out” than at 15 seconds. Dilation lag can be documented with flash photos at 5 and then again at 15 seconds; better yet, infrared-sensitive video can be used.76 Facial Anhidrosis. Hemifacial sweating disorders or hemifacial blanching or flushing sometimes draw attention to a sympathetic deficit, but these phenomena are not consistently present in Horner’s syndrome. Again, aberrant regeneration of damaged fibers may contribute to the uncertainty.49,65 Conjunctival and Lid Hyperemia. Conjunctival and lid hyperemia is related to the defect in ipsilateral innervation of all extracranial blood vessels. Loss of vasoconstrictive tone leaves the conjunctiva hyperemic. A fresh Horner’s syndrome is frequently mistaken for conjunctivitis.69 Pupil Pharmacology as Related to Autonomic Denervation. There are many pitfalls in trying to interpret the response of pupils to eyedrops. Individual variability of responsiveness to drugs makes it important to compare the effect of drops in both eyes. Emotional upset or anxiety will dilate the pupils, and drowsiness will constrict them. Both states interfere with evaluation.

Dry eyes, numb eyes, exposed eyes, and eyes that recently have been touched by a wisp of cotton, a topical anesthetic, a tonometer, or a contact lens all are likely to have damaged corneal epithelium.4 Such an eye is likely to soak up more of any drug than an eye with a normal corneal epithelium, which acts as a barrier to absorption. Variation in corneal permeability is especially important when supersensitivity tests are being done. Weak solutions may cause pupillary reactions that would not occur but for highly permeable corneas. In general, the more pigment in the iris, the more slowly the drug takes effect and the longer its action lingers.39,40,46 This is probably due to binding of the drug to iris melanin, from which it is then slowly released. There are wide individual variations in the pupillary responses to topical drugs. The range of responses among blue eyes includes the average response of dark brown eyes,46 so it is probably not just a matter of melanin; there may be genetic factors that influence corneal penetration. Sympathetic Pharmacology Cocaine (5% to 10%) is applied to the conjunctiva as a topical anesthetic, as a mydriatic, and as a test for Horner’s syndrome.11,18,74 Its mydriatic effect is due to accumulation of norepinephrine at the adrenergic receptors (synaptic sites) of the dilator cells. Cocaine causes the transmitter substance to build up at the neuroeffector junction because it prevents the reuptake of the norepinephrine back into the nerve endings. The action of cocaine is analogous to the effect of an anticholinesterase at the cholinergic synapse—it interferes with the mechanism for the prompt disposition of the chemical mediator. Cocaine has no direct action on the effector cell and does not release norepinephrine from the nerve ending. Cocaine

208

Function of the Peripheral Nervous System

does not block the physiologic release of norepinephrine from the nerve endings; it only blocks the norepinephrine reuptake mechanism so that norepinephrine accumulates at the effector junction until the muscle cell of the iris dilator fires continuously. If the nerve action potential is interrupted at any place in the sympathetic pathways, as in Horner’s syndrome, norepinephrine will not accumulate and the pupil will fail to dilate with cocaine drops. The duration of cocaine mydriasis is variable; it may last for more than 24 hours and there is no “rebound miosis.” No more than one drop of 10% cocaine or two drops of the 5% solution should be placed in each eye. More cocaine has a toxic effect on the corneal epithelium, and if too much of it is used, the patient may have an uncomfortable eye after the anesthetic effect wears off. Epinephrine (adrenalin) directly stimulates the receptor sites of the pupil dilator. When epinephrine is applied to the conjunctiva in a weak 1:1000 (0.1%) solution, it does not penetrate into the normal eye in sufficient quantity to have any obvious mydriatic effect. If there has been postganglionic denervation and the receptors have been made supersensitive, this weak solution may dilate the pupil, but the effect is not dependable. Phenylephrine (Neo-Synephrine) is also a directly active adrenergic mydriatic; its action is almost exclusively a direct ␣-adrenergic stimulation of the effector cell. The 1% solution has a moderate mydriatic effect; put in both eyes, it is a better test for adrenergic supersensitivity in postganglionic Horner’s syndrome than is 0.1% epinephrine. Both 5% and 2.5% solutions of phenylephrine are commonly used mydriatics. The pupil recovers in 8 hours and shows a “rebound miosis” lasting for several days. Higher concentrations such as 10% phenylephrine, when absorbed by the nasal mucosa, may produce hypertensive crises, especially in patients with

generalized autonomic neuropathy, if even two drops are placed in each eye.69 Tyramine and especially hydroxyamphetamine (Paredrine) act adrenergically by releasing norepinephrine from the postganglionic nerve endings. This appears to be their only effective action. This effect permits them to be used to separate postganglionic from central and preganglionic Horner’s syndrome.5,70,71 Botulinum toxin blocks the release of acetylcholine, and hemicholinium interferes with synthesis of acetylcholine at both the preganglionic and the postganglionic nerve endings, thus interrupting the parasympathetic pathway in two places. The outflow of sympathetic impulses also is interrupted by systemic doses of these drugs, since the chemical mediator in sympathetic ganglia is also acetylcholine.69

The Size of the Normal Pupil The pupil is not normally quite at the center of the iris but is slightly nasal and inferior. Pupil size varies with age (Fig. 9–5).28 Sympathetic innervation of the dilator muscle appears to be acquired at term. This is based on the observation that the pupils of some premature infants dilate in response to phenylephrine but not to hydroxyamphetamine.16 When a premature baby is awakened, the pupils are small (3.6 mm ⫾ 0.9 mm), according to Isenberg and co-workers,16 and they show no light reaction until a gestational age of 31 weeks is reached. The light reaction gradually improves until it is about 2 mm at term.47 The miosis in newborn babies is due to the fact that they sleep 22 hours out of every 24 and also because the eyeball is not full-grown. By adolescence

FIGURE 9–5 Pupillary size in darkness at various ages. The 1263 subjects were chosen at random from a population survey. Each point represents the average diameter of the two pupils taken together. The ordinate shows horizontal diameter (in millimeters); the abscissa shows age in years. Note the wide scatter among individuals and the obvious age trend. (From Loewenfeld, I. E.: Pupillary changes related to age. In Thompson, H. S. [ed.]: Topics in Neuro-Ophthalmology. Baltimore, Williams & Wilkins, p. 124, 1979, with permission.)

209

The Pupil

the pupils are at their largest. Thereafter they gradually become smaller until age 60, when size levels off.28

associated with hippus, but gradually it has become accepted as a normal phenomenon.60,69

Simple Anisocoria

Pupil Cycle Time

It is common to have a small amount of pupillary inequality. About 20% of the normal population have an easy-to-detect anisocoria of 0.4 mm or more at any given time (based on flash photos taken in dim light). This pupillary inequality is not constant in any given individual; it may increase or decrease or even reverse sides within days or hours. Smaller amounts of asymmetry are even more common. When subjects are examined twice daily for 5 days, 70% will have anisocoria of 0.3 mm or more at least once during this period (Fig. 9–6).24 “Simple” anisocoria decreases in bright light, has not been associated with any disease process, and produces no symptoms. The alternative terms physiologic anisocoria and essential anisocoria are also used. Since the light, dark, and near reactions in both eyes are normal, no further evaluation for a peripheral neuropathic cause for the anisocoria is warranted.

A small beam of light focused on the edge of the pupil induces regular oscillations of the pupil that one can time with a stopwatch. The period of the average complete cycle is called the pupil cycle time.34,68 This technique requires a slit lamp and has been used to look at the iris with sympathetic and parasympathetic denervation.3,31 In both cases the pupil cycle time is slowed. Although pupil cycle time testing may broadly point to an autonomic neuropathy that affects the innervation of the intraocular muscles, it is not useful in distinguishing sympathetic from parasympathetic damage. It is a labor intensive, nonspecific marker of iris muscle dysfunction.

Pupillary Unrest and Hippus The normal iris moves almost continuously, even when there is constant illumination and accommodation. This movement is known as physiologic pupillary unrest or hippus. It is thought to be due to moment-to-moment fluctuations in the activity of the sympathetic and parasympathetic innervation of the iris muscles. Hippus is most obvious in young patients with large pupils and is greatest in moderately bright light. The frequency of the oscillation increases with the intensity of the light. During the 19th century many neurologic diseases were

FIGURE 9–6 The amount of clinically visible anisocoria in a group of 128 normal subjects photographed morning and evening for 5 days. Anisocoria of 0.4 mm is easily discerned and is seen in 19% at any given time; 41% of the subjects had it all the time or some of the time. Thus anisocoria, which is clinically identifiable, is present in one in five patients at any moment, but this is not a fixed population. The mnemonic PERRLA (pupils equal, round, and reactive to light and accommodation) is inaccurate at least 20% of the time. (Data from Lam, B. L., Thompson, H. S., and Corbett, J.: Prevalence of simple anisocoria. Am. J. Ophthalmol. 104:69, 1987.)

Iris Damage and Its Effect on the Light, Near, and Drug Responses Blunt injury to the globe may cause a traumatic iridoplegia. Several factors seem to be at work: (1) the chamber angle may be recessed, tearing the branches of the short ciliary nerves that serve the iris muscle; (2) the sphincter muscle itself may be injured so that it will not constrict in response to pilocarpine; and (3) the short ciliary nerves may be damaged if the choroid is ruptured. These injuries may produce a segmental palsy of the sphincter, a light-near dissociation, and either an undersensitivity or oversensitivity of the sphincter to pilocarpine, depending on the exact nature of the damage. An attack of acute angle-closure glaucoma often produces a pupil of moderate size that fails to react to light or to pilocarpine. This is due to permanent ischemic damage to the iris muscles.

210

Function of the Peripheral Nervous System

EFFERENT PUPILLARY DEFECTS Defects of Parasympathetic Innervation Differential Diagnosis of Sphincter Dysfunction (Fig. 9–7) Dorsal midbrain damage rarely causes isolated pupillary defects. When it does, usually both pupils are affected, and there is often a light-near dissociation. The most common associated ocular motility disorder in this setting is the dorsal midbrain syndrome. Compression of the dorsal midbrain or an intrinsic lesion is the usual cause. Pupil abnormalities caused by ventral midbrain or fascicular lesions are also exceedingly rare in isolation, usually occurring with the other elements of ventral midbrain dysfunction. The pupil fibers are grouped superiorly and medially on the oculomotor nerve as it exits the brainstem, and rarely, small mesencephalic infarcts will affect pupil fibers alone. Tonic Pupil Damage to the ciliary ganglion or short ciliary nerves produces a very characteristic combination of signs consisting of the following: (1) a poor pupillary reaction to light that, under slit-lamp examination, can be seen to be a regional or sector palsy of the iris sphincter (Fig. 9–8);

(2) accommodative paresis, which is frequently incomplete and therefore also “segmental”; and (3) cholinergic supersensitivity of the denervated muscles. The usual recommendation of 0.1% pilocarpine to test for an Adie’s tonic pupil has been recently challenged by a study that recommends 0.0625% pilocarpine. This dose will constrict a tonic pupil but will not constrict a normal pupil. In a study of 11 patients with dilated pupils caused by oculomotor nerve palsy and dilated pupils resulting from Adie’s tonic pupil, no difference was seen between the two groups using 0.1% pilocarpine as the supersensitivity stimulus. Both demonstrated denervation supersensitivity, conclusively showing that this phenomenon clinically occurs in both preganglionic and postganglionic conditions. Denervation supersensitivity cannot be used to separate tonic pupils from those caused by third cranial nerve preganglionic damage because it is seen in both.25 Often the pupillary response to near vision is unusually strong and tonic. Patients who show this constellation of signs are said to have tonic pupils. A tonic pupil can result from any postganglionic, parasympathetic denervation of the intraocular muscles.29 This can, of course, be due to a local infection or injury or a widespread generalized or autonomic peripheral neuropathy, as with diabetes or alcoholism.56,58 If otherwise unexplained, it is called Adie’s syndrome.13,29,66

FIGURE 9–7 Pharmacology of the sympathetic and parasympathetic nerve pathways of the pupil.

The Pupil

211

FIGURE 9–8 A tonic pupil showing segmental constriction of the iris sphincter between 11:00 and 12:30 and 2:30 and 3:30. Partial segmental denervation produces, with hippus, the flickering so-called vermiform movements of the edge of the iris as it is held by illumination. The segments that constrict in response to light are not the same as those that constrict in response to a near stimulus.

Adie’s Syndrome Adie’s syndrome is probably the most common pathologic cause of anisocoria and unilateral poor light reaction. Almost all denervated sphincters in Adie’s syndrome have sectors of residual light reaction when examined under slit lamp.66 The majority of these patients are young women, but Adie’s pupil can be seen in both sexes and at any age. In children younger than 10 years, Adie’s pupil is almost always preceded by a history of chickenpox. The damage in Adie’s tonic pupil occurs in the ciliary ganglion or to the short posterior ciliary nerves. The reaction of the pupil to light is absent in 10% of patients. The reaction of the pupil to a near effort is usually slow and strong and is often tonic. When the patient is asked to refixate at a distance, the tonically constricted pupil redilates very slowly. This is true of any parasympathetically denervated pupil.

FIGURE 9–9 Depiction of unilateral (left eye) anisocoria seen in Adie’s syndrome. Initially, there is a large discrepancy between the two pupils that, over time, becomes less and less. In this example, the right eye then becomes involved at 31/2 years. The anisocoria becomes greater, but gradually both pupils become small, unreactive to light, and tonically reactive to a near stimulus. Involvement of the second eye occurs at a rate of 4% per year. (From Thompson, H. S. [ed.]: Topics in NeuroOphthalmology. Baltimore, Williams & Wilkins, 1979, with permission.)

The basic principles of parasympathetic denervation are the same whatever the cause: 1. Partial denervation produces segmental dysfunction of the iris sphincter depending on which motor units are preserved and which are denervated. 2. Reinnervation is likely to be disproportionately populated by ciliary muscle nerves, thereby producing a pupil that responds better to a near effort than it does to light. 3. While the affected pupil is larger than the normal pupil at first, it gradually becomes smaller until, in all but the brightest of light, it is smaller than that of the unaffected eye. This has been called a “little old Adie’s” (Fig. 9–9). If both eyes are affected, both pupils become small and either weakly or imperceptibly reactive to light. Thus

212

Function of the Peripheral Nervous System

they would seem to be indistinguishable from the light-near dissociation that is associated with syphilis. However a “little old Adie’s” has a tonic near response and the Argyll Robertson pupil does not.17 Drug-Induced and Toxic Mydriasis Atropinic drugs and plants that are rich in the belladonna alkaloids can all cause unilateral or bilateral pupil dilation and paralysis of accommodations. The features of an anticholinergic mydriasis are (1) a 360-degree paralysis of the iris sphincter (no sector loss such as that seen with neuropathic damage), and (2) no constriction in response to 1% pilocarpine eyedrops—a concentration that will vigorously constrict the normal and the denervated iris sphincter.71,72 An iron-containing foreign body in the globe can slowly leach ferric ions into the tissues and poison the intraocular muscles. The result is a dilated and unreactive pupil.35 None of the heavy metals that typically cause peripheral axonal neuropathies has been implicated in pupillary problems. Patients with such intoxication should be tested for sector palsies by careful slit-lamp examination. Light-Near Dissociation Sometimes a striking disparity can be seen between the pupillary response to light and the pupillary response to near vision.67 The light reflex arc may be damaged, but the near reaction pathway may be untouched. This is the case in which there is damage to the afferent limb of the pupillary arc, including the retina, optic nerve, chiasm, an optic tract, brachium of the superior colliculus, or pretectal nucleus. In contrast, the light reflex and the near response may both be damaged in the efferent pathway, and then the near reaction may be restored by aberrant regeneration. Since 97% of the cells in the ciliary ganglion are directed to the ciliary muscle, random regeneration of these cells will be sure to send some fibers to the iris sphincter.69 The pupillary near reaction is frequently forgotten in the course of routine examination. For a near response examination to be accurate, the patient must be able to see what he or she is looking at as a near target, and the study should be done with good room illumination. The patient should be given an accommodative target to look at—an object of interest or one with fine detail on it. Occasionally the response is better if another sensory input is used, such as clicking one’s fingernails. Proprioceptive input, such as using the patient’s own thumbnail, can sometimes elicit a near response when nothing else works. The examiner should watch for the eyes to converge since this reveals how hard the patient is trying. The near response is triggered by blurred or disparaged imagery but has a large volitional component, and verbal encouragement may help. Recognizing a Light-Near Dissociation. It can be difficult to know whether there is a light-near dissociation. When is the near response better than the light response?

When the clinician examines the patient with a pocket light, there are usually three levels of light available: darkness with a light shining tangentially on the pupils from below, room light, and room light with an additional bright light in the eyes. Having the patient look in the distance and shining a bright hand light in the eye three or four times, each time for only 1 or 2 seconds, reveals how small the pupils will become with just a light stimulus. A true light-near dissociation can be present only if the near response (tested in moderate light) is better than the best constriction that bright light can produce (Table 9–1).

The Pupil as a Measure of Autonomic Dysfunction Despite clinical observations and reports of studies of pupil defects in disease (Table 9–2), many problems arise when the pupils are used as an indicator of autonomic dysfunction. The chief difficulty is that the autonomic neuropathy being sought is a generalized process that is likely to affect both pupils. The clearest pupillary signs are the ones that contrast one pupil with the other. Unilateral denervation is easy to identify, but bilateral partial denervation is hard to distinguish from normal. There is hardly a test for Horner’s syndrome, for example, that does not involve comparing one eye with the other. Any test that depends on pupil size or pupillary inequality must contend with the fact that there is a certain amount of simple anisocoria in all of us; we are not, it seems, made with more precision than is necessary. This anisocoria can change from one minute to the next. In addition, one has to be certain that both irises are free of structural damage from prior disease or injury. The fact that pupil size varies with age also must be sidestepped. Both pupils tend to enlarge together with excitement and constrict together with drowsiness, so whenever possible one pupil should be used as a control. Table 9–1. Equipment Useful in Testing and Recording Pupil Function Polaroid CU-5 1:1 camera Filament-free, bright light source (Welch Allyn handle and Finhoff transilluminator) Magnifier with built-in millimeter rule scale to accurately measure pupil photos Accommodative targets Cocaine 4%–10%—ordinarily dispensed in small individual droppers by prescription. Must be disposed of after use. Hydroxyamphetamine 1% (Paredrine)—now available from Pharmics (2350 South Redwood Road, Salt Lake City, UT 84119; Fax: 801-972-4139) Phenylephrine 1% Pilocarpine 0.025% Slit lamp

The Pupil

Table 9–2. Peripheral Neuropathies with Pupil Involvement Acute pandysautonomia1,2,30,62,78,79 Adie’s syndrome13,29,65,66 Alcoholism61,64 Amyloidosis9,51 Botulism10,22,32 Charcot-Marie-Tooth disease20,52 Diabetes mellitus12,15,56–58 Giant cell arteritis6,45 Hereditary sensory neuropathy33 Holmes-Adie syndrome44,54,59 Landry-Guillain-Barré disease77 Miller Fisher variant19,38 Ross syndrome7,14,50 Sarcoidosis23 Shy-Drager syndrome55 Syphilis8,27

Any test using an autonomically acting drug must take into account the influence of the iris pigment, the integrity of the corneal surface and its innervation, the presumed completeness of the denervation, and the amount of reinnervation. Some of the simplest tests for the integrity of innervation of the iris muscles either are not available to the neurologist or are not in common use. For example, the high-magnification view of the iris afforded by a slit lamp provides a clear indication of segmental denervation of the iris sphincter and can also reveal signs of aberrant regeneration. A simple infrared-sensitive video of the pupils gives a valuable view of both pupils in the dark, in the light, and at near. For heavily pigmented irises, a video camera can make the difference between seeing and not seeing the pupillary signs.76 An iris suffering from denervation of its muscles is likely to be smaller in the dark than is appropriate for the patient’s age, and it will probably react relatively weakly to light. The trouble is that both eyes are likely to be affected, and it will be hard to be certain that they are abnormal. If a segment of the iris sphincter is palsied in response to light but still responds to near or to eye movement (a slit-lamp observation), that is evidence of denervation and aberrant reinnervation. If the pupil dilates poorly in the dark and has a clearly stronger reaction to a near stimulus than to light (seen best with infrared video), that indicates abnormality.

REFERENCES 1. Andersen, O., Lindberg, J., Modigh, K., and Reske-Nielsen, E.: Subacute dysautonomia with incomplete recovery. Acta Neurol. Scand. 48:510, 1972.

213

2. Appenzeller, O., and Kornfeld, M.: Acute pandysautonomia: clinical and morphologic study. Arch. Neurol. 29:334, 1973. 3. Blumen, S. C., Feiler-Ofry, V., and Korczyn, A. D.: The pupil cycle time in Horner’s syndrome. J. Clin. Neuro-ophthalmol. 6:232, 1986. 4. Carlson, D. W., and Tychsen, L.: Touching the cornea enhances pharmacologic dilation of the pupil, mainly in the dark iris. Aviat. Space Environ. Med. 60:994, 1989. 5. Cremer, S. A., Thompson, H. S., Digre, K. B., and Kardon, R. H.: Hydroxyamphetamine mydriasis in normal subjects. Am. J. Ophthalmol. 110:66, 1990. 6. Currie, J., and Lessell, S.: Tonic pupil with giant cell arteritis. Br. J. Ophthalmol. 68:135, 1984. 7. Esterly, N. B., Cantolino, S. J., Alte, B. P., and Brusilow, S. W.: Pupillotonia, hyporeflexia, and segmental hypohidrosis: autonomic dysfunction in a child. J. Pediatr. 73:852, 1968. 8. Fletcher, W. A., and Sharpe, J. A.: Tonic pupils in neurosyphilis. Neurology 36:188, 1986. 9. Frewin, D. B., Gilmore, H. R., Ho, J. Q., and Scroop, G. C.: Clinical, physiological and pathological observations in a case of progressive autonomic nervous system degeneration associated with Holmes-Adie syndrome and peripheral neuropathy. Australas. Ann. Med. 17:141, 1968. 10. Friedman, D. I., Fortanasce, V. N., and Sadun, A. A.: Tonic pupils as a result of botulism. Am. J. Ophthalmol. 109:236, 1990. 11. Friedman, J. R., Whiting, D. W., Kosmorsky, G. S., and Burde, R. M.: The cocaine test in normal patients. Am. J. Ophthalmol. 98:808, 1984. 12. Friedman, S. A., Feinberg, R., Podolak, E., and Bedell, R. H.: Pupillary abnormalities in diabetic neuropathy: a preliminary study. Ann. Intern. Med. 67:977, 1967. 13. Harriman, D. G., and Garland, H.: The pathology of Adie’s syndrome. Brain 91:401, 1968. 14. Hedges, T. R., and Gerner, E. W.: Ross’ syndrome (tonic pupil plus). Br. J. Ophthalmol. 59:387, 1975. 15. Hreidarsson, A. B.: Pupil size in insulin-dependent diabetes: relationship to duration, metabolic control, and long-term manifestations. Diabetes 31:442, 1982. 16. Isenberg, S. J., Dang, Y., and Jotterand, V.: The pupils of term and preterm infants. Am. J. Ophthalmol. 108:75, 1989. 17. Kardon, R. H., Corbett, J. J., and Thompson, H. S.: Segmental denervation and reinnervation of the iris sphincter as shown by infrared videographic transillumination. Ophthalmology 105:313, 1998. 18. Kardon, R. H., Denison, C. E., Brown, C. K., and Thompson, H. S.: Critical evaluation of the cocaine test in the diagnosis of Horner’s syndrome. Arch. Ophthalmol. 108:384, 1990. 19. Keane, J. R.: Tonic pupils with acute ophthalmoplegic polyneuritis. Ann. Neurol. 2:393, 1977. 20. Keltner, T. L., Swisher, C. N., Gay, A., and Hepler, R. S.: Myotonic pupils in Charcot-Marie-Tooth disease: successful relief of symptoms with 0.025% pilocarpine. Arch. Ophthalmol. 93:1141, 1975. 21. Kerrison, J. B., Biousse, V., and Newman, N. J.: Isolated Horner’s syndrome and syringomyelia. J. Neurol. Neurosurg. Psychiatry 69:3, 2000. 22. Konig, H., Gassman, H. B., and Jenzer, G.: Ocular involvement in benign botulism B. Am. J. Ophthalmol. 80:430, 1975.

214

Function of the Peripheral Nervous System

23. Kupersmith, M. J., and Aleksic, S. N.: Bilateral pupillary cholinergic supersensitivity in a case of sarcoidosis. Neuro-ophthalmology 4:15, 1984. 24. Lam, B. L., Thompson, H. S., and Corbett, J. J.: The prevalence of simple anisocoria. Am. J. Ophthalmol. 104:69, 1987. 25. Leavitt, J. A., Wayman, L. L., Hodge, D. O., and Brubaker, R. F.: Pupillary response to four concentrations of pilocarpine in normal subjects: application to testing for Adie tonic pupil. Am. J. Ophthalmol. 135:259, 2003. 26. Lepore, F. E.: Enophthalmos and Horner’s syndrome. Arch. Neurol. 40:460, 1983. 27. Loewenfeld, I. E.: The Argyll Robertson pupil, 1869–1969. A critical survey of the literature. Surv. Ophthalmol. 14:199, 1969. 28. Loewenfeld, I. E.: Pupillary changes related to age. In Thompson, H. S. (ed.): Topics in Neuro-Ophthalmology. Baltimore, Williams & Wilkins, p. 124, 1979. 29. Loewenfeld, I. E., and Thompson, H. S.: The tonic pupil: a re-evaluation. Am. J. Ophthalmol. 63:46, 1967. 30. Low, P. A., Dyck, P. J., Lambert, E. H., et al.: Acute panautonomic neuropathy. Ann. Neurol. 13:412, 1983. 31. Manor, R. S., Yassur, Y., Siegal, R., and Ben-Sira, I.: The pupil cycle time test: age variations in normal subjects. Br. J. Ophthalmol. 65:750, 1981. 32. Miller, N. R., and Moses, H.: Ocular involvement in wound botulism. Arch. Ophthalmol. 95:1788, 1977. 33. Miller, R. G., Nielsen, S. L., and Sumner, A. J.: Hereditary sensory neuropathy and tonic pupils. Neurology 26:931, 1976. 34. Miller, S. D., and Thompson, H. S.: Edge-light pupil cycle time. Br. J. Ophthalmol. 62:495, 1978. 35. Monteiro, M. L., Ulrich, R. F., Imes, R. K., et al.: Iron mydriasis. Am. J. Ophthalmol. 97:794, 1984. 36. Morris, J. G. L., Lee, L., and Kim, C. L.: Facial sweating in Horner’s syndrome. Brain 107:751, 1984. 37. Nielsen, P. J.: Upside down ptosis in patients with Horner’s syndrome. Acta Ophthalmol. 61:952, 1983. 38. Okajima, T., Imamura, S., Kawasaki, S., et al.: Fisher’s syndrome: a pharmacological study of the pupils. Ann. Neurol. 2:63, 1977. 39. Patil, P. M., and Jacobowitz, D.: Unequal accumulation of adrenergic drugs by pigmented and nonpigmented iris. Am. J. Ophthalmol. 78:470, 1974. 40. Patil, P. N.: Cocaine-binding by the pigmented and the nonpigmented iris and its relevance to the mydriatic effect. Invest. Ophthalmol. 11:739, 1972. 41. Pilley, S. F., and Thompson, H. S.: Pupillary “dilatation lag” in Horner’s syndrome. Br. J. Ophthalmol. 59:731, 1975. 42. Pomeranz, H.: Isolated Horner syndrome and syrinx of the cervical spinal cord. Am. J. Ophthalmol. 133:702, 2002. 43. Pruett, R. C.: Internal ophthalmoplegia after panretinal therapy. Arch. Ophthalmol. 97:2212, 1979. 44. Purcell, J. J. Jr., Krachmer, J. H., and Thompson, H. S.: Corneal sensation in Adie’s syndrome. Am. J. Ophthalmol. 84:496, 1977. 45. Rabinowich, L., and Mehler, M. F.: Parasympathetic pupillary involvement in biopsy-proven temporal arteritis. Ann. Ophthalmol. 20:400, 1988. 46. Richardson, R. W.: Comparing the mydriatic effect of tropicamide with respect to iris pigmentation. J. Am. Optom. Assoc. 53:885, 1982.

47. Robinson, J., and Fielder, A. R.: Pupillary diameter and reaction to light in preterm neonates. Arch. Dis. Child. 65:35, 1990. 48. Rogell, G. D.: Internal ophthalmoplegia after argon laser panretinal photocoagulation. Arch. Ophthalmol. 97:904, 1979. 49. Rosenberg, M. L.: The friction sweat test as a new method for detecting facial anhidrosis in patients with Horner’s syndrome. Am. J. Ophthalmol. 108:443, 1989. 50. Ross, A. T.: Progressive selective sudomotor denervation: a case with coexisting Adie’s syndrome. Neurology 8:809, 1958. 51. Rubinow, A., and Cohen, A. S.: Scalloped pupils in familial amyloid polyneuropathy. Arthritis Rheum. 29:445, 1986. 52. Salisachs, P., and Lapresle, J.: Argyll-Robertson-like pupils in the neural type of Charcot-Marie-Tooth disease. Eur. Neurol. 16:172, 1977. 53. Schidte, S. N.: Effects on choroidal nerves after panretinal xenon arc and argon laser photocoagulation. Acta Ophthalmol. 62:244, 1984. 54. Selhorst, J. B., Madge, G., and Ghatak, N.: The neuropathology of the Holmes-Adie syndrome. Ann. Neurol. 16:138, 1984. 55. Shy, G. M., and Drager, G. A.: A neurologic syndrome associated with orthostatic hypertension: a clinical pathologic study. Arch. Neurol. 2:511, 1960. 56. Sigsbee, B., Torkelson, R., and Kadis, G.: Parasympathetic denervation of the iris in diabetes mellitus: a clinical study. J. Neurol. Neurosurg. Psychiatry 37:1031, 1974. 57. Smith, S. A., and Smith, S. E.: Evidence for a neuropathic aetiology in the small pupil of diabetes mellitus. Br. J. Ophthalmol. 67:89, 1983. 58. Smith, S. E., Smith, S. A., Brown, P. M., et al.: Pupillary signs in diabetic autonomic neuropathy. Br. Med. J. 2:924, 1978. 59. Spector, R. H., and Bachman, D. L.: Bilateral Adie’s tonic pupil with anhidrosis and hyperthermia. Arch. Neurol. 41:342, 1984. 60. Stark, L., Campbell, F. W., and Atwood, J.: Pupil unrest: an example of noise in a biological servomechanism. Nature 182:857, 1958. 61. Tan, E. T., Lambie, D. G., Johnson, R. H., and Whiteside, E. A.: Parasympathetic denervation of the iris in alcoholics with vagal neuropathy. J. Neurol. Neurosurg. Psychiatry 47:61, 1984. 62. Thomashefsky, A. J., Horwitz, S. J., and Feingold, M. H.: Acute autonomic neuropathy. Neurology 22:251, 1972. 63. Thompson, B. M., Corbett, J. J., Lanning, B. K., and Thompson, H. S.: Pseudo-Horner’s syndrome. Arch. Neurol. 39:108, 1982. 64. Thompson, H. S.: Adie’s syndrome: some new observations. Trans. Am. Ophthalmol. Soc. 75:587, 1977. 65. Thompson, H. S.: Diagnosing Horner’s syndrome. Trans. Am. Acad. Ophthalmol. Otolaryngol. 83:840, 1977. 66. Thompson, H. S.: Segmental palsy of the iris sphincter in Adie’s syndrome. Arch. Ophthalmol. 96:1615, 1978. 67. Thompson, H. S.: Light-near dissociation of the pupil. Ophthalmologica 189:21, 1984. 68. Thompson, H. S.: The pupil cycle time. J. Clin. Neuroophthalmol. 7:38, 1987.

The Pupil 69. Thompson, H. S.: The pupil. In Moses, R. A., and Hart, W. (eds.): Adler’s Physiology of the Eye and Clinical Application. St. Louis, C. V. Mosby, 1992. 70. Thompson, H. S., Digre, K. B., Kardon, R. H., and Cremer, S. A.: Hydroxyamphetamine mydriasis in normal subjects [comment]. Am. J. Ophthalmol. 110:71, 1990. 71. Thompson, H. S., and Mensher, J. H.: Adrenergic mydriasis in Horner’s syndrome: hydroxyamphetamine test for diagnosis of postganglionic defects. Am. J. Ophthalmol. 72:472, 1971. 72. Thompson, H. S., Newsome, D. A., and Loewenfeld, I. E.: The fixed dilated pupil: sudden iridoplegia or mydriatic drops? A simple diagnostic test. Arch. Ophthalmol. 86:21, 1971. 73. Van der Wiel, H. L., and Van Gijn, J.: Horner’s syndrome: criteria for oculosympathetic denervation. J. Neurol. Sci. 56:293, 1982.

215

74. Van der Wiel, H. L., and Van Gijn, J.: Diagnosis of Horner’s syndrome: use and limitations of the hydroxyamphetamine test. J. Neurol. Sci. 73:311, 1986. 75. Van der Wiel, H. L., and Van Gijn, J.: No enophthalmos in Horner’s syndrome. J. Neurol. Neurosurg. Psychiatry 50:498, 1987. 76. Verdick, R. E., and Thompson, H. S.: Infrared videography of the eyes. J. Ophthalmol. Photog. 13:19, 1991. 77. Williams, D., Brust, J. C., Abrams, G., et al.: Landry-GuillainBarre syndrome with abnormal pupils and normal eye movements: a case report. Neurology 29:1033, 1979. 78. Yee, R. D., Trese, M., Zee, D. S., et al.: Ocular manifestations of acute pandysautonomia. Am. J. Ophthalmol. 81:740, 1976. 79. Young, R. R., Asbury, A. K., Adams, R. D., and Corbett, J. L.: Pure pan-dysautonomia with recovery. Trans. Am. Neurol. Assoc. 94:355, 1969.

10 Neural Control of Cardiac Function DAVID J. PATERSON AND JOHN H. COOTE

Organization of Cardiac Vagal and Sympathetic Nervous System Origin of Cardiac Sympathetic Tone Origin of Cardiac Parasympathetic Tone Cardiac Plexus and the “Cardiac Brain” Hypothesis Neurochemical Transmission in the CNS Concluding Remarks on the CNS

Chemical Transmission at the Cardiac Neuroeffector Junction History Postjunctional Signaling Communication between the Cardiac Sympathetic and Parasympathetic Systems Accentuated Antagonism

It is well established that the central and peripheral nervous systems play an important role in regulating the performance of the healthy and diseased heart. This regulation is far reaching, affecting excitability, contraction, blood flow, and cardiac metabolism. Many diseases of the cardiovascular system are also diseases of the autonomic nervous system (dysautonomia), where cardiac dysautonomias may actually influence morbidity and mortality associated with the primary disease itself.17 This chapter is concerned with the neural control of cardiac physiologic function and its relation to disorders of the autonomic nervous system. To appreciate the functional significance of the subject area requires some understanding as to how the cardiac nervous system is organized.

ORGANIZATION OF CARDIAC VAGAL AND SYMPATHETIC NERVOUS SYSTEM The healthy heart is subject to a stream of nerve impulse traffic in both vagus and sympathetic nerves. Resting heart rate (HR) is determined primarily by the magnitude of parasympathetic activity acting directly on the heart or interacting with the sympathetic nerves through several mechanisms in the brain and at the level of the heart. The tonic activity of both systems arises in the central nervous system (CNS) (Fig. 10–1).

Assessing Cardiac Neural Drive: Physiological Outcome Impairment of Cardiac Neural Control: Clinical Outcome

Origin of Cardiac Sympathetic Tone The sympathetic neurons supplying the heart lie in the intermediolateral cell column of gray matter in the upper thoracic segments (T4-T5) of the spinal cord. These fibers are relatively short as they emerge in the white rami communicantes to enter the thoracic sympathetic ganglia, because the ganglia are located close to the spinal cord. They then ascend into the neck via the paravertebral sympathetic chain. The synapse between the pre- and postganglionic cardiac neuron occurs somewhere in the upper thoracic and cervical ganglia. Postganglionic fibers pass gray rami communicantes before forming the cardiac nerves that innervate both pacemaking tissue and the ventricle. The right sympathetic nerves predominantly innervate the sinoatrial node (SAN)137 to increase discharge from the pacemaker (positive chronotrophy). The left sympathetic nerves innervate the atrioventricular node to increase conduction (dromotropism), the excitability of the bundle of His–Purkinje fiber conducting system, and contractility (inotropism) in the atrium and ventricle.83,137 Experimental animal studies suggest that cardiac sympathetic neurons are arranged as a longitudinal column of cells apparently distinct from other columns of sympathetic neurons with different end organ targets.119 They have extensive dendritic fields onto which converge numerous spinal cord and supraspinal afferents. A major input is that from the rostral ventrolateral medulla 217

218

Function of the Peripheral Nervous System

Cerebellum

Baroreceptors IX nerve

ICC

NTS NA

RVL

Brainstem

SAN

FIGURE 10–1 Neural map showing main afferent and efferent pathways that regulate cardiac excitability. The nucleus tractus solitarius (NTS) receives input from baroreceptors that is conveyed by glossopharyngeal (IXth cranial nerve) afferents. It connects with the nucleus ambiguus (NA), which projects via the vagus to the sinoatrial (SA) node of the heart. The connection from NTS to NA might not be direct, as indicated by the dashed projection. Also illustrated are the rostral ventrolateral medulla (RVL) and the intermediolateral cell column (ICC) of the spinal cord. The cerebellum is shown for orientation.

(RVLM), wherein again it is likely that cardiac sympathetic neurons exist as a distinct group.102 It is the input activity in the RVLM projection that most likely determines the level of tonic discharge in the preganglionic neurons.21 The consensus opinion is that this tonic activity is largely dependent on multiple oscillating networks of neurons in the lateral tegmentum of the medulla that fire with a 2- to 6-Hz rhythm that becomes entrained to the arterial baroreceptor input (which is inhibitory and in synchrony with the cardiac cycle).7 This activity is enhanced during inspiration, an effect that is injected at the level of the RVLM.21 Neurochemical transmission between RVLM neuron terminals and spinal sympathetic preganglionic neurons is via the release of glutamate acting on both N-methyl-D-aspartate (NMDA) and non-NMDA receptors.24 Inhibition by the arterial baroreceptor input is mediated by ␥-aminobutyric acid (GABA)151 released by neurons projecting to RVLM neurons from the caudal

ventrolateral medulla, where the neurons receive a direct projection from the nucleus tractus solitarius (NTS).110,139 A further projection from the NTS to GABA interneurons in the spinal cord to inhibit sympathetic preganglionic neurons has also been proposed.85 Numerous other chemical transmitters play a part in modifying the pattern and magnitude of discharge of sympathetic neurons, but little is known about specific influences on cardiac sympathetics. At the level of the RVLM, glutamate, angiotensin II, and vasopressin are excitatory, whereas GABA and nitric oxide (NO) are inhibitory.43 In the spinal cord serotonin, norepinephrine (NE), epinephrine, dopamine, oxytocin, and vasopressin may be released from supraspinal terminals. The monoamines can exert both excitatory and inhibitory actions either indirectly via interneurons or directly via different postsynaptic receptors.18,41 Supramedullary neurons project directly to the spinal cord and release peptides, possibly together with glutamate, which is excitatory. There is also evidence that oxytocin release selectively affects cardiac sympathetic neurons.158 The discharge of sympathetic neurons in the spinal cord is also dependent on NO, which mainly acts to reduce their discharge by causing the release of glycine from spinal interneurons.155,157 A summary of the neuronal pathways regulating cardiac sympathetic preganglionic neurons is illustrated in Figure 10–2.

Origin of Cardiac Parasympathetic Tone Preganglionic cardiac vagal efferents predominantly project from the nucleus ambiguus in humans, although there are also functional projections in many species from the dorsal motor nucleus of the vagus and intermediate zone of the medulla oblongata.57,67 These fibers do not diverge as extensively as those of the sympathetic division, and after exiting the brainstem travel in the Xth cranial nerve, descending in the neck caudally to the common carotid artery. Vagal preganglionic fibers synapse to intracardiac ganglia that are in close proximity to their neuroeffector site. These ganglia in humans are located in the posterior aspect of the atria within the subepicardial layers near the SAN and atrioventricular node. The pattern of vagal innervation is not as evenly distributed compared to the sympathetic innervation.65,71,137 Right-sided innervation is dominant given the location of the SAN as the primary pacemaker for the vagus to decrease excitability and rate (bradycardia). For many years it was thought there was no significant functional vagal innervation beyond the supraventricular structure in humans. This is simply not true.84 There is good evidence for innervation into the septum and subendocardial layers of the human left ventricle. Functionally, vagal stimulation can decrease excitability and contraction of the ventricle. These effects are accentuated in the presence of elevated sympathetic activity and independent of changes in HR brought about by vagal stimulation.86

219

Neural Control of Cardiac Function

NTS

Supramedullary and medullary spinal projecting neurons

Arterial baroreceptor

Lat. tegmental oscillator

Inspiratory N

+

Glutamate

RVLM

+

Cardiac presympathetic neuron

–

+

GABA CVLM

FIGURE 10–2 Schematic summarizing putative neuronal pathways and chemical neurotransmitters determining the tonic activity of cardiac sympathetic nerves. CVLM ⫽ caudal ventrolateral medulla; GABA ⫽ ␥-aminobutyric acid; NTS ⫽ nucleus tractus solitarius; RVLM ⫽ rostral ventrolateral medulla; ⫹, excitation; ⫺, inhibition.

Evidence from a variety of sources indicates that cardiac vagal preganglionic neurons are inherently silent, displaying no spontaneous pacemaker activity. Therefore, their discharge is determined by the activity and chemical transmitters released by neurons that synapse with them.40,104 A summary of the neuronal pathways regulating cardiac parasympathetic preganglionic neurons is illustrated in Figure 10–3. Chief among the excitatory inputs are those related to stimulation of arterial baroreceptors and chemoreceptors, whereas powerful phasic inhibitory inputs arise from central respiratory neurons and airway receptors. As a consequence, in anesthetized animals the firing pattern observed in single cardiac vagal efferent fibers has both a cardiac-related rhythm and a respiratory rhythm.23,76,96,117,136 As stated in the previous section, the effect of these inputs is the opposite to that exerted on cardiac sympathetic neurons. In an elegant series of intracellular experiments in anesthetized animals, the membrane potential of cardiac vagal preganglionic neurons revealed that they exhibit depolarizing potentials correlated with the cardiac cycle, interposed with hyperpolarizations during inspiration. It was shown that these neurons were

+

Monoamines peptides

+

Glutamate

Cardiac preganglionic neuron

–

GABA Heart

receiving inhibitory postsynaptic potentials at a time when the central inspiratory drive was greatest. The cells were most excitable immediately after end inspiration and relatively less excitable as expiration proceeded.40 These studies explained the observation in humans that the cardiac vagal neuron responsiveness to increasing carotid baroreceptor activity (by stretching the arteries with brief neck suction applied externally during controlled-frequency breathing) was greater during expiration than inspiration.32 This effect is referred to as respiratory “gating”95 and is probably the main cause of respiratory sinus arrhythmia.31 However, studies in a variety of vertebrates suggest that there may be two types of cardiac vagal preganglionic neurons: one that displays phasic activity that is cardiorespiratory related and important for coupling cardiac output to ventilation, and another that fires tonically and responds more to states of emergency.141,142 Similar types of cardiac vagal neurons may exist in mammals because, in an exquisite study in anesthetized cats, a population of cardiac vagal neurons was revealed that were only weakly or not at all affected by the respiratory gate.20 As well as arterial baroreceptor, chemoreceptor, and respiratory inputs, a number of other afferent systems can

220

Function of the Peripheral Nervous System

NTS

Lung stretch receptor NO

GABA

– Arterial baroreceptor Arterial chemoreceptor

NO

+

+ + NO

5-HT

NMDA Non NMDA

Inspiratory neuron

+ GABA

GABAA

Cardiac vagal neuron

–

M1

Heart

influence the excitability of cardiac vagal preganglionic neurons when situations demand (e.g., exercise). These are illustrated in the diagram in Figure 10–4.

Cardiac Plexus and the “Cardiac Brain” Hypothesis Although it is widely assumed that neurons in the brainstem and spinal cord are the sole source of cardiac neuronal control, the intrathoracic paravertebral ganglia (containing sympathetic efferent postganglionic neurons) and intrinsic cardiac ganglia (containing parasympathetic efferent postganglionic neurons) are now recognized as being more than efferent relay stations to the heart.5 At the base of the heart lie the cardiac plexuses that are formed by preganglionic and postganglionic sympathetic fibers. There are two divisions within each plexus that are closely interconnected with several small ganglia located within it. The ventral (superficial) portion is formed by the cardiac branch of the left superior cervical ganglion and the lower cervical cardiac branches of the left vagus nerve. The dorsal (deep) aspect is made up of the cardiac branches from the cervical and upper thoracic sympathetic ganglia and the vagus and recurrent laryngeal nerves.106

– Acetyl choline

FIGURE 10–3 Schematic summarizing putative neuronal pathways and chemical neurotransmitters determining the tonic activity of cardiac vagal nerves. GABA ⫽ ␥-aminobutyric acid; 5-HT ⫽ serotonin; M1 ⫽ muscarinic receptor; NMDA, nonNMDA ⫽ two types of glutamate receptor; NO ⫽ nitric oxide; NTS ⫽ nucleus tractus solitarius; ⫹, excitation; ⫺, inhibition.

The idea of a complex nervous system within the heart (“little brain in the heart”) has been recognized for some time, although its precise role is now only being fully appreciated. The principal cells in cardiac ganglia send their axons to pacemaker cells, conducting tissue, and the ventricles. Some ganglia act as interneurons and terminate within the ganglia, whereas others project to other ganglia in the intracardiac plexus. Armour and Ardell’s recent text reviews the role the cardiac neuraxis plays in processing cardiovascular sensory information in order to maintain optimal neural output.5 Whether these local networks of neurons are important during the neural remodeling that takes place with myocardial disease, aging, and neurogenic hypertension has not been firmly established. Nevertheless, the concept of local neural circuits presents an attractive proposition that deserves exploration.

NEUROCHEMICAL TRANSMISSION IN THE CNS In electrophysiologic experiments in rats in which the NTS region was stimulated and postsynaptic responses recorded, cardiac vagal neurons demonstrated excitatory postsynaptic currents mediated by glutamate.98 It seems

221

Neural Control of Cardiac Function

+ INSPIRATORY ACTIVITY

– –

Arterial chemoreceptors

+

– Arterial baroreceptors

+

–

Lung stretch receptors

Superior laryngeal receptors

Post-inspiratory neurons Muscle ergoreceptors

+

FIGURE 10–4 Diagram illustrating the interaction of a variety of peripheral sensory receptors in inspiratory activity and cardiac vagal activity as discussed in the text. ⫹, excitation; ⫺, inhibition.

+

likely that the physiologic role of this pathway is to mediate baroreceptor- and chemoreceptor-induced vagal bradycardia. NO is also likely to be an important mediator of excitation. In anesthetized rats microinjection of NO donors or substrates into the nucleus ambiguus results in a profound vagal bradycardia, and this effect can be prevented by prior microinjection of NO synthesis inhibitors into the same region.38,124 In anesthetized animals NO may be tonically released, because NO synthesis inhibitors microinjected into the nucleus ambiguus on their own lead to a small increase in HR.124 This effect of NO on cardiac vagal efferents is in contrast to its endogenous role in the NTS, the primary relay center for the baroreceptor afferents. Here stimulation of NO release leads to a reduction of a baroreceptor-elicited vagal bradycardia. This effect in the NTS is dependent on NO stimulating GABA release from interneurons that inhibit transmission in the baroreceptor pathway.115 GABA neurons are also located within the nucleus ambiguus, where they make synaptic contact with cardiac vagal preganglionic neurons.28,99 GABA applied in the vicinity of these neurons inhibits their firing. The GABAA receptor antagonist bicuculline prevents this action and also results in an increased cardiac vagal activity, suggesting that there is a tonically active GABA neuron input.66 The functional implication of this input is unclear, but it could mediate the GABAergic inhibitory effects on cardiac vagal neurons shown to occur on stimulation of the NTS.152 A physiologic role of this pathway may be to transmit the cardiac inhibitory action of lung stretch receptors, which have a primary relay in the NTS. A further possibility is that the GABA neurons transmit the inhibition of vagal tone that occurs during physical exercise.

–

–

+

+

+

–

CARDIAC VAGAL PREGANGLIONIC NEURON

Acetylcholine (ACh) has been identified as a putative inhibitory transmitter in the nucleus ambiguus. Its local application reduces the discharge of identified cardiac vagal preganglionic neurons, an action blocked by atropine, indicating a muscarinic receptor is involved.40 Atropine applied alone enhances the discharge of the neurons, particularly during inspiration when they are normally silent, so that there is a complete suppression of respiratory modulation. Such evidence confirms the idea that inspiratory neurons that are cholinergic are directly inhibiting cardiac vagal preganglionic neurons.40 These results also provide an explanation for the observation that very low doses of atropine given in both experimental animals and humans result in augmentation of cardiac vagal activity, probably by an action in the CNS.69,120 The straightforward explanation of the actions of ACh may be more complex than this, because recent in vitro studies in rat brain slices indicate that nicotine, a cholinergic agonist, enhances glutamate neurotransmission to activate cardiac vagal neurons.109 However, until in vivo studies are done, it is too early to discuss the functional significance of these data. Serotonin-containing nerve terminals make synaptic contact with cardiac vagal neurons, and it appears that serotonin excites these neurons via activation of 5-hydroxytryptamine (5-HT)1A receptors and 5-HT7 receptors. Functionally it seems that part of this input is utilized by cardiopulmonary receptors that, when stimulated, elicit a bradycardia as part of the classic BezoldJarisch reflex.152 Microinjections of enkephalins and opiate agonists into the nucleus ambiguus of dogs cause profound bradycardia,77 and intravenous or intracerebroventricular injections of these compounds excite cardiac vagal efferent fibers.61 It appears there is an endogenous opioid system

222

Function of the Peripheral Nervous System

responsible for this enhancing effect on cardiac vagal efferent activity, because recent studies show that low doses of opioid antagonists into the CNS, but not intravenously, block the decrease in HR that accompanies hemorrhagic shock.3 A direct facilitatory action of opioids on cardiac vagal preganglionic neurons was confirmed in a recent study of these neurons recorded in vitro in slices of the rat medulla. This effect of the endorphins is by an inhibitory modulation of voltage-gated calcium currents that secondarily may cause a reduction in calcium-activated potassium currents.63 The effect of endorphins may be further enhanced by a presynaptic action to cause a reduction in GABA release. The important role of GABA in regulating cardiac vagal preganglionic neuron activity is also emphasized by observing the actions of anesthetics.64 A variety of anesthetics cause a clear depression of activity in cardiac vagal efferent fibers, resulting in increased HR.60 Studies on single cardiac vagal neurons recorded in vitro indicate a possible mechanism for this effect. They show that pentobarbital increases the duration of spontaneous inhibitory postsynaptic currents at concentrations that are clinically relevant. This action is via the GABAA receptor because it is prevented if the receptor subunits are modified to make them insensitive.63 The GABAA receptor is also a target for the benzodiazepine drugs, which increase its sensitivity by binding to a specific protein subunit of the receptor.123 As a consequence, they augment the effect of GABA, and it has been shown that in humans intravenous benzodiazepines such as midazolam increase HR by decreasing cardiac vagal tone.37

Concluding Remarks on the CNS This account may give the impression that more details are known about the properties and organization of cardiac vagal preganglionic neurons than cardiac sympathetic neurons. This is because the former are easier to identify as a select group, whereas the characteristics of cardiac sympathetic neurons in general have to be derived from the much larger literature on the sympathetic nervous system. Nonetheless, we can deduce with reasonable certainty— and this has been shown experimentally—that the major afferent regulators of the two populations of cardiac neurons elicit reciprocal actions, although unusually both cardiac vagal and sympathetic activity can increase simultaneously under certain experimental conditions.74 Missing from this account is a description of the role of endogenous brain peptides such as vasopressin and angiotensin. Details of these are outside the scope of this account, but it should not be forgotten that a peptide such as angiotensin can indirectly inhibit cardiac vagal neurons144 and directly excite presympathetic neurons in the RVLM.87 These actions may go awry in disease such as hypertension or heart failure and could explain some of the beneficial

effects of angiotensin-converting enzyme inhibitors that might cross the blood-brain barrier to reduce the production of angiotensin, and hence increase cardiac vagal tone while also decreasing cardiac sympathetic tone.144

CHEMICAL TRANSMISSION AT THE CARDIAC NEUROEFFECTOR JUNCTION Because of its relative ease of access, the cardiac neuroeffector junction became the synapse where the concept of neurotransmission was proved, and in so doing established one of the major physiologic principles of the 20th century, which pharmacologists elegantly exploited for therapeutic purposes. It is worthwhile highlighting briefly the evolution of chemical transmission in a modern text to place recent findings in perspective.

History In the 17th century, Willis first described the vagi or “wandering nerves,” and observed a projection to the aortic arch and speculated that it “may react to changes in the pulse” because bilateral denervation caused “great trembling” in the dog heart. His pupil Richard Lower went on to show that vagal activation affected heartbeat, but it was not until the 1800s that Weber provided conclusive evidence for the right vagus causing cardiac inhibition. This effect could be inhibited by atropine and mimicked by muscarine, strongly suggesting that the vagus releases a chemical neurotransmitter. The Nobel Prize–winning work of Otto Loewi in 1921 proved this concept to be true. Stimulation of the right vagus nerve of an isolated frog heart decreased HR, and the transfer of the perfusate to a second frog heart produced bradycardia in the donor. The chemical substance in the perfusate responsible for this action was named “vagusstoff” and was identified as ACh by Dale in the 1930s. A similar fascinating path of discovery occurred in parallel for the sympathetic nervous system. In 1543 Vesalius distinguished the sympathetic nerves from the vagus, with Willis ascribing some cardiac function to them. However, it was not until Oliver and Schafer, in 1895, had discovered the action of adrenalin and Gaskell, in 1886, divided the nervous system into two distinct antagonistic systems that Langley, in 1898, coined the term autonomic nervous system. T. R. Elliot’s classic experiment in 1905 in Langley’s laboratory in Cambridge suggested that adrenaline or its immediate precursor might be the mediator at the sympathetic neuroeffector junction. This study paved the way for von Euler in 1946 to conclusively prove that norepinephrine (NE) was the sympathetic transmitter, for which he received the Nobel Prize for Medicine or Physiology in 1970.

Neural Control of Cardiac Function

We now know that postganglionic parasympathetic ACh and postganglionic sympathetic NE are packaged into vesicles, and, on depolarization of the nerve terminal, influx of calcium through voltage-gated calcium channels promotes their fusion to the neuronal membrane, causing the release of the transmitter. Almost immediately (2.5 s) after ACh is released it is hydrolyzed by acetylcholinesterase,89,93 whereas NE has a longer active half-life and must be taken up by the sympathetic varicosity by active processes.75 At rest the inhibitory action of the vagus on the SAN dominates over the excitatory action of the sympathetic.82 If vagal tone is removed following the muscarinic antagonist atropine, HR will increase from around 70 beats/min to 130 beats/min in healthy humans. ␤-Adrenergic blockade with propranolol will reduce HR to 58 beats/min78 (Fig. 10–5).

Postjunctional Signaling Acetylcholine diffuses across the synaptic cleft to bind to SAN cell muscarinic M2 receptors, coupled to inhibitory heterotrimeric G proteins. This brings about changes in a range of pacemaking currents that contribute to the spontaneous depolarization of these cells during diastole and the rhythmic generation of action potentials. The upstroke of the SAN action potential is carried by the L-type calcium current (ICaL)73,90 and repolarization by the delayed rectifier potassium current (IK).48,72,126 Deactivation of IK during diastole unmasks a background sodium current (IbNa), which

Vagus nerve Sympathetic varicosity NCa NCa

↑Ca2+ ↑Ca2+ NE +

+ Norepinephrine (NE) β1 M2

ACh

Acetylcholine (ACh)

Cardiac myocyte

HR

HR

Sympathetic stimulation

Vagal stimulation

FIGURE 10–5 Postganglionic sympathovagal signaling on cardiac myocyte. Diagram illustrates voltage activation of neuronal calcium channels, resulting in increased intracellular calcium (Ca2⫹) that facilitates the release of neurotransmitter to either excite (norepinephrine) or inhibit (acetylcholine) cardiac excitability.

223

contributes to diastolic depolarization.45 This is assisted by activation of several other currents, including the hyperpolarization-activated current (If), the sustained inward current (Ist), the T-type calcium current (ICaT), and current generated by the sodium-calcium exchanger (INCE).12,105,156 M2-receptor stimulation is coupled via G␤␥ to an inwardly rectifying potassium current (IKACh), increasing the open probability of the channel proteins GIRK1 and GIRK4.125 This hyperpolarizes SAN cells and increases the time taken to reach action potential threshold. The latency of response to increase R-R intervals is around 200 ms and is a direct result of activation of IKACh.159 These receptors are also coupled to G␣i2, which inhibits the activity of adenylate cyclase. If, one of many currents contributing to the diastolic depolarization, can be directly modulated by nucleotides, such as cyclic GMP (cGMP) and cyclic AMP (cAMP).25 By inhibiting adenylate cyclase and decreasing cAMP levels, acetylcholine reduces If by shifting its activation curve to more negative potentials.26,27 The decrease in cAMP also reduces protein kinase A–dependent phosphorylation of L-type calcium channels. It is not clear how changes in ICaL may alter HR. This may occur via changing intracellular calcium handling and therefore calcium-dependent currents and exchangers involved in diastolic depolarization, by altering action potential threshold, or, more controversially, by contributing to the diastolic depolarization itself.121 The role of these currents in normal pacemaking and its cholinergic modulation in different cells throughout the SAN is an area of ongoing debate, but they nevertheless contribute to the indirect modulation of HR during vagal activation via second messengers (Fig. 10–6).62 Released NE from postganglionic sympathetic varicosities binds to ␤-adrenergic receptors on both pacemaking cells and ventricular myocytes. ␤-Receptor stimulation of these cells is coupled to stimulatory G protein (Gs). When the agonist binds to the receptor, it displaces GDP with GTP and activates the stimulatory alpha subunit (G␣s-GTP) to promote an increase in adenylyl cyclase and the formation of cAMP. This cyclic nucleotide acts as a second messenger to modulate the inward calcium current via protein kinase A.138,154 This in turn facilitates Ca2⫹induced Ca2⫹ release from the sarcoplasmic reticulum (SR) to provide calcium for binding to contractile proteins to increase force of contraction.51 A rise in calcium will also increase HR because the upstroke of the pacemaker potential is calcium dependent.62 cAMP can act directly on If to increase its activation kinetics to more positive potentials, thereby increasing pacemaking. It also acts on phospholamban phosphorylation to increase calcium reuptake by the SR, thereby facilitating relaxation (lusitropic effect).50 Although NE activation of Gs can be directly coupled to ICaL, functional activation takes several (1 to 3) seconds compared to the millisecond activation of the vagus. This is because the sympathetic system

224

Function of the Peripheral Nervous System

ACh β1 Right vagus nerve

Gs

+

AC

Gi

–

M2

Gk

NOS-III Sinoatrial node cell

NO ? – PDE2

Sinoatrial node

PKA

cAMP

sGC cGMP +

PDE3 ICaL

IKACh

If

0 pA -500

+50

Sinoatrial node cell action potentials ACh

0 mV -50 -100

0

300 ms

FIGURE 10–6 Cholinergic modulation of sinoatrial node pacemaking currents. Stimulation of the right vagus nerve leads to the release of acetylcholine (ACh), which binds to muscarinic M2 receptors on sinoatrial node pacemaking cells. This decreases the rate of spontaneous action potential generation by direct G protein (Gk), gating of an inward rectifying potassium current (IKACh), and G protein (Gi) inhibition of adenylyl cyclase (AC) to decrease cAMP- and protein kinase A (PKA)–dependent stimulation of the hyperpolarization-activated current (If) and the L-type calcium current (ICaL). M2 receptor stimulation may also activate endothelial nitric oxide synthase (NOS-III), which increases cGMP levels via NO-dependent stimulation of soluble guanylyl cyclase (sGC). This may inhibit ICaL or If in the presence of high levels of ␤1-adrenergic receptor stimulation via cGMP-dependent stimulation of phosphodiesterase 2 (PDE2) and a decrease in cAMP, although this is controversial. (From Herring, N., Danson, E. J. F., and Paterson, D. J.: Cholinergic control of heart rate by nitric oxide is site specific. News Physiol. Sci. 17:204, 2002, with permission.)

predominantly mobilizes second messenger pathways to activate ion channels, whereas the functional action of the vagus does not rely on the involvement of second messenger phosphorylation pathways to decrease HR.

COMMUNICATION BETWEEN THE CARDIAC SYMPATHETIC AND PARASYMPATHETIC SYSTEMS The close proximity of sympathovagal postganglionic terminal varicosities allows for axoaxonic presynaptic communication. Both NE and ACh can inhibit their own release (autoinhibition) and reciprocally modulate each other’s release. Norepinephrine can bind presynaptically to ␣2 receptors to inhibit its own release by decreasing adenylyl cyclase–cAMP kinase activation of neuronal

calcium channels. Cross talk occurs when ACh binds to presynaptic M2 receptors on sympathetic varicosities to inhibit the release of NE, whereas NE can inhibit ACh release via prejunctional a receptors on vagal terminals.49,81,92,154 Putative cotransmitters have also been identified in both sympathetic and vagal terminals. Sympathetic varicosities contain neuropeptide Y, and vasoactive intestinal polypeptide (VIP) has been identified in vagal terminals.2,153 Because they do not fulfill all the traits of classic neurotransmitters, these peptides probably act as neuromodulators since they are not released in significant quantity unless high-frequency stimulation is performed. Functionally neuropeptide Y can inhibit the release of both NE and ACh94; conversely, VIP can cause a positive chronotropic and inotropic response. Indeed, VIP may be responsible for “vagal-induced tachycardia” when stimulation is stopped.153 ATP can also act on presynaptic

Neural Control of Cardiac Function

purinergic receptors to modulate release of transmitters. Emerging evidence suggests that the gaseous molecule NO is an important messenger because it acts both as a paracrine agent (cotransmission) and as an autocrine agent to modulate neurotransmission. Functionally, NO activates the second messenger cascade involving cGMP signaling to phosphodiesterase and protein kinases. Presynaptically NO inhibits the release of NE130 and decreases sympathetic-induced tachycardias.16 Conversely, it facilitates the release of ACh55 to accentuate vagal-induced bradycardia.54 NO may also have the capacity to diffuse to the postsynaptic cell to inhibit or excite pacemaking. Importantly, its differential action is now recognized as being isoform and site specific, because microdomains of its enzyme nitric oxide synthase (NOS) seem to be critical for encoding function.115

Accentuated Antagonism Both limbs of the cardiac autonomic nervous system act in a synergistic manner, resulting in the phenomenon of accentuated antagonism, whereby the negative chronotropic and inotropic action of the vagus is more pronounced when sympathetic tone is high.79,80 The mechanism underlying this probably involves NO-cGMP, although the exact location of its action is not yet firmly established. Some authors,46,47 although not all,146 support the idea that postjunctional NO is critical because NOS inhibition abolishes inhibition of ICaL by ACh analogues only when ␤-adrenergic tone is high in pacemaking cells. However, recent work suggests that the main functional role of accentuated antagonism occurs presynaptically where NO facilitates the release of ACh during nerve stimulation—effects not mimicked by bath-applied transmitter.115 Disruption of this pathway both anatomically and physiologically may be central to the vagal dysfunction that is seen in hypertension. Conversely, upregulation of peripheral parasympathetic signaling by NO-cGMP is a key component of enhanced vagal responsiveness brought about by physical training.22 Understanding this pathway is clearly important because impaired cardiac parasympathetic activity is a powerful negative prognostic indicator that can predict sudden cardiac death.17

ASSESSING CARDIAC NEURAL DRIVE: PHYSIOLOGIC OUTCOME At a given point in time, HR depends on the net excitatory and inhibitory action of sympathovagal balance. This is controlled by the CNS, which in turn depends on reflex inputs from cardiopulmonary receptors, arterial baroreflexes, and muscle ergoreceptors. Assessing cardiac neural drive noninvasively often relies on methods that can dissect out variabil-

225

ity in both HR and arterial pressure. Rapid changes in rate are predominately due to the vagus, with slower responses being attributed to the sympathetic.84,122 One simple measure is to study the effect of respiration on R-R interval. Inspiration is associated with a decrease and expiration with an increase in R-R interval (respiratory sinus arrhythmia). This response is abolished by atropine but unaffected by propranolol, suggesting involvement of vagal efferents.35,68 However, there is poor reproducibility of responses when respiratory frequency and tidal volume are high.4 Rapid changes in body posture8,35 as well as dynamic36 and isometric39 exercise can also be used to unmask the role of vagal efferents on HR variability. Quantification of HR variability using fast Fourier transform spectral analysis is useful in unmasking autonomic impairment in various pathophysiologic states.53,103,107,111,150 Two main peaks are identified with spectral analysis of cardiovascular waveforms: a high-frequency peak (0.2 to 0.4 Hz) that coincides with breathing frequency and is predominately vagal in origin, and a low-frequency fluctuation (0.03 to 0.15 Hz) that relates to sympathetic activity.53,113,116 Changes in the low-frequency/high-frequency ratio of the R-R interval variability are thought to reflect altered cardiac sympathovagal balance.10,19,52,128 Spectral analysis of cardiovascular waveforms from Holter monitoring studies have shown that patients with heart failure or diabetic neuropathy or following heart transplantation have low HR variability with an impaired high-frequency component (vagal dysfunction) and augmented low-frequency component (enhanced sympathetic activity).9–11,19,34,70,128 In patients with heart failure, augmented sympathetic responses have been correlated with enhanced circulating catecholamines and sympathetic activity (detected by microneurographic recordings).70 This approach has proved popular in the clinical environment and is often used as a prognostic indicator in patients suffering from several cardiac pathologies.11 However, there are pitfalls when using spectral analysis of HR variability during exercise; in particular, evidence indicates that, as the cardiac vagal tone falls with increasing levels of exercise, a greater percentage of the residual power of the high-frequency component may be due to non-neural mechanisms.14 Several methods are available to activate cardiac autonomic reflexes in humans. Infusion of vasoactive drugs that either increase or decrease vascular resistance can activate the vagal or sympathetic limbs of the arterial baroreflex, respectively. The advantage of this technique is that it is relatively easy to use and is minimally invasive, requiring an intravenous infusion or bolus of agents. Arterial pressure can be recorded directly via a catheter, although a Finipres gives a reasonably accurate record of changes in pressure.30,100,135 The main assumption in using vasoactive drugs is that they have no direct action on the heart itself. Phenylephrine or angiotensin has been used to raise blood pressure, although the latter can

226

Function of the Peripheral Nervous System

result in a small increase in HR and also cause the prejunctional release of NE.1 Nitrovasodilators, which release NO, are also not without problems; recent work has shown them to have a direct positive chronotropic effect.56,108 Interpretation of results using nitrovasoactive compounds must therefore be viewed with caution if they are infused over several minutes, because they do not provide an accurate quantitative assessment of cardiac neural drive.15 Furthermore, they only assess the cardiac component and not the vascular component of the reflex. The neck chamber technique overcomes many of the problems of vasoactive drugs. This ingenious approach that was developed by Ernsting and Parry in 1957 allows an airtight chamber applied to the subject’s neck to either increase or decrease carotid sinus transmural pressure, thereby activating or deactivating the high-pressure receptors.33 The downside is that the technique is relatively cumbersome and may not provide a discrete baroreceptor stimulus because transmission of negative pressure is incomplete, with only 65% of the applied pressure being sensed by the carotid arteries.29,98 Both head-up tilt and lower body negative pressure can displace significant blood volume to activate both low- and high-pressure receptors to test neural control of HR. Conversely, the Valsalva maneuver, cold pressure test, and isometric exercise can all be used to activate sympathetic efferent activity. These techniques are discussed in more detail in other chapters.

IMPAIRMENT OF CARDIAC NEURAL CONTROL: CLINICAL OUTCOME Is cardiac sympathovagal innervation clinically important? When the body needs to be in the rest-and-digest phase, the parasympathetic nervous system dominates; conversely, when alertness and activity are required, the sympathetic nervous system prevails. It is generally accepted that high sympathetic activity can facilitate the occurrence of ventricular arrhythmia, whereas vagal activation increases electrical stability and minimizes the incidences of ventricular tachyarrhythmias and fibrillation.118,138 Several pathophysiologic states or trauma can disrupt cardiac sympathovagal communications. Cardiac autonomic disorders can either be localized or generalized. Claude Bernard in 1854 probably described the first case of generalized autonomic failure when he showed that transection of the cervical spinal cord caused a marked fall in blood pressure as a result of loss of sympathetic vasoconstrictor tone, decreasing peripheral vasculature resistance. At the level of the heart, surgical (heart transplantation) or pharmacologic (double autonomic blockade with atropine and propranolol) denervation unmasks the intrinsic rate (approximately 90 beats/min) resulting from the inherent automaticity of the SAN,

which can function independently from the autonomic nervous system. Following heart transplantation or surgical sympathectomy, patients can experience localized autonomic abnormality that may result in life-threatening events, as in chronotropic denervation supersensitivity to adrenergic stimulation. Myocardial ischemia and infarction also create supersensitivity to catecholamines.59,131,147 These enhanced sympathoexcitatory responses lead to greater autonomic heterogeneity, increasing the propensity for arrhythmia and sudden cardiac death.147,160 Various peripheral afferent neuropathies, such as ascending polyneuritis (as in Guillain-Barré syndrome), porphyria, tabes dorsalis, and carotid sinus hypersensitivity, and efferent neuropathies such as diabetic autonomic neuropathy can result in fixed tachycardia at rest and sympathovagal imbalance as detected by power spectral analysis of heart variability.42,88,91,134 In many cases the high-frequency component of the power spectrum is absent, indicating vagal impairment with enhanced indices of sympathetic tone, all decreasing ventricular fibrillation threshold.6,44,97 In diabetics with autonomic neuropathy, there is a high incidence of sudden cardiac death, although cardiovascular reflexes are abnormal in this group before peripheral autonomic neuropathy appears.91 Whether the high mortality and morbidity are a direct consequence of dysautonomia affecting afferent and/or efferent limbs of the reflex arch or a defect in central processing has not been firmly established. Cardiac dysautonomia is also present in other disease states, such as heart failure, hypertension, ischemic heart disease, myocardial infarction, and Shy-Drager syndrome (neurologic manifestations involving striatonigral degeneration and Parkinsonian features, olivopontocerebellar atrophy, or the combination of the two, known as multiple system atrophy).101 Emerging evidence from the new subspeciality of neurocardiology strongly supports a major role for the CNS in the etiology of many cardiac arrhythmias via dysautonomia. Subarachnoid hemorrhage,114,145 cerebral vascular trauma especially to the prefrontal areas, and epilepsy145 may all cause profound alterations in cardiac rhythm. Stimulation of the defense area (hypothalamic region and amygdala) initiates arrhythmia in pigs,13,133 an effect blocked by propranolol.132 Stereotactic surgery in awake patients undergoing implantation of deep brain–stimulating electrodes for Parkinson’s disease has shown that activation of midbrain nuclei causes increases in HR and arterial blood pressure, indicating that complex functional circuits are directly coupled to autonomic outflow from this part of the brain.143 Excessive activation of the sympathetic nervous system can cause marked prolongation of the Q-T interval, as can localized central lesions.112,129,140,147 This facilitates the occurrence of afterdepolarizations caused by abnormal calcium loading in ventricular myocytes and can be amplified by hypokalemia (which further lengthens action potential

Neural Control of Cardiac Function

duration)149 and digitalis. When psychological stress,147 exercise, or even rapid-eye-movement sleep148—all events associated with high sympathetic activity—is experienced by a patient with long Q-T syndrome (disruption of the HERG protein that is responsible for the repolarization current IKr127), there is an increased incidence of potentially lethal ventricular arrhythmia. ␤-Adrenergic blockade minimizes the adverse affects of abnormal sympathetic activation to reduce sympathetic outflow and inhibit the action of catecholamines at the neuroeffector site. Interestingly, vagal stimulation and exercise training have been shown to be as effective as ␤-adrenergic blockade in protecting against the arrhythmogenic effect of myocardial ischemia.58 Indeed, high cardiac vagal tone and responsiveness, both characteristics of the athletic phenotype, are now becoming well-established independent positive prognostic markers against sudden cardiac death.17 Whether a centrally induced asymmetrical increase in cardiac sympathetic activity is involved in the etiology of sudden infant death syndrome and sudden adult death syndrome is not known.

REFERENCES 1. Abboud, F. M., and Chapleau, M. W.: Effects of pulse frequency on single unit activity during sine-wave and natural pulses in dogs. J. Physiol. (Lond.) 401:295, 1988. 2. Allen, J. M., Gjorstrup, P., Bjorkman, J. A., et al.: Studies on cardiac distribution and function of neuropeptide Y. Acta Physiol. Scand. 126:405, 1986. 3. Ang, K. K., McRitchie, R. J., Minson, J. B., et al.: Activation of spinal opioid receptors contributes to hypotension after haemorrhage in conscious rats. Am. J. Physiol. 276: H1552, 1999. 4. Angelone, A., and Coulter, N. A.: Respiratory sinus arrhythmia: a frequency dependent phenomenon. J. Appl. Physiol. 19:479, 1964. 5. Armour, J. A., and Ardell, J. L.: Preface. In Armour, J. A., and Ardell, J. L. (eds.): Basic & Clinical Neurocardiology. Oxford, UK, Oxford University Press, p. v, 2004. 6. Armour, J. A., Hageman, G. R., and Randall, W. C.: Arrhythmias induced by local cardiac nerve stimulation. Am. J. Physiol. 223:1068, 1972. 7. Barman, S. M., Gebber, G. L., and Orer, H. S.: Medullary lateral tegmental field: an important source of basal sympathetic nerve discharge in the cat. Am. J. Physiol. 278:R995, 2000. 8. Bellavere, F., and Ewing, D. J.: Autonomic control of the immediate heart rate response to lying down. Clin Sci. 62:57, 1982. 9. Bigger, J. T. Jr., Fleiss, J. L., Steinman, R. C., et al.: Correlation among time and frequency domain measures of heart rate period variability two weeks after acute myocardial infarction. Am. J. Cardiol. 69:891, 1992. 10. Bigger, J. T. Jr., Fleiss, J. L., Steinman, R. C., et al.: Frequency domain measures of heart period variability and mortality after myocardial infarction. Circulation 85:164, 1992.

227

11. Bigger, J. T. Jr., La Rovere, M. T., Steinman, R. C., et al.: Comparison of baroreflex sensitivity and heart period variability after myocardial infarction. J. Am. Coll. Cardiol. 14:1511, 1989. 12. Brown, H., and Difrancesco, D.: Voltage-clamp investigations of membrane currents underlying pace-maker activity in rabbit sino-atrial node. J. Physiol. (Lond.) 308:331, 1980. 13. Carpeggiani, C., Landisman, C., Montaron, M. F., and Skinner, J. E.: Cryoblockade in the limbic brain (amygdala) prevents or delays ventricular fibrillation after coronary artery occlusion in psychologically stressed pigs. Circ. Res. 70:600, 1992. 14. Casadei, B., Moon, J., Johnston, J., et al.: Is respiratory sinus arrhythmia a good index of cardiac vagal tone in exercise? J. Appl. Physiol. 81:556, 1996. 15. Casadei, B., and Paterson, D. J.: Should we still use nitrovasodilators to test baroreflex sensitivity? J. Hypertens. 18:3, 2000. 16. Choate, J. K., and Paterson, D. J.: Nitric oxide inhibits the positive chronotropic and inotropic responses to sympathetic nerve stimulation in the isolated guinea-pig atria. J. Auton. Nerv. Syst. 75:100, 1999. 17. Cole, C. R., Blackstone, E. H., Pashkow, F. J., et al.: Heartrate recovery immediately after exercise as a predictor of mortality. N. Engl. J. Med. 1341:1351, 1999. 18. Coote, J. H.: The organisation of cardiovascular neurones in the spinal cord. Rev. Physiol. Biochem. Pharmacol. 110:148, 1988. 19. Cripps, T. R., Malik, M., Farrell, T. G., and Camm, A. J.: Prognostic value of reduced heart rate variability after myocardial infarction: clinical evaluation of a new analysis method. Br. Heart J. 65:14, 1991. 20. Daly, M de B., and Kirkman, E.: Differential modulation by pulmonary stretch afferents of some reflex cardioinhibitory responses in the cat. J. Physiol. (Lond.) 417:323, 1989. 21. Dampney, R. A. L.: Functional organisation of central pathways regulating the cardiovascular system. Physiol. Rev. 74:323, 1994. 22. Danson, E. J. F., and Paterson, D. J.: Enhanced neuronal nitric oxide synthase expression is central to cardiac vagal phenotype in exercise-trained mice J. Physiol. (Lond.) 546:225, 2003. 23. Davidson, N. S., Goldner, S., and McCloskey, D. I.: Respiratory modulation of baroreceptor and chemoreceptor reflexes affecting heart-rate and cardiac vagal efferent nerve activity. J. Physiol. (Lond.) 259:523, 1976. 24. Deuchars, S. A., Morrison, S. F., and Gilbey, M. P.: Medullaryevoked EPSPs in neonatal rat sympathetic preganglionic neurones in vitro. J. Physiol. (Lond.) 487:453, 1995. 25. DiFrancesco, D., and Tortora, P.: Direct activation of cardiac pacemaker channels by intracellular cyclic AMP. Nature 351:145, 1991. 26. DiFrancesco, D., and Tromba, C.: Inhibition of the hyperpolarization-activated current (If) induced by acetylcholine in rabbit sino-atrial node myocytes. J. Physiol. (Lond.) 405:477, 1988. 27. DiFrancesco, D., and Tromba, C.: Muscarinic control of the hyperpolarization-activated current (If) in rabbit sino-atrial node myocytes. J. Physiol. (Lond.) 405:493, 1988.

228

Function of the Peripheral Nervous System

28. DiMicco, J. A., Gale, K., Hamilton, B. L., and Gillis, R. A.: GABA receptor control of parasympathetic outflow to heart: characterisation and brainstem location. Science 204:1106, 1979. 29. Eckberg, D. L.: Temporal response patterns of the human sinus node to brief carotid baroreceptor stimuli. J. Physiol. (Lond.) 259:769, 1976. 30. Eckberg, D. L.: Parasympathetic cardiovascular control in human disease: a critical review of methods and results. Am J. Physiol. 239:H581, 1980. 31. Eckberg, D. L.: The human respiratory gate. J. Physiol. (Lond.) 548:339, 2003. 32. Eckberg, D. L., and Orsham, C. R.: Respiratory and baroreceptor reflex interactions in man. J. Clin. Invest. 59:780, 1977. 33. Ernsting, J., and Parry, D. J.: Some observations on the effects of stimulating the stretch receptors in the carotid artery of man. J. Physiol. (Lond.) 137:45P, 1957. 34. Ewing, D. J., Campbell, I. W., Murray, A., et al.: Immediate heart-rate response to standing: simple test for autonomic neuropathy in diabetes. Br. Heart J. 1:145, 1978. 35. Ewing, D. J., and Clarke, B. F.: Diagnosis and management of diabetic autonomic neuropathy. Br. Med. J. 285:916, 1982. 36. Fagraeus, L., and Linnarsson, D.: Autonomic origin of heart rate fluctuations at the onset of muscular exercise. J. Appl. Physiol. 40:679, 1976. 37. Farmer, M. R., Vaile, J. C., Osman, F., et al.: A central ␥-aminobutyric acid mechanism in cardiac vagal control in man revealed by studies with intravenous midazolam. Clin. Sci. 95:241, 1998. 38. Fletcher, J., Moody, W. E., Chowdhary, S. C., and Coote, J. H.: Enhancement of the baroreceptor heart rate reflex by nitric oxide in the nucleus ambiguus of anaesthetised rats [abstract]. J. Physiol. (Lond.) 544:33P, 2002. 39. Friedman, D. B., Jensen, F. B., Mitchell, J. H., and Secher, N. H.: Heart rate and arterial blood pressure at the onset of static exercise in man with complete neural blockade. J. Physiol. (Lond.) 423:543, 1990. 40. Gilbey, M. P., Jordan, D., Richter, D. W., and Spyer, K. M.: Synaptic mechanisms involved in the inspiratory modulation of vagal cardio-inhibitory neurones in the cat. J. Physiol. (Lond.) 356:65, 1984. 41. Gladwell, S. J., and Coote, J. H.: Inhibitory and indirect excitatory effects of dopamine on sympathetic preganglionic neurones in the neonatal rat spinal cord in vitro. Brain Res. 818:397, 1999. 42. Greenland, P., and Griggs, R. C.: Arrhythmic complications in the Guillain-Barré syndrome. Arch. Intern. Med. 140:1053, 1980. 43. Guyenet, P. G., and Stornetta, R. L.: The presympathetic cells of the rostral ventrolateral medulla (RVLM): anatomy, physiology and role in the control of the circulation. In Dun, N. J., Machado, B. H., and Pilowsky, P. M. (eds.): Neural Mechanisms of Cardiovascular Regulation. Boston, Kluver, p. 187, 2004. 44. Hageman, G. R., Goldberg, J. M., Armour, J. A., and Randall, W. C.: Cardiac dysrhythmias induced by autonomic nerve stimulation. Am. J. Cardiol. 32:823, 1973. 45. Hagiwara, N., Irisawa, H., Kasanuki, H., and Hosoda, S.: Background current in sino-atrial cells of the rabbit heart. J. Physiol. (Lond.) 448:53, 1992.

46. Han, X., Shimoni, Y., and Giles, W. R.: An obligatory role for nitric oxide in autonomic control of mammalian heart rate. J. Physiol. (Lond.) 476:309, 1994. 47. Han, X., Shimoni, Y., and Giles, W. R.: A cellular mechanism for nitric oxide-mediated cholinergic control of mammalian heart rate. J. Gen. Physiol. 106:45, 1995. 48. Harris, J. M., and Hutter, O. F.: The action of acetylcholine on the movements of potassium in the sinus venosus of the heart [abstract]. J. Physiol. (Lond.) 133:8, 1956. 49. Hartzell, H. C.: Distribution of muscarinic acetylcholine receptors and presynaptic nerve terminals in amphibian heart. J. Cell. Biol. 86:6, 1980. 50. Hartzell, H. C., and Fischmeister, R.: Direct regulation of cardiac Ca2⫹ channels by G proteins: neither proven nor necessary? Trends Pharmacol. Sci. 13:380, 1992. 51. Hartzell, H. C., Mery, P. F., and Fischmeister, R.: Sympathetic regulation of cardiac calcium current is due exclusively to cAMP-dependent phosphorylation. Nature 351:573, 1991. 52. Hayano, J., Sakakibara, Y., Yamada, M., et al.: Decreased magnitude of heart rate spectral components in coronary artery disease: its relation to angiographic severity. Circulation 81:1217, 1990. 53. Hayano, J., Sakakibara, Y., Yamada, A., et al.: Accuracy of assessment of cardiac vagal tone by heart rate variability in normal subjects. Am. J. Cardiol. 67:199, 1991. 54. Herring, N., Golding S., and Paterson, D. J.: Pre-synaptic NO-cGMP pathway modulates vagal control of heart rate in isolated guinea-pig atria. J. Mol. Cell. Cardiol. 32:1795, 2000. 55. Herring, N., and Paterson D. J.: NO-cGMP pathway facilitates acetylcholine release and bradycardia during vagal nerve stimulation in the guinea pig in-vitro. J. Physiol. (Lond.) 535:507, 2001. 56. Hogan, N. A., Casadei, B., and Paterson, D. J.: Nitric oxide donors can activate heart rate independent of autonomic activation. J. Appl. Physiol. 87:97, 1999. 57. Hopkins, D. A.: The dorsal motor nucleus of the vagus nerve and the nucleus ambiguus: structure and connections. In Hainsworth, R., McWilliam, P. N., and Mary, D. A. S. G. (eds.): Cardiogenic Reflexes. New York, Oxford University Press, p.185, 1987. 58. Hull, S. S., Vanoli, E., Adamson, P. B., et al.: Exercise training confers anticipatory protection from sudden death during acute myocardial ischaemia. Circulation 89:548, 1994. 59. Inoue, H., and Zipes, D. P.: Results of sympathetic denervation in the canine heart: supersensitivity that may be arrhythmogenic. Circulation 75:877, 1987. 60. Inoue, K., and Arndt, J. O.: Efferent vagal discharge and heart rate in response to methohexitone, althesin, ketamine and etomidate in cats. Br. J. Anaesth. 54:1105, 1982. 61. Inoue, K., Nashan, B., and Arndt, J. O.: (D-met2,Pro5) enkephalinamide activates cardioinhibitory efferents in anaesthetised dogs. Eur. J. Pharmacol. 110:233,1985. 62. Irisawa, H., Brown, H. F., and Giles, W.: Cardiac pacemaking in the sinoatrial node. Physiol. Rev. 73:197, 1993. 63. Irnaten, M., Aicher, S. A., Wang, J., et al.: Mu-opioid receptors are located postsynaptically and endorphin-1 inhibits voltage-gated calcium currents in premotor cardiac parasympathetic neurones in the rat nucleus ambiguus. Neuroscience 116:573, 2003.

Neural Control of Cardiac Function 64. Irnaten, M., Walwyn, W. M., Wang, J., et al.: Pentobarbital enhances GABAergic neurotransmission to cardiac parasympathetic neurons, which is prevented by expression of GABAA epsilon sub unit. Anesthesiology 97:717, 2002. 65. James, T. N., and Spence, C. A.: Distribution of cholinesterase within the sinus node and AV node of the human heart. Anat. Rec. 155:151, 1966. 66. Jordan, D., and Spyer, K. M.: Central neural mechanisms mediating respiratory-cardiovascular interactions. In Taylor, E. W. (ed.): Neurobiology of the Cardiorespiratory System. Manchester, UK, Manchester University Press, p. 322, 1987. 67. Kalia, M.: Brain stem localization of vagal preganglionic neurons. J. Auton. Nerv. Syst. 3:451, 1981. 68. Katona, P., and Jih, F.: Respiratory sinus arrhythmia: noninvasive measure of parasympathetic cardiac control. J. Appl. Physiol. 39:801, 1975. 69. Katona, P. G., Lipson, D., and Dauchot, P. J.: Opposing central and peripheral effects of atropine on parasympathetic cardiac control. Am. J. Physiol. 232:H146, 1997. 70. Kienzle, M. G., Ferguson, D. W., Birkett, C. L., et al.: Clinical, hemodynamic and sympathetic neural correlates of heart rate variability in congestive heart failure. Am. J. Cardiol. 69:761, 1992. 71. King, T. S., and Coakley, J. B.: The intrinsic nerve cells of the cardiac atria of mammals and man. J. Anat. 92:353, 1958. 72. Kodama, I., Boyett, M. R., Nikmaram, M. R., et al.: Regional differences in effects of E-4031 within the sinoatrial node. Am. J. Physiol. 276:H793, 1999. 73. Kodama, I., Nikmaram, M. R., Boyett, M. R., et al.: Regional differences in the role of the Ca2⫹ and Na⫹ currents in pacemaker activity in the sinoatrial node. Am. J. Physiol. 272: H2793, 1997. 74. Kollai, M., and Koizumi, K.: Cardiovascular reflexes and interrelationships between sympathetic and parasympathetic activity. J. Auton. Nerv. Syst. 4:135, 1981. 75. Kopin, I. J., Hertting, G., and Gordon, E. K.: Fate of norepinephrine-H3 in the isolated perfused rat heart. J. Pharmacol. Exp. Ther. 138:34, 1962. 76. Kunze, D. L.: Reflex discharge patterns of cardiac vagal efferent fibres. J. Physiol. (Lond.) 222:1, 1972. 77. Laubie, M, Schmitt, H., and Vincent, M.: Vagal bradycardia produced by microinjections of morphine-like drugs into the nucleus ambiguus in anaesthetised dogs. Eur. J. Pharmacol. 59:287, 1979. 78. Leon, D. F., Shaver, J. A., and Leonard, J. J.: Reflex heart rate control in man. Am. Heart J. 80:729, 1970. 79. Levy, M. N.: Sympathetic-parasympathetic interactions in the heart. Circ. Res. 29:437, 1971. 80. Levy, M. N.: Cardiac sympathetic-parasympathetic interactions. Fed. Proc. 43:2598, 1984. 81. Levy, M. N., and Blattberg, B.: Effect of vagal stimulation on the overflow of norepinephrine into the coronary sinus during cardiac sympathetic nerve stimulation in the dog. Circ. Res. 38:81, 1976. 82. Levy, M. N., and Martin, P. J.: Neural control of the heart. In Berne, R. M., and Sperelakis, N. (eds.): Handbook of Physiology, Sect. 2. The Cardiovascular System, Vol. 1. The Heart. Bethesda, MD, American Physiological Society, p. 581, 1979.

229

83. Levy, M. N., Ng, M. L., and Zieske, H.: Functional distribution of the peripheral cardiac sympathetic pathways. Circ. Res. 19:650, 1966. 84. Levy, M. N., and Schwartz, P. J. (eds.): Vagal Control of the Heart: Experimental Basis and Clinical Implications. New York, Futura, 2004. 85. Lewis, D. I., and Coote, J. H.: Chemical mediators of spinal inhibition of rat sympathetic neurones on stimulation in the nucleus tractus solitarii. J. Physiol. (Lond.) 486:48, 1995. 86. Lewis, M. E., Al-Khalidi, A. H., Bonser, R. S., et al.: Vagus nerve stimulation decreases left ventricular contractility in vivo in the human and pig heart. J. Physiol. (Lond.) 534:547, 2001. 87. Li, Y. W., and Guyenet, P. G.: Angiotensin II decreases a resting K⫹ conductance in rat bulbospinal neurons of the C1 area. Circ. Res. 78:274, 1996. 88. Lichtenfield, P.: Autonomic dysfunction in the GuillainBarré syndrome. Am. J. Med. 50:772, 1971. 89. Lindmar, R., Löffelholz, K., and Weide, W.: Interstitial washout and hydrolysis of acetylcholine in the perfused heart. Naunyn Schmiedebergs Arch. Pharmacol. 318:295, 1982. 90. Lipsius, S. L., and Vassalle, M.: Dual excitatory channels in the sinus node. J. Mol. Cell. Cardiol. 10:753, 1978. 91. Lloyd-Mostyn, R. H., and Watkins, P. J.: Defective innervation of heart in diabetic autonomic neuropathy. Br. Med. J. 3:15, 1975. 92. Löffelholz, K., and Muscholl, E.: A muscarinic inhibition of the noradrenaline release evoked by postganglionic sympathetic nerve stimulation. Naunyn Schmiedebergs Arch. Pharmacol. 265:1, 1969. 93. Löffelholz, K., and Pappano, A. J.: The parasympathetic neuroeffector junction of the heart. Pharmacol. Rev. 37:1, 1985. 94. Löffelholz, K., and Weide, W.: Aminopyridines and the release of acetylcholine. Trends Pharmacol. Sci. 3:147, 1982. 95. Lopes, O. U., and Palmer, J. H.: Proposed respiratory ‘gating’ mechanism for cardiac slowing. Nature 264:454, 1976. 96. Lovick, T. A.: Differential control of cardiac and vasomotor activity by neurones in nucleus paragigantocellularis lateralis in the cat. J. Physiol. (Lond.) 389:23, 1987. 97. Malliani, A., Schwartz, P. J., and Zanchetti, A.: Neural mechanisms in life-threatening arrhythmias. Am. Heart J. 100:705, 1980. 98. Mancia, G., Ferrari, A., Gregorini, L., et al.: Circulatory reflexes from carotid and extracarotid baroreceptor areas in man. Circ. Res. 41:309, 1977. 99. Maqbool, A., Batten, T. F. C., and McWilliam, P. N.: Ultrastructural relationship of GABAergic terminals and cardiac vagal preganglionic motoneurones and vagal afferents in cat: combined HRP tracing-immunogold labelling study. Eur. J. Neurosci. 3:501, 1991. 100. Mary, D. A. S. G., and Hainsworth, R.: Methods for the study of cardiovascular reflexes. In Hainsworth, R., and Mark, A. L. (eds.): Cardiovascular Reflex Control in Health and Disease. London, W. B. Saunders, p. 1, 1993. 101. Mathias, C. J.: Disturbances of cardiovascular control in autonomic disorders. In Hainsworth, R., and Mark, A. L. (eds.): Cardiovascular Reflex Control in Health and Disease. London, W. B. Saunders, p. 425, 1993.

230

Function of the Peripheral Nervous System

102. McAllen, R. M., and Dampney, R. A. L.: Vasomotor neurones in the rostral ventrolateral medulla are organised topographically with respect to type of vascular bed but not body region. Neurosci. Lett. 110:91, 1990. 103. McAreavey, D., Neilson, J. M. M., Ewing, D. J., and Russell, D. C.: Cardiac parasympathetic activity during the early hours of acute myocardial infarction. Br. Heart J. 62:165, 1989. 104. Mendelowitz, D.: Firing properties of identified parasympathetic cardiac neurons in nucleus ambiguus. Am J Physiol 271:H2609–H2614, 1996. 105. Mitsuiye, T., Shinagawa, Y., and Noma, A.: Sustained inward current during pacemaker depolarization in mammalian sinoatrial node cells. Circ. Res. 87: 88, 2000. 106. Mizeres, N. J.: The cardiac plexus in man. Am. J. Anat. 112:141, 1963. 107. Murray, A., Ewing, D. J., Campbell, I. W., et al.: RR interval variations in young male diabetics. Br. Heart J. 37:882, 1975. 108. Musialek, P., Lei, M., Brown, H. F., et al.: Nitric oxide can increase heart rate by stimulating the hyperpolarizationactivated inward current If. Circ. Res. 81:60, 1997. 109. Neff, R. A., Humphrey, J., Mihalevich, M., and Mendelowitz, D.: Nicotine enhances presynaptic and postsynaptic glutaminergic neurotransmission to activate cardiac parasympathetic neurons. Circ. Res. 83:1241, 1998. 110. Neff, R. A., Mihalevich, M., and Mendelowitz, D.: Stimulation of NTS activates NMDA and non-NMDA receptors in rat cardiac vagal neurones in the nucleus ambiguus. Brain Res. 792:277, 1998. 111. Oppenheim, A. V., and Schafer, R. W.: Digital Signal Processing. Englewood Cliffs, NJ, Prentice Hall, 1975. 112. Oppenheimer, S. M., Cechetto, D. F., and Hachinski, V. C.: Cerebrogenic cardiac arrhythmias: cerebral electrocardiographic influences and their role in sudden death. Arch. Neurol. 47:513, 1990. 113. Pagani, M., Lombardi, F., Guzzetti, S., et al.: Power spectral analysis of heart rate and arterial pressure variabilities as a marker of sympatho-vagal interaction in man and conscious dog. Circ. Res. 59:178, 1986. 114. Parizel, G.: On the mechanism of sudden death with subarachnoid hemorrhage. J. Neurol. 220:71, 1979. 115. Paton, J. F. R., Kasparov, S., and Paterson, D. J.: Site-specific differential modulation of cardiac autonomic control by nitric oxide. Trends Neurosci. 25:626, 2002. 116. Pomeranz, B., Macaulay, R. J. B., Caudill, M. A., et al.: Assessment of autonomic function in humans by heart rate spectral analysis. Am. J. Physiol. 248:H151, 1985. 117. Potter, E. K.: Inspiratory inhibition of vagal responses to baroreceptor and chemoreceptor stimuli in the dog. J. Physiol. (Lond.) 316:177, 1981. 118. Prystowsky, E. N., Jackman, W. M., Rinkenberger, R. L., et al.: Effect of autonomic blockade on ventricular refractoriness and atrioventricular nodal conduction in humans: evidence supporting a direct cholinergic action on ventricular muscle refractoriness. Circ. Res. 49:511, 1981. 119. Pyner, S., and Coote, J. H.: Rostroventrolateral medulla neurons preferentially project to target specified sympathetic preganglionic neurones. Neuroscience 83:617, 1998.

120. Raczkowska, M., Eckberg, D. L., and Ebert, T. J.: Muscarinic cholinergic receptors modulate vagal cardiac responses in man. J. Auton. Nerv. Syst. 7:271, 1983. 121. Rigg, L., Heath, B. M., Cui, Y., and Terrar, D. A.: Localisation and functional significance of ryanodine receptors during beta-adrenoreceptor stimulation in the guineapig sino-atrial node. Cardiovasc. Res. 48: 254, 2000. 122. Rosenblueth, A., and Simeone, F. A.: The interregulations of vagal and accelerator effects on the cardiac rate. Am. J. Physiol. 110:42, 1934. 123. Rudolph, U., Crestani, F., Benke, D., et al.: Benzodiazepine actions mediated by specific ␥-aminobutyric acid A receptor subtypes. Nature 401:796, 1999. 124. Ruggeri, P., Battaglia, A., Ermirio, R., et al.: Role of nitric oxide in the control of the heart rate within the nucleus ambiguus of rats. Neuroreport 11:481, 2000. 125. Sakmann, B., Noma, A., and Trautwein, W.: Acetylcholine activation of single muscarinic K⫹ channels in isolated pacemaker cells of the mammalian heart. Nature 303:250, 1983. 126. Sanguinetti, M. C., Jiang, C., Curran, M. E., and Keating, M. T.: A mechanistic link between an inherited and an acquired cardiac arrhythmia: HERG encodes the Ikr potassium channel. Cell 81:299, 1995. 127. Sanguinetti, M. C., and Keating, M. T.: Role of delayed rectifier potassium channels in cardiac repolarisation and arrhythmias. News Physiol. Sci. 12:152, 1997. 128. Saul, J. P., Arai, Y. H., Berger, R. D., et al.: Assessment of autonomic regulation in chronic congestive heart failure by heart rate spectral analysis. Am. J. Cardiol. 61:1292, 1988. 129. Schwartz, P. J., Periti, M., and Malliani, A.: The long Q-T syndrome. Am. Heart J. 89:378, 1975. 130. Schwarz, P., Diem, R., Dun, N. J., and Forstermann, U.: Endogenous and exogenous nitric oxide inhibits norepinephrine release from rat heart sympathetic nerves. Circ. Res. 77:841, 1995. 131. Shepherd, J. T.: Cardiac mechanoreceptors. In Fozzard, H. A., Haber, E., Jennings, R. B., and Katz, A. M. (eds.): The Heart and Cardiovascular System. New York, Raven Press, p. 1535, 1986. 132. Skinner, J. E.: Neurocardiology shows that the central, not the peripheral, action of propranolol reduces mortality following coronary artery occlusion in the conscious pig. Integr. Physiol. Behav. Sci. 26:85, 1991. 133. Skinner, J. E., and Reed, J. C.: Blockade of frontocorticalbrain stem pathway prevents ventricular fibrillation of ischaemic heart. Am. J. Physiol. 240:H156, 1981. 134. Smith, S. A., and Smith, S. E.: Heart rate variations in the Guillain-Barré syndrome. Br. Med. J. 281:1009, 1980. 135. Smyth, H. S., Sleight, P., and Pickering, G. W.: Reflex regulation of arterial pressure during sleep in man: a quantitative method of assessing baroreflex sensitivity. Circ. Res. 24:109, 1969. 136. Spyer, K. M.: Neural organisation and control of the baroreceptor reflex. Rev. Physiol. Biochem. Pharmacol. 88:23, 1981. 137. Stotler, W. A., and McMahon, R. A.: The innervation and structure of the conductive system of the human heart. J. Comp. Neurol. 87:57, 1947. 138. Stull, J. T., and Mayer, S. E.: Biochemical mechanisms of adrenergic and cholinergic regulation of myocardiac

Neural Control of Cardiac Function

139.

140. 141. 142.

143.

144.

145.

146.

147. 148.

149.

contractility. In Berne, R. M., and Sperelakis, N. (eds.): Handbook of Physiology, Sect. 2. The Cardiovascular System, Vol. 1. The Heart. Bethesda, MD, American Physiological Society, p. 741, 1979. Sun, M. K.: Central neural organisation and control of sympathetic nervous system in mammals. Prog. Neurobiol. 47:465, 1995. Talman, W. T.: Cardiovascular regulation and lesions of the central nervous system. Ann. Neurol. 18:1, 1985. Taylor, E. W.: The evolution of efferent vagal control of the heart in vertebrates. Cardioscience 5:173, 1994. Taylor, E. W., Jordan, D., and Coote, J. H.: Central control of the cardiovascular and respiratory systems and their interactions in vertebrates. Physiol. Rev. 79:855, 1999. Thornton, J. M., Aziz, T., Schlugman, D., and Paterson, D. J.: Electrical stimulation of the midbrain increases heart rate and arterial blood pressure in awake humans. J. Physiol. (Lond.) 593:615, 2002. Vaile, J. C., Chowdhary, S., Osman, F., et al.: Effects of angiotensin II (AT1) receptor blockade on cardiac vagal control in heart failure. Clin. Sci. 101:559, 2001. Van Buren, J. M., and Ajmone-Marsan, C.: A correlation of autonomic and EEG components in temporal lobe epilepsy. Arch. Neurol. 3:683, 1960. Vandecasteele, G., Eschenhagen, T., Scholz, H., et al.: Muscarinic and beta-adrenergic regulation of heart rate, force of contraction and calcium current is preserved in mice lacking endothelial nitric oxide synthase. Nat. Med. 5:331, 1999. Verrier, R. L., and Lown, B.: Behavioural stress and cardiac arrhythmias. Annu. Rev. Physiol. 46:155, 1984. Verrier, R. L., Muller, J. A., and Hobson, J. A.: Sleep, dreams, and sudden cardiac death: the case for sleep as an autonomic stress test for the heart. Cardiovasc. Res. 31:181, 1996 Volders, P. G. A., Vos, M. A., Szabo, B., et al.: Progress in the understanding of cardiac early after depolarisations and torsades de pointes: time to revise current concepts. Cardiovasc. Res. 46:376, 2000.

231

150. Vybiral, T., Bryg, R. J., Maddens, M. E., and Boden, W. E.: Effect of passive tilt on sympathetic and parasympathetic components of heart rate variability in normal subjects. Am. J. Cardiol. 63:1117, 1989. 151. Wang, J., Irnaten, M., and Mendelowitz, D.: Characterisation of spontaneous and evoked GABA synaptic currents in cardiac vagal neurons in rats. Brain Res. 889:78, 2001. 152. Wang, Y., and Ramage, A. G.: The role of central 5-HT1A receptors in the control of B-fibre cardiac and bronchoconstrictor vagal preganglionic neurones in anaesthetised cats. J. Physiol. (Lond.) 536:753, 2001. 153. Wharton, J., Gulbenkian, S., Merighi, A., et al.: Immunohistochemical and ultrastructural localisation of peptide-containing nerves and myocardial cells in the human atrial appendage. Cell Tissue Res. 254:155, 1988. 154. Wikberg, J. E. S., and Lefkowitz, R. J.: Adrenergic receptors in the heart: pre- and postsynaptic mechanisms. In Randall, W. C. (ed.): Nervous Control of Cardiovascular Function. New York, Oxford University Press, p. 95, 1984. 155. Wu, S. Y., and Dun, N. J.: Potentiation of IPSCs by nitric oxide in immature rat sympathetic preganglionic neurones in vitro. J. Physiol. (Lond.) 495:479, 1996. 156. Yanagihara, K., and Irisawa, H.: Inward current activated during hyperpolarization in the rabbit sinoatrial node cell. Pflugers Arch. 385:11, 1980. 157. Yang, Z., Smith, L., and Coote, J. H.: Paraventricular nucleus activation of renal sympathetic neurones is synaptically depressed by nitric oxide and glycine acting at a spinal level. Neuroscience 124:421, 2004. 158. Yashpal, K., Gauthier, S., and Henry, J. L.: Oxytocin administered intrathecally preferentially increases heart rate rather than arterial pressure in the rat. J. Auton. Nerv. Syst. 20:167, 1987. 159. Yatani, A., Hamm, H., Codina, J., et al.: A monoclonal antibody to the subunit of Gk blocks muscarinic activation of atrial K⫹ channels. Science 241:828, 1988. 160. Zipes, D. P., Levy, M. N., Cobb, L. A., et al.: Sudden cardiac death: neural-cardiac interactions. Circulation 76(Suppl. I): I-202, 1987.

12 Neurobiology of the Enteric Nervous System JACKIE D. WOOD

Overview Historical Perspective ENS Conceptual Model Histoanatomy Enteric Neuronal Morphology Enteric Sensory Neurons Visceral Hypersensitivity in Irritable Bowel Syndrome Chemoreception Mechanoreception Interneurons

Enteric Motor Neurons Secretomotor Neurons Excitatory Musculomotor Neurons Inhibitory Musculomotor Neurons Disinhibitory Motor Disorders Chronic Intestinal Pseudo-obstruction Sphincteric Achalasia Neuronal Electrical Behavior Resting Membrane Potentials Action Potentials

OVERVIEW The enteric nervous system (ENS) contains as many neurons as the spinal cord. An estimated 100 million neurons are required for programmatic control of digestive functions (e.g., specialized motility patterns and glandular secretion). The required number of neurons, together with accompanying glia, would greatly expand the volume of the central nervous system (CNS) if placed there. Rather than packing the neural control circuits exclusively within the skull or vertebral column and transmitting control signals over long, unreliable transmission lines to the gut, vertebrate animals have evolved with the neural control networks distributed along the length of the digestive tract in close apposition to the muscles, glands, and blood vessels that must be controlled and whose activity must be integrated for effective whole-organ function. The ENS integrates contraction of the musculature, glandular secretion, and intramural blood flow into organized patterns of digestive behavior. Activity in the ENS may either evoke contractions in the gut musculature or inhibit contractile activity as required for the organization of specific patterns of organ motility. Likewise, the ENS organizes absorption from and secretion into the intestinal lumen and integrates these functions with motility. Integrative microcircuitry in the ENS also controls blood

Synaptic Transmission Fast Excitatory Postsynaptic Potentials Slow Synaptic Excitation Inhibitory Postsynaptic Potentials Presynaptic Facilitation Presynaptic Inhibition Conclusions

flow within the gut’s walls and the distribution of flow to the musculature, mucosa, and lamina propria. Malfunctions of integrative ENS control of the gut’s effector systems are becoming increasingly recognized as underlying factors in gastrointestinal disorders, especially in the functional gastrointestinal disorders.173–175 The musculature, mucosal secretory epithelium, and blood and lymphatic vasculature are the gut’s effector systems. Moment-to-moment behavior of any of the specialized compartments of the gut (e.g., stomach, small and large intestine) reflects neurally integrated activity of the effector systems. The ENS not only controls the activity of each effector, but also coordinates the activity of each to achieve organized behavior of the whole organ. Coordinated activity is achieved through timing of excitatory and/or inhibitory neural inputs to each effector system individually. Coordination of mucosal secretion and muscle contractile activity in a timed sequence during intestinal propulsion of luminal contents is one example of ENS integrative function. Secretion of H2O, electrolytes, and mucus occurs first, followed by propulsive contractions in the musculature. A simultaneous increase in neurally controlled blood flow to the secretory glands supports the enhanced secretory response. The ENS is viewed as a “minibrain” with a library of programs for multiple patterns of small or large intestinal 249

250

Function of the Peripheral Nervous System

behavior. A specific program determines motor behavior in the postprandial state, and another establishes the pattern of intestinal motility that characterizes the fasting state. The specialized motility pattern seen in the upper third of the small intestine during emesis reflects output of another of the programs in the library. During emesis, peristaltic propulsion in the upper third of the small intestine is reversed for rapid movement of the luminal contents toward the stomach. The emetic program can be called up from the library either by commands from the brain or by local sensory detection of threatening substances in the lumen.74

HISTORICAL PERSPECTIVE Modern progress in understanding the neurobiology of the ENS started in the mid-20th century with a transition in conceptual thinking from the traditional view that all control was central, via sympathetic and parasympathetic neural pathways, to a concept that emphasized a local minibrain as the chief determinant of intestinal behavior. The prevailing concept in the first half of the 20th century followed the classic concepts of neural control of other autonomic systems (e.g., heart, blood vasculature, airways, and urinary bladder). Ganglia in the gut wall were presumed to be typical parasympathetic terminal ganglia. Postganglionic neurons with their cell bodies in the terminal ganglia were expected to initiate responses of the musculature or secretory glands through the direct release of acetylcholine at neuromuscular and neuroglandular junctions. All of the excitatory impulses for evoking

Outmoded concept Brain stem Spinal cord

Spinal cord

Preganglionic parasympathetic nerve

Preganglionic sympathetic nerve

muscle contraction and glandular secretion were postulated to be transmitted to the gut along parasympathetic autonomic projections starting in either the brainstem or the sacral region of the spinal cord. The sympathetic division of the autonomic nervous system was the accepted inhibitory component of gut innervation in the early concept. Postganglionic neurons, with their cell bodies in prevertebral sympathetic ganglia, were thought to release norepinephrine to suppress secretion and contraction of the musculature. A brake-accelerator analogy of opposing inhibitory and excitatory input, like that in the heart, was invoked to explain neural control of gut motility. A balance favoring parasympathetic excitation was expected to result in increased activity; balance favoring sympathetic inhibition accounted for suppression of activity. The ganglia in the gut were interpreted as relaydistribution centers for signals transmitted from the CNS (Fig. 12–1). In this model, all of the decision making and automated feedback control functions resided in neural networks in the CNS. No intelligence was attributed to the nervous system in the gut. Sir Johannes Newport Langley, the British autonomic physiologist of Cambridge University’s Trinity College, recognized during the first quarter of the 20th century that the ganglia of the gut did more than robotically relay and distribute information from the CNS. Langley was conceptually unable to reconcile the large disparity between the 2 ⫻ 108 neurons in the gut and the few hundred efferent fibers in the vagus nerves other than to conclude that the nervous system of the gut possessed

Current concept Brain stem Spinal cord

Spinal cord

Parasympathetic nervous system

Sympathetic nervous system

Enteric nervous system (brain-in-the-gut) Interneurons Postganglionic parasympathetic neuron

Postganglionic sympathetic neuron Motor neurons Musculature Glands Blood vessels

Musculature Glands Blood vessels

FIGURE 12–1 Outmoded and current concepts of the functional organization of the autonomic control of the digestive tract.

251

Neurobiology of the Enteric Nervous System

integrative functions independent of the CNS.75 This conclusion was reinforced by his demonstration of organized propulsion of intraluminal boluses in intestinal segments detached from the body in vitro.76 Based on these observations, he proposed that the nervous system in the gut is a distinct third division of the autonomic nervous system and named it the enteric nervous system. Langley’s insightful concept languished for nearly five decades. Now, in the new millennium, three divisions— parasympathetic, sympathetic, and enteric—are accepted by scientists the world over as the construct for the autonomic nervous system.40,41 After Langley, the utmost advance in basic understanding of neuropathy in the digestive tract has been the enlightened realization that the nervous system of the gut is a minibrain with intelligent neural networks. Emergence of the heuristic model of the ENS as a “little brain” has opened the way for basic research that follows many of the approaches used to gain understanding of the neurobiology of the brain and spinal cord. The outcome has been improved insight into the disordered physiology that is manifest in a variety of symptoms, including those of a functional nature. Issues related to functional gastrointestinal disorders today are reminiscent of the historical progression of insight into functional disorders of the nervous system. Functional disorders of the CNS, like those of the digestive tract, were designated as such out of ignorance of the pathophysiologic basis of the disorder. Parkinson’s disease was regarded as an idiopathic disorder in the years before the synaptic microcircuits of the basal ganglia and the importance of the neurotransmitter dopamine were understood for the brain. The history of progress in basic brain research underlies a conviction that expanded understanding of the neural networks in the ENS has potential for uncovering unrealized explanations for gastrointestinal disorders currently described as idiopathic in nature, as well as rational approaches to therapy.

ENS CONCEPTUAL MODEL The heuristic model for the ENS (Fig. 12–2) is the same as for all integrative nervous systems, whether the “simple” nervous systems of invertebrate animals or the vertebrate CNS. Like other integrative nervous systems, the ENS functions with sensory neurons, interneurons, and motor neurons. Sensory neurons have receptor regions specialized for detecting changes in stimulus energy. The gut’s sensory receptors are classified according to the kinds of energy they detect. These include thermo-, chemo-, and mechanoreceptors. The receptor regions transform changes in stimulus energy into signals coded by action

Brain

Effector systems Muscle Sensory neurons

Interneurons

Motor neurons

Enteric nervous system

Secretory Glands Blood vessels

Organ system behavior

FIGURE 12–2 Sensory neurons, interneurons, and motor neurons are synaptically interconnected to form the integrated microcircuits of the enteric nervous system. Like the central nervous system, sensory neurons, interneurons, and motor neurons are connected synaptically for flow of information from sensory neurons to interneuronal integrative networks to motor neurons to effector systems. The enteric nervous system organizes and coordinates the activity of each effector system into meaningful behavior of the whole organ. Bidirectional communication occurs between the central and the enteric nervous system.

potentials that subsequently are transmitted along sensory nerve fibers to other points in the nervous system for processing. Interneurons are interconnected to form networks that process sensory information and control the behavior of motor neurons. Multiple synaptic connections among interneurons organize the neurons into “logic” circuits that decipher action potential codes from sensory neurons and neural signals from elsewhere in the nervous system. Some of the interneuronal circuits are integrative; others are reflex circuits. Integrative circuits are circuits that process input information and reconfigure it into functional output signals that are not direct transforms of the input signal. Plasticity resides in integrative circuits. Reflex circuits, in comparison, are “hardwired,” much like spinal motor reflexes, to generate stereotyped outputs in response to input from specific kinds of sensory receptors. Motor neurons are the final common pathways for transmission of control signals from integrative and reflex circuits to the effector systems. Motor neurons may be stimulated to fire by excitatory synaptic input derived from interneuronal circuitry, or their activity may be suppressed by inhibitory synaptic input from the interneuronal circuitry. In the digestive tract, motor neurons may release excitatory or inhibitory neurotransmitters at their junctions with the muscles and thereby actively excite muscular contraction or actively inhibit contraction.

252

Function of the Peripheral Nervous System

HISTOANATOMY

tation for sustaining the neural networks in a functional state as the ENS is subjected to the mechanical forces and deformations that occur during contraction of the muscle or expansion and stretching of the wall as the lumen fills. Most of the motor neurons that innervate the circular and longitudinal muscle coats reside in the myenteric plexus.12,13 The submucosal plexus is a ganglionated plexus situated throughout the submucosal space between the mucosa and circular muscle coat (Fig. 12–3). It is most prominent as a ganglionated network in the small and large intestine. No ganglionated submucosal plexus exists in the esophagus, and ganglia are sparse in the submucosal space of the stomach. In larger mammals (e.g., pig, human), the submucosal plexus consists of an inner submucosal network (Meissner’s plexus) located at the serosal side of the muscularis mucosae and an outer plexus (Schabadasch’s plexus) adjacent to the luminal side of the circular muscle coat. In human small and large bowel, a third intermediate plexus lies between Meissner’s and Schabadasch’s plexuses.157 Motor innervation of the intestinal crypts and villi is derived from ganglion cells in the submucosal plexus. Some of the neurons in submucosal ganglia send fibers to

Cell bodies of ENS neurons are in ganglia buried inside the walls of the digestive tract. The ganglia are interconnected by interganglionic fiber tracts to form ganglionated plexuses. Interganglionic fiber tracts consist mainly of projections from ganglion cell bodies in one ganglion to connect synaptically with neurons in neighboring ganglia. The ganglionated plexuses form a distributed continuum around the circumference and along the length of each of the specialized organs of the gut. Myenteric and submucosal plexuses are the prominent ganglionated plexuses of the ENS (Fig. 12–3). The myenteric plexus, known also as Auerbach’s plexus, is the major plexus of the ENS. It is located between the longitudinal and circular muscle coats of most all regions of the gastrointestinal tract. It is a two-dimensional array of flat, disclike neurons, ganglia, and interganglionic fiber tracts sandwiched between the longitudinal and circular muscle coats (Fig. 12–3). Unlike other autonomic ganglia, the cell bodies of the ganglion cells are not in grapelike clusters. They lie edge to edge like a single layer of coins placed in a two-dimensional plane. The single-layered, two-dimensional organization is postulated to be an adap-

A

Enteric nervous system Myenteric plexus Circular muscle Submucosal plexus

Longitudinal muscle

Mucosa

80 μm

B

80 μm

Myenteric plexus

C

Submucosal plexus

FIGURE 12–3 Histoanatomic organization of the enteric nervous system. A, Ganglia and interganglionic fiber tracts of the myenteric and submucosal plexuses are notable structures of the enteric nervous system. The myenteric plexus is distributed between the circular and longitudinal muscle coats; the submucous plexus is between the mucosa and circular muscle coat. B, Tyrosine hydroxylase stain of the myenteric plexus as it appears on the underside of the longitudinal muscle coat after microdissection from guinea pig small intestine. C, Tyrosine hydroxylase stain of the submucosal plexus as it appears in a preparation obtained from guinea pig small intestine.

Neurobiology of the Enteric Nervous System

the myenteric plexus and also receive synaptic input from axons projecting from the myenteric plexus.138,139 The interplexus connections link the two networks into a functionally integrated nervous system. Structure, function, and neurochemistry of enteric ganglia differ significantly from other autonomic ganglia. Unlike other autonomic ganglia in which function is mainly as relay-distribution centers for signals transmitted from the CNS, enteric ganglia are interconnected to form a nervous system with mechanisms for integration and processing of information like those found in the brain and spinal cord. Most properties of the ENS mimic those of the CNS. Some examples are 1. Complex integrative functions: Like the brain, the ENS executes complex integrative functions that control single effector systems and coordinates the activity of multiple systems into patterns of overall behavior that achieve homeostasis. Program libraries are present in both the ENS and CNS, but not in other autonomic ganglia. 2. Sensory neurons: Sensory neurons transmit information to the neural circuits of the ENS for processing in the same manner that the CNS receives and processes sensory signals. 3. Interneurons: Interneurons are synaptically interconnected into information processing circuits in both the ENS and CNS, but generally not in other autonomic ganglia. 4. Motor neurons: Enteric motor neurons are the final common pathways for flow of information from interneuronal processing circuits to the effector systems. Their function is analogous to that of alpha and gamma motor neurons from the spinal cord to skeletal muscles. 5. Glial elements: The supporting cells of the ENS are like glial cells in the brain but, unlike the Schwann cells found associated with other autonomic ganglia. Enteric glial elements structurally and chemically resemble astroglia of the brain. 6. Synaptic neuropil: Synaptic neuropils are present in both enteric ganglia and the CNS. Neuropils are tangled meshworks of fine nonmyelinated nerve fibers that are segregated in regions of the ganglia away from the cell bodies that give rise to the fibers. The presence of a neuropil is significant because, in all integrative nervous systems, most of the information processing occurs in microcircuits within a synaptic neuropil. 7. Multiple synaptic mechanisms: As in the brain and spinal cord, the neural circuits of the ENS express multiple forms of neurotransmission, which is involved in the handling of information. 8. Multiple neurotransmitters: Most of the neurotransmitters found in the brain and spinal cord are also expressed in the ENS.

253

9. Absence of connective tissue: Collagen, which forms part of the connective tissue for the structural support of autonomic ganglia, is absent from the interior of enteric ganglia, as it is in the CNS. 10. Reduced extracellular space: Unlike other autonomic ganglia, extracellular space is reduced by close packing of glial support elements. 11. Isolation from blood vessels: No blood vessels enter enteric ganglia of the intestine, and a blood-ganglion barrier analogous to the blood-brain barrier is inserted between the vasculature and the synaptic circuits of the ganglia.

ENTERIC NEURONAL MORPHOLOGY Dogiel types I and II are the principal morphologic classes of neurons in the ENS. The German neuroanatomist A. S. Dogiel described two morphologic types of enteric ganglion cells that are named for him. These are Dogiel types I and II, both of which are found in myenteric and submucosal plexuses. Both types of neurons are distributed in a two-dimensional plane (i.e., the cell bodies are not stacked one on the other). Dogiel type I neurons have cell bodies with multiple short processes and a single long process (Fig. 12–4). These are flat neurons with the processes extending for short distances from the cell body in the circumferential and longitudinal planes of the wall. The short processes are dendrites, which receive synaptic input; the long process is an axon. Axons of Dogiel type I neurons project for relatively long distances through interganglionic fiber tracts and many rows of ganglia. Still, the projections are not long in absolute length; the longest known projections of any enteric ganglion cell are only approximately 2 to 3 cm. Dogiel I neurons projecting in a specific direction are identified by expression of specific neurotransmitters. Aborally projecting Dogiel I axons release nitric oxide and vasoactive intestinal polypeptide (VIP) as neurotransmitters; those projecting in the oral direction release substance P and acetylcholine. Some of the Dogiel type I neurons are motor neurons to the musculature and secretory epithelium; others are interneurons. The cell bodies of ENS neurons with Dogiel type II morphology have smooth surfaces with long and short processes arising in a variety of configurations (Fig. 12–4). The long processes may extend through interganglionic fiber tracts across several rows of ganglia in the circumferential, oral, or aboral direction. Shorter processes may only project within the home ganglion. Almost all Dogiel II neurons in the myenteric plexus project a process to the submucosa/mucosa.36 Terminals of Dogiel II projections into the mucosal regions are fired by mechanical and chemical stimulation of the mucosa (Fig. 12–5).8 Dogiel II

254

Function of the Peripheral Nervous System

FIGURE 12–4 Dogiel morphologic types I and II neurons. A, Dogiel morphologic type I neuron in the submucosal plexus of guinea pig small intestine. Dogiel I neurons have multiple short dendrites at the edges of the cell body and a single axon. B, Dogiel morphologic type II neuron in the myenteric plexus of the guinea pig small intestine. Dogiel II neurons have multiple long processes of variable length that ramify throughout the ganglion and into interganglionic fiber tracts. The neurons in A and B were marked by the injection of biocytin from micropipettes that also served as recording microelectrodes.

neurons are involved at an early stage in the handling of sensory information, as well as interneuronal functions in the microcircuits of the ENS. They are identified by expression of immunoreactivity for the neurotransmitters substance P and acetylcholine and the calcium-binding protein calbindin.

ENTERIC SENSORY NEURONS Sensory neurons belong to one of the three functional categories of enteric neurons illustrated in Figure 12–2. Sensory afferent fibers with neuronal cell bodies in nodose and dorsal root ganglia transmit information from the gastrointestinal tract and gallbladder to the CNS for processing. These afferents also transmit a steady stream of information to the local processing circuits in the ENS and to prevertebral sympathetic ganglia. Mechano-, chemo-, and thermoreceptors are present in the ENS. Mechanoreceptors sense changes in mechanical forces in the mucosa, musculature, serosal surface, and mesentery. They supply both the ENS and the CNS with information on stretch-related tension and muscle length in the wall and on the movement of luminal contents as they brush past the mucosal surface. Mesenteric receptors generate coded information on gross movements of the organ. Chemoreceptors generate information on the concentration of nutrients, osmolarity, and pH in the luminal contents. Recording of nerve impulses in afferent fibers as they exit the gut shows most sensory receptors to be multimodal in that they respond to both mechanical and chemical stimuli.

Thermoreceptors supply the brain with deep-body temperature data used in body temperature regulation and probably account also for the sensations of temperature change in the lumen. Presence in the gastrointestinal tract of pain receptors (nociceptors) equivalent to those connected with C fibers and A␦ fibers elsewhere in the body is likely, but not unequivocally confirmed except for the gallbladder.18 The biophysical principles of sensory function in the gastrointestinal tract are like those found elsewhere in the body. Each sensory neuron has at least one neurite with a generator region that converts a change in stimulus energy into an action potential code. The frequency of action potential discharge is a code for transmission of information on intensity of the stimulus to processing circuits in the ENS and CNS. Information on contractile state of the musculature and distention of the gut wall is coded by mechanoreceptors. Whether the cell bodies of intramuscular and mucosal mechanoreceptors reside in dorsal root ganglia, in enteric ganglia, or in both places is an unresolved question.36,170 Stretch-sensitive mechanoreceptors have pathophysiologic importance because a significant proportion of patients diagnosed with the irritable bowel syndrome (IBS) experience abnormal sensitivity to stretch that is perceived as abdominal pain.77,162

Visceral Hypersensitivity in Irritable Bowel Syndrome The primary causes of abdominal pain of digestive tract origin are distention and exaggerated muscular contraction. Stimuli such as pinching and burning of the mucosa

255

Neurobiology of the Enteric Nervous System

Slow excitatory inputs Neural

Paracrine AH Dogiel II neuron

Myenteric plexus

Myenteric plexus Circular muscle coat Submucosa 5-HT3 Serotonergic receptors

5-HT Mucosa

Mechanosensitive enterochromaffin cell

FIGURE 12–5 Functional neuroanatomy of AH/Dogiel morphologic type II neurons in the neural networks of the ENS. The projectional geometry of the neurites underscores the significance of the cell body as a gating determinant of neurite-to-neurite communication across the cell soma (see Figs. 12–4B and 12–9). Most AH neurons project one or more neurites to the mucosa. The terminals in the mucosa fire in response to acidic pH and in response to serotonin released by mechanical stimulation of mucosal enterochromaffin cells. Like elements in a neural network, AH neurons are stimulated to fire repetitively by slow excitatory inputs. Firing of the neuronal cell body is transformed into slow synaptic output at synapses with neighboring AH neurons in the network (see Fig. 12–4B). The somal gate is closed in the absence of slow synaptic input. In this state, inbound information from the mucosa is not gated to neurites across the cell body. In circumstances in which the somal gate is opened by either slow synaptic or paracrine excitatory input, information from the mucosa is gated in the direction of slow synaptic outputs to AH neurons in the circuit (see Fig. 12–10).

do not evoke pain. Hypersensitivity of the mechanosensors for stretch (distention) and contractile tension are implicated as pain factors in patients diagnosed with IBS. Heightened sensitivity to distention and conscious awareness of gastrointestinal sensations experienced by patients with IBS are generalized phenomena that extend throughout their digestive tract, including the esophagus.120 The locus of neurologic disturbance responsible for the hypersensitivity in patients with IBS is not fully understood. The available evidence suggests disordered neurophysiology in either the peripheral nervous system, the CNS, or both. A peripheral neural hypothesis states that mechanosensitive afferents in the gut wall become hypersensitive to mechanical stimuli and, as a result, transmit erroneous information to the CNS. The alternative hypothesis is for the mechanosensors to be transmitting accurate information, with misinterpretation in the central processing circuits being responsible for the perception of

abnormal sensations from the gut. Results obtained with modern methods of brain imaging suggest abnormal processing of sensory information in the amygdala, anterior cingulate cortex, and prefrontal cortex in patients with IBS.89,118,135 The logical conclusion is that sensory physiology may be disordered at peripheral and/or central levels of neural organization in patients with IBS who experience gastrointestinal hypersensitivity to mechanical stimuli. Hyposensory perception, particularly in the rectosigmoid region, is at the opposite extreme of gastrointestinal sensory abnormality. Sensory suppression, either in the intramural connections for rectoanal stretch reflexes or in the transmission pathway from the rectosigmoid to conscious perception of distention in the CNS, is implicated as an underlying factor in the pathogenesis of chronic constipation and associated symptoms.81 Hyperglycemia represents another of the pathologic disturbances that can lead to sensory neuropathy and suppressed sensations of rectosigmoid fullness that may lead to constipation.19

Chemoreception Sensory transduction of chemical information related to luminal nutrient concentrations in the small bowel involves paracrine signaling from enteroendocrine cells to afferent terminals in the intestinal wall. Mucosal enteroendocrine cells “taste” the luminal contents and release cholecystokinin (CCK) in direct proportion to the concentration of products of protein digestion and lipids. Amounts of CCK released are a transform of the nutrient concentration, which in turn acts to evoke concentrationdependent firing in vagal afferent neurons. Vagal afferents in the stomach and upper small bowel bifurcate for simultaneous transmission of the same information to the CNS and ENS.7 The firing frequency transmitted by the afferents becomes a transform of the stimulus strength for simultaneous decoding in the brain and enteric processing circuitry. The neuronal receptors for CCK belong to the CCK-A subtype.29,136 Vagal sensory mechanisms involving CCK are implicated in the functions involved in the sensations of fullness, satiety, and control of food intake.136 Disordered vagal sensory function is suspect in the pathophysiology of functional dyspepsia.146,151 Firing of impulses in vagal afferent terminals, spinal afferent terminals, and the terminals of the projections of myenteric Dogiel type II neurons in the intestinal mucosa can be initiated by the action of 5-hydroxytryptamine (5-HT) released from mucosal enterochromaffin cells.8,71,151 Distortion of the mucosal surface by mechanical shearing forces or the presence of noxious chemical agents stimulates the enterochromaffin cells to release 5-HT, which then becomes a paracrine signal to serotonergic receptors on spinal afferent and Dogiel type II terminals in the mucosa and elsewhere in the gut wall (see Fig. 12–5). The serotonergic receptors on vagal afferent

256

Function of the Peripheral Nervous System

terminals, on spinal afferent terminals, and on the mucosal terminals of myenteric Dogiel type II neurons belong to the 5-HT3 receptor subtype. Stimulation of the 5-HT3 receptors on vagal afferent terminals is a trigger for nausea and emesis and accounts for the efficacy of 5-HT3 antagonists (e.g., odansetron) as antiemetics. Histologic examination of mucosal biopsies taken from patients with IBS show elevated numbers of enterochromaffin cells and mucosal mast cells, both of which are paracrine sources of serotonin.11,112 Measurements of elevated postprandial concentrations of serotonin in the hepatic portal blood of patients with IBS suggest that elevated release and excessive stimulation of 5-HT3 receptors on intestinal afferent nerve terminals might account for the abdominal pain and discomfort experienced by these patients. Efficacy of 5-HT3 receptor antagonists (e.g., alosetron) in the treatment of abdominal pain in women with the diarrheapredominant form of IBS adds to the findings that suggest a role for disordered release of serotonin and its action to stimulate firing in sensory afferents as underlying pathologic factors in IBS.15

Mechanoreception Low- and high-threshold mechanoreceptors code mechanosensory information from the gastrointestinal and biliary tracts. Low-threshold splanchnic afferents respond to innocuous levels of distention; high-threshold afferents respond to distending volumes (e.g., balloon distention) only at strengths above a set threshold. The latter highthreshold afferents are predominantly unmyelinated C fibers. Low-threshold afferents are myelinated fibers of the A␦ class. Low-threshold mechanoreceptors are believed to be the afferent component of autonomic regulatory reflexes. Whether some degree of activity in lowthreshold pathways reaches the level of conscious perception is uncertain; nevertheless, it is likely that some nonpainful sensations such as abdominal fullness and the presence of gas are derived from the low-threshold– related activity. High-threshold afferents are thought to be the sensory correlates of sharp, localized pain in organs such as the gallbladder, where pain is the only consciously perceived sensation.17 Combined involvement of both lowand high-threshold fibers in nociperception cannot be excluded. Cervero and Janig18 suggested for organs such as the urinary bladder and colon, where distention can evoke sensations ranging from mild fullness to intense pain, that activation of differing proportions of low- and high-threshold mechanoreceptors accounts for the gradation of sensations. Acute visceral pain may emerge from activation of highthreshold nociceptive fibers, whereas chronic forms of visceral pain may be attributed to sensitization of both types of mechanoreceptors by inflammatory conditions or ischemia. Application of acetic acid or other irritants to the large

intestinal mucosa in animals lowers the threshold and sensitizes both the high- and low-threshold distention-sensitive afferents. Involvement of another class of splanchnic afferents termed silent nociceptors is also suspect in chronic abdominal pain. Silent nociceptors are splanchnic afferents that do not discharge impulses in response to the strongest of distending stimuli. Inflammatory mediators sensitize these normally silent receptors. Spontaneous firing and responses to normally innocuous mechanical distention can be recorded in previously unresponsive splanchnic afferents after the bowel is inflamed by experimentally applied irritants or inflammatory substances. Reports that a significant fraction of patients with IBS progress to the development of IBS-like symptoms after an acute bout of infectious enteritis is reminiscent of the basic scientific findings that sensory afferents become sensitized in inflammatory states.25,26 Gwee et al.45 reported that 23% of patients with acute gastroenteritis progressed to IBS-like symptoms within 3 months. Nevertheless, the question of whether the association between acute infectious enteritis and IBS reflects low-level inflammation (e.g., microscopic enteritis) and chronic sensitization of intestinal afferents by inflammatory mediators is not fully resolved. Primary sensory afferent terminals in the intestine express receptors for several different messenger substances, including inflammatory mediators. Receptors for 5-HT, bradykinin, ATP, adenosine, prostaglandins, leukotrienes, and proteases are among those expressed by the afferent terminals.71 Accumulation of any of these substances has potential for increasing the sensitivity of intestinal afferents, especially in disordered conditions such as inflammation and ischemia. Mediators released during mast cell degranulation can sensitize “silent” nociceptors in the large intestine. In animal models, mast cell degranulation results in a reduced threshold for pain responses to balloon distention, and treatment with mast cell–stabilizing drugs prevents this lowering of the pain threshold.92

INTERNEURONS Interneurons represent another of the three functional categories of enteric neurons illustrated in Figure 12–2. Enteric interneurons are synaptically interconnected into integrated circuits that process sensory information and direct the activity of motor neurons to the musculature, secretory epithelium, and blood vasculature. They have individualized properties, including specific morphologic characteristics of the cell somas, directionality of projections, length of axonal projections, electrical and synaptic behavior, and neurotransmitter codes. Multiple synaptic connections contact the cell bodies of interneurons and their axons and dendrites in the ganglionic neuropil. Synaptic connections are made between

Neurobiology of the Enteric Nervous System

axons and cell bodies, between axons and dendrites, and from axon to axon. Whereas the bulk of information processing in integrative nervous systems occurs in the synaptic neuropil, only the cell bodies of enteric neurons are presently accessible for electrophysiologic study of electrical and synaptic behavior. As in the brain, the interneuronal microcircuits account for the higher functions that are required for integrated control and coordination of behavior of effector systems within the complete organ.

ENTERIC MOTOR NEURONS Motor neurons innervate the effector systems. The motor neuron pool of the ENS contains both excitatory and inhibitory neurons. Excitatory motor neurons release neurotransmitters that stimulate muscle contraction and secretion. Inhibitory motor neurons release neurotransmitters that suppress muscular contraction.

Secretomotor Neurons Secretomotor neurons are excitatory motor neurons in the ENS that innervate the intestinal crypts of Lieberkühn (Fig. 12–6). They are uniaxonal neurons with characteristic shapes described as Dogiel type I. When secretomotor neurons fire, they release acetylcholine and/or VIP as neurotransmitters at their junctions with the epithelium of the crypts. Collaterals from secretomotor axons innervate submucosal arterioles (Fig. 12–6). Collateral innervation of the blood vessels links blood flow to secretion by releasing

Secretomotor neuron

dila Va so

(+)

tati

on

Elevated blood flow

Intestinal crypt ¨ of Lieberkuhn

Arteriole

Secretion

FIGURE 12–6 Intestinal crypts of Lieberkühn are innervated by secretomotor neurons. Secretomotor neurons in the submucosal plexus have a single axon that projects to the mucosal epithelium (see Fig. 12–4A). When the neuron fires, neurotransmitters that evoke secretion are released at the junctions with the crypts. Axon collaterals to blood vessels simultaneously dilate the vessels to increase blood flow in support of stimulated secretion.

257

acetylcholine simultaneously at neuroepithelial and neurovascular junctions. Acetylcholine acts at the blood vessels to release nitric oxide from the endothelium, which in turn dilates the vessels and increases blood flow in support of stimulated secretion. Secretomotor neurons have receptors that receive excitatory and inhibitory synaptic input from other neurons in the integrative circuitry of the ENS and from sympathetic postganglionic neurons. They are influenced also by paracrine chemical messages from non-neural cell types in the mucosa and submucosa (e.g., enterochromaffin cells and immune/inflammatory cells). Activation of the excitatory receptors on secretomotor neurons stimulates the neurons to fire and release their transmitters at the junctions with the crypts and regional blood vessels. The overall result of secretomotor neuronal firing is stimulation of the secretion of H2O, electrolytes, and mucus from the crypts into the intestinal lumen. Inhibitory inputs hyperpolarize the membrane potential of secretomotor neurons and thereby decrease the probability of firing. The physiologic effect of inhibiting secretomotor firing is suppression of mucosal secretion. Postganglionic neurons of the sympathetic nervous system are an important source of inhibitory input to the secretomotor neurons. Norepinephrine released from sympathetic axons acts at ␣2a noradrenergic receptors to inhibit firing of the secretomotor neurons. Inhibition of secretomotor firing reduces the release of excitatory neurotransmitters at the junctions with epithelial cells in the crypts. The end result is reduced secretion of water and electrolytes. Suppression of secretion in this manner is part of the mechanism involved in sympathetic nervous “shutdown” of gut function in homeostatic states in which blood is shunted from the splanchnic to the systemic circulation. Knowledge of the neurobiology of submucosal secretomotor neurons is key to understanding the pathophysiology of secretory diarrhea as well as constipation. In general, secretomotor hyperactivity is associated with neurogenic secretory diarrhea; hypoactivity is associated with decreased secretion and a constipated state. Suppression of secretomotor firing by antidiarrheal agents (e.g., opiates, clonidine, and somatostatin analogues) is manifest as harder, drier stools. Stimulation by chemical mediators, such as VIP, serotonin, and histamine, is manifest as more liquid stools. Watery diarrhea may be caused by several different pathophysiologic events. For example, secretomotor neurons may be overly stimulated by excessive serotonin release from mucosal enterochromaffin cells, by histamine release from inflammatory/immune cells in the mucosa/submucosa, or by circulating VIP released from a VIPoma outside the gut. Release of histamine or other inflammatory mediators associated with diarrhea not only stimulates the firing of secretomotor neurons, but also simultaneously acts at presynaptic inhibitory receptors to suppress the release of norepinephrine from the

258

Function of the Peripheral Nervous System

postganglionic sympathetic axons that provide inhibitory input to secretomotor neurons.80 Thus two pathologic factors are involved in the production of neurogenic secretory diarrhea. One is overstimulation of secretomotor neuronal firing and the other is presynaptic suppression of norepinephrine from sympathetic postganglionic neurons that results in removal of the sympathetic braking action on the neurons. Secretomotor neurons have excitatory receptors for several neurotransmitters, including acetylcholine, VIP, substance P, and serotonin (see review by Wood170). One of the serotonergic receptors belongs to the 5-HT3 receptor subtype.33 Efficacy of blockade of 5-HT3 receptors by a 5-HT3 antagonist in the treatment of diarrhea in the diarrhea-predominant population of women with IBS14,15 suggests that overstimulation of secretomotor neurons by serotonin is a significant factor in this form of IBS. Observations that IBS symptoms are commonly exacerbated in the postprandial state20,156 and evidence for elevated postprandial release of serotonin6 raise suspicion that overactive release of serotonin from the enterochromaffin cell population in the mucosa underlies the diarrheal symptoms of IBS. This is reinforced by suggestions of elevated numbers of serotonin-containing enterochromaffin cells and also mast cells in the mucosa of patients with IBS.11,112

Excitatory Musculomotor Neurons Acetylcholine and substance P are the principal neurotransmitters released from excitatory motor neurons to evoke contraction of the muscles.28,47,84 The receptors for acetylcholine on the musculature belong to the muscarinic receptor subtype. Receptors for substance P at neuromuscular junctions are the NK-1 subtype. Axons of excitatory musculomotor neurons to the intestinal circular muscle coat project in the orad direction and excitatory musculomotor axons to the longitudinal muscle coat project in the aborad direction.12,13 Axons of excitatory gastric motor neurons project in the orad direction to both the musculature and mucosa.119,127

Inhibitory Musculomotor Neurons Inhibitory motor neurons release neurotransmitters that suppress contractile activity of the musculature. Early evidence implicated a purine nucleotide, possibly ATP, as the inhibitory transmitter released by enteric inhibitory musculomotor neurons. Consequently, for the gut, the term purinergic neuron temporarily became a synonymous term for inhibitory motor neuron. ATP was joined subsequently by VIP, pituitary adenylate cyclase–activating peptide, and nitric oxide as candidate inhibitory transmitters released by enteric inhibitory musculomotor neurons.43,44,65,97,98 Enteric musculomotor neurons that express VIP and/or

nitric oxide synthase innervate the circular muscle coats of the stomach, intestines, gallbladder, and the various smooth muscle sphincters. Inhibitory musculomotor neurons appear as an evolutionary adaptation for neural control of the specialized self-excitatory properties of the musculature. Electrically conducting junctions connect the smooth muscle fibers one to another to form a functional electrical syncytium that is reminiscent of cardiac muscle. Like cardiac muscle, action potentials and pacemaker potentials spread from muscle fiber to muscle fiber in three dimensions and trigger a contraction as they enter each successive muscle fiber. An interconnected network of non-neuronal pacemaker cells, known as interstitial cells of Cajal (ICCs), extends around the circumference and throughout the longitudinal axis of the small and large intestine.124 The ICC network generates electrical pacemaker potentials (also called electrical slow waves) that trigger action potentials and their associated contractions in the intestinal circular muscle coat. The pacemaker potentials spreading from the ICC networks become an extrinsic stimulus to which the circular muscle responds. The physiologic characteristics of the musculature as a self-excitable electrical syncytium suggest that the pacemaker network should continuously evoke contractions that spread in three dimensions throughout the extent of the syncytium, which in effect is the entire length of the intestine. Nevertheless, in the normal bowel, long stretches of intestine can exhibit no contractile activity and appear as if the musculature were in a paralytic state. Ongoing firing of impulses by enteric inhibitory motor neurons accounts for the occurrence of contractile silence (i.e., physiologic ileus) in the self-excitable electrical syncytium that forms the intestinal circular muscle coat. The circular muscle is able to respond to a pacemaker potential only when the inhibitory musculomotor neurons in a segment of intestine are switched to an inactive state by input from other neurons in the interneuronal control circuits. When the inhibitory motor neurons are firing and releasing their inhibitory neurotransmitters at the neuromuscular junctions, the muscle cannot be excited to contractile threshold by the omnipresent ICC pacemaker activity. Likewise, action potentials and associated contractions can only propagate into regions of the musculature where the inhibitory innervation is “turned off.” Consideration of the physiologic properties of the musculature suggests that inhibitory musculomotor neurons and control of their activity by the integrative microcircuits of the ENS have evolved as the mechanism that determines when the ongoing slow waves initiate a contraction, as well as the distance and direction of propagation once the contraction has been initiated. Inhibitory motor neurons to the circular muscle discharge continuously, with action potentials and contractions in the muscle occurring only when the inhibitory

Neurobiology of the Enteric Nervous System

neurons are “switched off” by input from interneurons in the control circuits. In the various smooth muscle sphincters found along the digestive tract, the inhibitory motor neurons are normally quiescent and are switched to an active state with timing appropriate for coordination of the opening of the sphincter with physiologic events in adjacent regions. When this occurs, the inhibitory neurotransmitter relaxes ongoing muscle contraction in the sphincteric muscle and prevents excitation-contraction in the adjacent muscle from spreading into and closing the sphincter. In nonsphincteric circular muscle, the state of activity of inhibitory motor neurons determines the length of a contracting segment by controlling the distance of spread of action potentials within the three-dimensional electrical geometry of the syncytium. Contraction can occur in segments in which ongoing inhibition has been switched off, while adjacent segments with continuing inhibitory neuronal input cannot contract. The boundaries of the contracted segment reflect the transition zone from inactive to active inhibitory musculomotor neurons. The directional sequence in which the inhibitory motor neurons are switched off establishes the direction of propagation of the contraction. Normally, they are switched off in the aboral direction, resulting in contractile activity that propagates in the aboral direction. In the abnormal conditions associated with emesis, the interneuronal microcircuitry must switch off the inhibitory musculomotor neurons in a reversed sequence to account for small intestinal propulsion that travels toward the stomach.74 Observations of continuous patterned discharge of action potentials by some of the myenteric neurons of dog, cat, guinea pig, and rabbit small intestine were early evidence for the continuous neuronal inhibition of the circular muscle.167 The continuous release of the inhibitory neurotransmitters VIP and nitric oxide coincident with the continuous inhibitory activity has been described for the canine small intestine.82,159 Spontaneously occurring inhibitory junction potentials that reflect the ongoing release of the inhibitory neurotransmitter can sometimes be recorded from the muscle. Nevertheless, in most preparations studied electrophysiologically in vitro, the ongoing inhibitory activity is manifest in the muscle as a steady hyperpolarization of the membrane potential and decreased input resistance. This reduces the probability that electrical pacemaker current from the ICCs will depolarize the muscle to the threshold for action potential discharge and the accompanying contraction. The steady inhibitory hyperpolarization probably reflects smoothing of many inhibitory junction potentials as a result of (1) the release of transmitter from multiple sites during asynchronous discharge of different nerve fibers, (2) long diffusion distances from transmitter release sites to the muscle, and (3) the electrical syncytial properties of the muscle. In general, any treatment or condition that ablates enteric inhibitory musculomotor neurons results in tonic

259

contracture and “achalasia” of the intestinal circular muscle. Several circumstances that involve functional ablation of the intrinsic inhibitory neurons are associated with conversion from a hypoirritable condition of the circular muscle to a hyperirritable state. These are experimental application of local anesthetics164; hypoxic vascular perfusion of an intestinal segment59; surgical ablation130; congenital absence, as in Hirschsprung’s disease165; paraneoplastic syndrome131; and inflammatory neuropathy.73 Most evidence suggests that subpopulations of inhibitory musculomotor neurons are tonically active, and that blockade or ablation releases the circular muscle from the inhibitory influence. The behavior of the muscle in these cases is tonic contracture and disorganized phasic contractile activity reminiscent of fibrillation in cardiac muscle.

DISINHIBITORY MOTOR DISORDERS The neuromuscular physiology of the intestine predicts that spasticity and achalasia will accompany any condition in which ablation of inhibitory musculomotor neurons occurs. Without inhibitory control, the selfexcitable syncytium of nonsphincteric regions will contract continuously and behave as an obstruction. This happens because the muscle is freed to respond to the pacemaker with contractions that propagate without amplitude, distance, or directional control. Contractions spreading in the uncontrolled syncytium collide randomly, resulting in fibrillation-like behavior in the affected intestinal segment. Loss or malfunction of inhibitory musculomotor neurons is the pathophysiologic basis of disinhibitory motor disease. It underlies several forms of chronic intestinal pseudo-obstruction and sphincteric achalasia (Table 12-1). Neuropathic degeneration in the ENS is a progressive disease that, in its earlier stages, can be manifest as symptoms that may be confused with IBS.

Chronic Intestinal Pseudo-obstruction Intestinal pseudo-obstruction is failure of intestinal propulsive motility in the absence of any mechanical form of obstruction. Both myopathic and neuropathic forms of

Table 12–1. Disinhibitory Motor Disease Pseudo-obstruction Chagas’ disease Paraneoplastic syndrome Hirschsprung’s disease Idiopathic Achalasia of the lower esophageal sphincter Biliary dyskinesia (sphincter of Oddi achalasia)

260

Function of the Peripheral Nervous System

chronic intestinal pseudo-obstruction are recognized. Degenerative changes in the musculature underlie the myopathic form, whereas neuropathic forms arise from degenerative changes in the ENS. Failure of propulsive motility in the affected length of bowel reflects loss of the neural microcircuits that program and control the repertoire of motility patterns required for the necessary functions of that region of bowel. Pseudo-obstruction occurs in part because contractile behavior of the circular muscle is hyperactive but disorganized in the denervated regions.140 Hyperactive-disorganized contractile activity, recorded with manometric catheters, is a diagnostic sign of the neuropathic form of chronic small bowel pseudo-obstruction in humans. The hyperactive and disorganized contractile behavior reflects the absence of inhibitory nervous control of the muscles that are self-excitable (autogenic) when released from the braking action imposed by inhibitory motor neurons. Chronic pseudo-obstruction is therefore symptomatic of the advanced stage of a progressive enteric neuropathy. Retrospective review of patients’ records suggest that some IBS-like symptoms can be an expression of early stages of the neuropathy.30,62,137 Degenerative noninflammatory and inflammatory ENS neuropathies are two kinds of disinhibitory motor disorders that culminate in pseudo-obstruction. Noninflammatory neuropathies can be either familial or sporadic.16 The mode of inheritance can be autosomal recessive or dominant. In the former, the neuropathologic findings include a marked reduction in the number of neurons in both myenteric and submucosal plexuses, and the presence of round, eosinophilic intranuclear inclusions in approximately 30% of the residual neurons. Histochemical and ultrastructural evaluations reveal the inclusions to be not viral particles, but rather proteinaceous material forming filaments.92,116 Members of two families have been described with intestinal pseudo-obstruction associated with the autosomal dominant form of ENS neuropathy.88,123 The numbers of enteric neurons were decreased in these patients, with no alterations found in the CNS or parts of the autonomic nervous system outside the gut. Degenerative inflammatory ENS neuropathies are characterized by a dense inflammatory infiltrate confined to enteric ganglia. Paraneoplastic syndrome, Chagas’ disease, and idiopathic degenerative disease are recognizable forms of pseudo-obstruction related to inflammatory neuropathies. Paraneoplastic syndrome is a form of pseudo-obstruction in which commonality of antigens between small cell carcinoma of the lungs and enteric neurons leads to autoimmune attack that results in loss of neurons.131 The majority of patients with symptoms of pseudo-obstruction in combination with small cell lung carcinoma have immunoglobulin G autoantibodies that react with their enteric neurons.78 These antibodies react with a variety of molecules expressed on the nuclear membrane of neurons, including

Hu and Ri proteins, and with cytoplasmic antigens (Yo protein) of Purkinje cells in the cerebellum.2,46,62,122 Immunostaining with sera from paraneoplastic patients shows a characteristic pattern of staining in enteric neurons.2 The detection of antienteric neuronal antibodies is a means to a specific diagnosis; nevertheless, the mechanisms by which the antibodies damage the neurons are unresolved. The association of enteric neuronal loss and symptoms of pseudo-obstruction in Chagas’ disease also reflects autoimmune attack on the neurons with accompanying symptoms that mimic the situation in achalasia and paraneoplastic syndrome. Trypanosoma cruzi, the blood-borne parasite that causes Chagas’ disease, has antigenic epitopes similar to enteric neuronal antigens.163 This antigenic commonality activates the immune system to assault the ENS coincident with its attack on the parasite. Idiopathic inflammatory degenerative ENS neuropathy occurs unrelated to neoplasms, infectious conditions, or other known diseases.30,73,137 De Giorgio et al.30 and Smith et al.137 described two small groups of patients with early complaints of symptoms similar to IBS that progressively worsened and were later diagnosed as idiopathic degenerative inflammatory neuropathy based on full-thickness biopsies taken during exploratory laparotomy that revealed chronic intestinal pseudo-obstruction. Each patient had inflammatory infiltrates localized to the myenteric plexus. Sera from the two cases reported by Smith et al.137 contained antibodies against enteric neurons similar to those found in secondary inflammatory neuropathies (i.e., antiHu), but with different immunolabeling patterns characterized by prominent cytoplasmic rather than nuclear staining. Recognition of the brain-like functions of the ENS leads to the conclusion that early neuropathic changes are expected to be manifest as functional symptoms that worsen with progressive neuronal loss. In diagnostic motility studies (e.g., manometry), degenerative loss of enteric neurons is reflected by hypermotility and spasticity140 because inhibitory motor neurons are included in the missing neuronal population.

Sphincteric Achalasia Sphincteric achalasia is one of the disinhibitory gastrointestinal motor disorders. Smooth muscle sphincters provide functional separation of the various specialized compartments of the digestive tract. The lower esophageal sphincter isolates the esophagus from the acidic environment of the stomach. The pyloric sphincter is a barrier to reflux from the duodenum into the stomach. The sphincter of Oddi guards against reflux from the duodenum into the biliary and pancreatic ducts. The internal anal sphincter contributes to the maintenance of fecal continence. Tonic contracture, which is an inherent myogenic property

Neurobiology of the Enteric Nervous System

of the sphincteric musculature, maintains closure of the orifices that separate the compartments. Enteric inhibitory musculomotor neurons innervate the smooth musculature that forms the sphincters. The function of the inhibitory innervation is to release inhibitory transmitters (e.g., VIP and nitric oxide) that act on the muscle to relax contractile tone and thereby open the sphincter. Inhibitory musculomotor neurons are the motor component of lower esophageal sphincter relaxation during swallowing, relaxation of the sphincter of Oddi to admit bile and pancreatic secretions into the duodenum, and internal anal sphincter relaxation for defecation and passing of flatus. Inhibitory control of sphincteric musculature differs from the tonically active inhibitory outflow to the intestinal musculature. Inhibitory musculomotor neurons to the sphincters are normally silent and are transiently activated with appropriate timing for opening of the sphincter and passage of luminal contents from one compartment to another. The sphincter is opened when the inhibitory musculomotor neurons are stimulated to firing threshold by excitatory synaptic input from interneurons in the ENS microcircuits. Closure of the sphincter returns on termination of the excitatory input and halt in firing of the inhibitory neurons. The requirement for presence of inhibitory musculomotor neurons to open sphincters predicts that ablation of the ENS will lead to sphincteric achalasia. Failure of development of the fetal ENS, including its population of inhibitory musculomotor neurons, accounts for the constricted terminal large intestinal segment and failure of relaxation of the internal anal sphincter in congenital megacolon (i.e., Hirschsprung’s disease).165 Achalasia in the lower esophageal sphincter leads to dilatation of the esophageal body and dysphagia. Achalasia in this group of patients is associated with loss of the inhibitory innervation of the sphincter and the presence of antimyenteric antibodies in the patient’s serum.160 The situation is similar for achalasia in the sphincter of Oddi, which leads to distention in the biliary tree and the characteristic pain localized in the right upper body quadrant. The hormone CCK is used in an effective test of the functional state of the inhibitory innervation of the sphincter of Oddi in patients with biliary pain of undiagnosed origin.55,121 Endoscopic placement of manometric catheters in the sphincter records changes in resting pressure in response to intravenous injection of CCK. Administration of CCK evokes relaxation in a normally innervated sphincter because excitatory receptors for CCK are expressed by the inhibitory musculomotor neurons that innervate the sphincteric smooth muscle. Conversely, CCK evokes a paradoxical contraction and increase in intraluminal pressure in a denervated sphincter. This occurs because the smooth muscle also expresses excitatory receptors for CCK, and contraction of the sphincter results from direct

261

stimulation of the muscle in the absence of input from inhibitory musculomotor neurons.

NEURONAL ELECTRICAL BEHAVIOR Two types of enteric neurons are distinguished by their electrophysiologic behavior in studies in which electrical activity is recorded with intracellular microelectrodes. The two types are identified as AH/type 2 and S/type 1. AH/type 2 and S/type 1 are an arbitrary combination of terms that recognize Nishi and North102 and Hirst et al.48 as the first to publish electrophysiologic descriptions of the two types of neurons. The names combine the alphabetical terms used by Hirst et al. and the numerical designations of Nishi and North. The terms are facilitating because AH/type 2 neurons usually have Dogiel type II multipolar morphology and S/type 1 neurons generally are unipolar like Dogiel type I neurons. AH/type 2 neurons are distinguished electrophysiologically by (1) higher resting membrane potential and lower input resistance than S/type 1 neurons; (2) no spike discharge to depolarizing current injection or discharge of one or two spikes only at the onset of intraneuronal injection of long-duration depolarizing current pulses (Fig. 12–7A) (3) absence of anodal-break excitation at the offset of hyperpolarizing current pulses; (4) prolonged postspike hyperpolarizing potentials; (5) Ca2⫹ contribution to the inward current of the action potential; and (6) tetrodotoxin-resistant action potentials. In addition, exposure to multivalent cationic Ca2⫹ entry blockers such as Mn2⫹, Mg2⫹, or Cd2⫹ depolarizes the neurons, increases input resistance, and augments excitability; and activation of adenylate cyclase and elevation of intracellular cyclic AMP (cAMP) depolarizes the cells, increases input resistance, and augments excitability. Aside from electrophysiologic properties, most AH neurons (⬃80%), but not S neurons, express the calcium-binding protein calbindin.60 AH neurons comprise the largest proportion of neurons in the intestinal myenteric plexus (⬃70%) and the smallest proportion of submucosal plexus neurons (⬍10%). Most of the multipolar AH-Dogiel type II neurons in the intestinal myenteric plexus project one of their long processes out of the ganglion to pass through the circular muscle coat and enter the mucosa (see Fig. 12–5). Because these terminals of AH-type neurons in the mucosa respond to release of serotonin from enterochromaffin cells in the mucosal epithelium, and because mechanical stimulation of the epithelial lining the lumen activates enterochromaffin cells to release serotonin, there has been an inclination toward calling them “intrinsic primary afferent neurons” and referring to them with the acronym IPANs.36,70 The term suggests sensory function. Nevertheless, the behavior of AH neurons overall does not fit the classic neurophysiologic descriptions of primary sensory afferent neurons

262

Function of the Peripheral Nervous System

A

B 20 mV

200 msec

C

20 mV

Resting Membrane Potentials

200 msec Action potential

Fast (⬍50-ms duration) nicotinic excitatory postsynaptic potentials (EPSPs) are evoked experimentally in most all S-type neurons. S-type neurons are more likely to show spontaneously occurring spike discharge than AH-type neurons. Nevertheless, AH neurons discharge spontaneously when activated by excitatory neurotransmitters or paracrine signals (see Slow Synaptic Excitation below).

10 mV

After-hyperpolarization

4 sec

FIGURE 12–7 Enteric neurons are classified as S/type I or AH/type II based on their electrophysiologic behavior. A, AH/type II neurons had low excitability reflected by discharge of only a single action potential during intraneuronal injection of a 200-ms depolarizing current pulse. B, The S/type I neuron had relatively high excitability reflected by repetitive discharge of action potentials during intraneuronal injection of a 200-ms depolarizing current pulse. The bottom traces in A and B show the current pulses and the top traces shows the neuronal responses to the depolarizing current pulses. C, AH/type II neurons are characterized by long-lasting membrane hyperpolarization (i.e., afterhyperpolarization) following the discharge of an action potential.

(i.e., dorsal root and nodose ganglion cells). “Intrinsic primary afferent” is therefore a confusing misnomer for students of neurophysiology. AH neurons are a unique population of neurons in the mammalian nervous system and would best be called “AH/Dogiel type II” or simply “enteric AH” to convey their functional significance in neural control of the bowel and to retain the term that has been used consistently since being applied first by Hirst et al. in 1974.48 S/type 1 neurons are distinguished by (1) lower resting membrane potentials than AH neurons; (2) higher input resistance than AH neurons; (3) elevated excitability reflected by repetitive spike discharge throughout longduration depolarizing current pulses (Fig. 12–7B); (4) increases in frequency of repetitive discharge in direct relation to increases in amplitude of experimentally evoked membrane depolarization; (5) anodal-break excitation (i.e., action potential discharge) at the offset of intraneuronally injected hyperpolarizing current pulses, another reflection of elevated excitability; (6) abolition of somal spikes by tetrodotoxin; (7) insensitivity to stimulation of adenylate cyclase by forskolin and elevation of cAMP; and (8) insensitivity to multivalent cationic Ca2⫹ entry blockers such as Mn2⫹, Mg2⫹, or Cd2⫹.

Potassium conductance is the main determinant of the resting membrane potential in enteric neurons. The resting potential is usually less than the potassium equilibrium potential, which is approximately ⫺90 mV. In AH neurons, a component of the resting potassium conductance, and consequently the resting potential, are dependent on the concentration of free intracellular Ca2⫹.1,42,110,154 A steady influx of Ca2⫹ is responsible for elevated intraneuronal levels that maintain Ca2⫹-activated K⫹ channels in an open state. This is reflected by high K⫹ conductance, lowered input resistance, and hyperpolarized membrane potential in AH neurons in the absence of any exposure to excitatory chemical signals. Neurotransmitters and paracrine neuromodulators act to increase or decrease the Ca2⫹-activated K⫹ conductance, and this is the key mechanism for up and down modulation of excitability and input-output relations in AH neurons. The functional significance of resting potentials less than the K⫹ equilibrium potential is provision of a mechanism whereby the membrane potential can be modulated in the hyperpolarizing or depolarizing direction as determined by whether the messenger substance acts to increase or decrease the K⫹ conductance. In enteric neurons, inhibitory signal substances such as opioid peptides, galanin, and adenosine decrease neuronal excitability by increasing K⫹ conductance and hyperpolarizing the membrane.95,96,115 Excitatory messengers such as substance P, serotonin, and histamine decrease resting potassium conductance, depolarize the membrane, and enhance excitability.66,69

Action Potentials Action potential generation in the cell bodies of AH neurons differs from spike-generating mechanisms in the cell bodies of S neurons and the AH neurons’ own neurites. The ionic mechanism of action potential generation in the cell bodies of AH neurons includes conductance changes for Ca2⫹, Cl⫺, Na⫹, and K⫹. Both Na⫹ and Ca2⫹ carry the inward current of the rising phase of the spike. A characteristic “shoulder,” reminiscent of the plateau on cardiac action potentials, is present at the onset of the falling phase of the spike in AH neurons.153 The “shoulder” reflects activation of voltage-gated Ca2⫹ conductance in N-type, highvoltage–activated Ca2⫹ channels3 as the membrane is

Neurobiology of the Enteric Nervous System

depolarized by inward Na⫹ current. The Na⫹ channels are typical of tetrodotoxin-sensitive channels found in neurons elsewhere.186 Application of tetrodotoxin reduces the rate of rise, the amplitude, and the threshold, but does not abolish the action potentials in AH neurons. The rate of rise of the spike in tetrodotoxin is increased by elevation of external Ca2⫹, and the pure Ca2⫹ spike in this case is abolished by multivalent ions that block Ca2⫹ entry.54,94,103,153,166 Dihydropyridines and other organic calcium entry blockers that suppress Ca2⫹ conductance in smooth and cardiac muscles do not affect the Ca2⫹ component of the spike in AH neurons.176 AH type neurons possess the full compliment of voltage-gated K⫹ currents generally found in neurons elsewhere. These include A-type, delayed rectifier, and inwardly rectifying currents.141,185 The falling phase of the spike is associated with time- and voltage-dependent activation of delayed rectifier K⫹ channels. Long-lasting hyperpolarizing afterpotentials are the hallmark of the action potentials in AH neurons and the basis for the term AH neuron (see Fig. 12–7B). The afterhyperpolarization activates slowly, beginning from 45 to 80 ms after termination of the spike, and lasts for up to 30 s. The amplitude of the afterpotential summates when two or more spikes are fired in close sequence. An increase in membrane conductance reflected by a decrease in the input resistance occurs during the hyperpolarizing afterpotentials. The amplitude of the afterpotential is increased by elevation of extracellular Ca2⫹ and is suppressed by multivalent ions that block Ca2⫹ entry. The amplitude is reduced and the duration of the hyperpolarizing afterpotential is shortened in bathing media with reduced Ca2⫹.103 The polarity of the hyperpolarizing afterpotential is reversed with current clamp of the membrane potentials to values greater than the K⫹ equilibrium potential of approximately ⫺90 mV. These early findings were evidence that an outward current carried by K⫹ ions generated the hyperpolarizing afterpotential. Ca2⫹-activated K⫹ channels carry the outward K⫹ current responsible for the hyperpolarizing afterpotential. Evidence for this comes from voltage-clamp results, which suggest that the slow outward current is coupled temporally to inward Ca2⫹ currents that are activated by depolarizing voltage steps.50 The amplitude of the current responsible for the hyperpolarizing afterpotential is a direct function of the amount of Ca2⫹ that enters the cell during depolarizing clamp steps or as the result of discharge of action potentials. Further evidence for the Ca2⫹ dependence of K⫹ current activation is the suppression of the current and the associated hyperpolarization produced by application of multivalent cations that block Ca2⫹ entry.42,54,103,109 This is consistent with the results obtained with optical imaging of intraneuronal Ca2⫹, which show elevations of intraneuronal free Ca2⫹ during the hyperpolarizing afterpotential.154

263

The hyperpolarizing afterpotentials function to lengthen the refractory period following the spike, and this automatically limits the frequency of spike discharge by the cell body. During slow synaptic excitation (discussed below), the neurotransmitter or paracrine modulator acts to reduce the hyperpolarizing afterpotential, and this allows the somal membrane to fire repetitively at higher frequency. Modulation of the afterpotential by chemical messengers is part of the overall mechanisms by which the excitability and input-output relations of AH neurons are governed.153 Somal action potentials in S-type neurons and spike generation by neurites of both AH- and S-type neurons are generated by classic mechanisms of activation and inactivation of time- and voltage-dependent Na⫹ and K⫹ conductance channels. The spikes are always abolished by sufficient concentrations of tetrodotoxin, and the rate of rise and amplitude are reduced with depleted Na⫹. Unlike AH neurons, mechanisms for generation of repetitive spike discharge are always activated and the cells will often discharge spontaneously, with the spikes preceded by ramplike prepotentials. In contrast to AH neurons, S neurons fire continuously during long-lasting depolarizing current pulses (see Fig. 12–7B), and the frequency of discharge is related directly to the size of the depolarizing pulse. Prominent positive afterpotentials are associated with the spikes and are not followed by Ca2⫹-dependent afterspike hyperpolarizing potentials like those in AH neurons. Unlike AH neurons, depletion of extracellular Ca2⫹ or activation of adenylate cyclase by forskolin does not elevate excitability. S-type electrophysiologic behavior is a characteristic of musculomotor and secretomotor neurons.

SYNAPTIC TRANSMISSION Fundamental mechanisms for chemically mediated synaptic transmission in the ENS are the same as elsewhere in the nervous system. Synaptic transmitters are released by Ca2⫹-triggered exocytosis from stores localized in vesicles at axonal terminals or transaxonal varicosities. Release is triggered by the depolarizing action of action potentials when they arrive at the release site and open voltageactivated Ca2⫹ channels. Once released, enteric neurotransmitters bind to their specific postsynaptic receptors to evoke inotropic or metabotropic synaptic events. When the receptors are directly coupled to the ionic channel, they are classified as inotropic. They are metabotropic receptors when their effects to open or close ionic channels are indirectly mediated by GTP-binding proteins and the induction of cytoplasmic second messengers (e.g., cAMP). Synaptic events in the ENS are basically the same as in the brain and spinal cord. EPSPs, inhibitory postsynaptic potentials (IPSPs), and presynaptic inhibition and

264

Function of the Peripheral Nervous System

facilitation are the principal synaptic events in the ENS. An enteric neuron may express mechanisms for both slow and fast synaptic neurotransmission. Fast synaptic potentials have durations in the millisecond range; slow synaptic potentials last for several seconds or minutes. Fast synaptic potentials are usually EPSPs. The slow synaptic events may be either EPSPs or IPSPs.

Fast Excitatory Postsynaptic Potentials Fast EPSPs were reported for the earliest intracellular studies of myenteric neurons, but were found only in S neurons in the work of Hirst et al.48 and Nishi and North.102 Fast EPSPs are depolarizing responses with durations less than 50 ms (Fig. 12–8A). Fast EPSPs were later reported to occur in AH and S neurons in both myenteric and submucosal plexuses. They appear to be the sole mechanism of transmission between vagal efferents and enteric neurons.125 Most of the fast EPSPs are mediated by acetylcholine acting at nicotinic postsynaptic receptors. The actions of 5-HT at the 5-HT3 serotonergic receptor subtype and purine nucleotides at P2X purinergic receptors behave much like fast EPSPs, and it is possible that some fast EPSPs are serotonergic or purinergic.37,57 Responses A

Action potential Stimulus artififact

mediated by 5-HT3 serotonergic receptors are found on neurons in both plexuses throughout the gastrointestinal tract, including the gastric corpus.129 Results obtained with patch-clamp recording methods show that the nicotinic and 5-HT3 serotonergic receptors are directly coupled to nonspecific cationic channels. Opening of these channels is responsible for the depolarizing event.32 Rapid desensitization is a characteristic of both the nicotinic and 5-HT3 operated channels, both of which are inotropic. Purinergic “fast” depolarizing responses are metabotropic and mediated by second messenger function of phospholipase A and elevation of inositol trisphosphate (IP3).37 Amplitudes of nicotinic fast EPSPs in the intestine become progressively smaller when they are evoked repetitively by focal electrical stimulation applied to the surface of the ganglion or interganglionic fiber tract. Decrease in EPSP amplitude occurs at stimulus frequencies as low as 0.1 Hz, and the rate of decline is a direct function of stimulus frequency. Rundown of this nature does not occur at the synapses in the stomach129,149 or gallbladder.85 The rundown phenomenon reflects presynaptic inhibition of acetylcholine release by additional transmitter substances broadly released by the electrical stimulus or by negative feedback involving autoinhibition of acetylcholine release B Stimulus artififact

IPSP

EPSPs

10 mV 10 mV

0.5 sec

10 msec

C

On Off Stimulus

40 mV 20 sec

FIGURE 12–8 Fast and slow excitatory postsynaptic potentials (EPSPs) and slow inhibitory postsynaptic potentials (IPSPs) are the principal kinds of synaptic events in enteric neurons. A, Two fast EPSPs were evoked by successive stimuli and are shown as superimposed records. Only one of the EPSPs reached threshold for discharge of an action potential. B, Slow IPSP evoked by stimulation of a sympathetic noradrenergic input to the neuron. C, Slow EPSP evoked by repetitive electrical stimulation of the synaptic input to the neuron. Slowly activating depolarization of the membrane potential continues for more than 2 minutes after termination of the stimulus. Repetitive discharge of action potentials reflects enhanced neuronal excitability during the slow EPSP.

Neurobiology of the Enteric Nervous System

(see Presynaptic Inhibition below). Rundown cannot be attributed to postsynaptic changes, because no decrease in the amplitude of the EPSPs occurs during repetitive applications of acetylcholine from microejection pipettes. Fast EPSPs function in the rapid transfer and transformation of neurally coded information between the elements of the enteric microcircuits. They are the bytes of information in the information-processing operations of the logic circuits. One of the fast EPSPs in Figure 12–8A reached threshold for discharge of an action potential, whereas the other EPSP did not reach threshold. Fast EPSPs do not reach threshold when the neuronal membranes are hyperpolarized during slow IPSPs. They are most likely to reach spike threshold when the membranes are depolarized during slow EPSPs. This effect of slow EPSPs and of slow EPSP-like paracrine mediators is an example of neuromodulation whereby the input-output relations of a neuron to one input is modified by a second synaptic or other kind of modulatory input.

Slow Synaptic Excitation Focal electrical stimulation applied experimentally to interganglionic fiber tracts or to the surfaces of ganglia in both myenteric and submucosal plexuses evokes slow EPSPs in both AH and S neurons. Slow EPSPs in AH neurons are accompanied by conversion from hypoexcitability with no action potential discharge to hyperexcitability with high-frequency discharge of action potentials. Neurons with slow EPSPs are found in the small and large intestine and gastric antrum, but not the gastric corpus or gallbladder. Slow EPSPs are restricted to the enteric microcircuits of specialized digestive compartments where propulsive motility is a significant function. Slowly activating membrane depolarization continuing for several seconds to minutes after termination of release of the neurotransmitter from the presynaptic terminal underlies the slow EPSP (Fig. 12–8C). Enhanced excitability reflected by long-lasting trains of action potentials is the hallmark of the event. Enhanced excitability during the slow EPSP is apparent experimentally as repetitive spike discharge during depolarizing current pulses and as anodal-break excitation at the offset of hyperpolarizing current pulses. AH neurons, which fire only a single spike at the beginning of a depolarizing current pulse in the inactivated state, will fire repetitively in response to depolarizing pulses when the slow EPSP is in effect. When activated by slow synaptic inputs, behavior of AH neurons is much like that of S neurons and may be confused as such if the AH neurons happen to be in an activated state as a result of ongoing release of the transmitter or a paracrine mediator (e.g., inflammatory/immune mediators). Postspike hyperpolarization in AH neurons is suppressed during slow EPSPs. Suppression of the afterhyperpolarization is part of the mechanism that permits

265

repetitive spike discharge at increased frequencies during the enhanced state of excitability. Slow EPSPs differ for AH neurons and S neurons. Slow EPSPs in AH neurons are generally associated with an increase in neuronal input resistance that reflects closure of ionic channels and decreased membrane ionic conductance. In S neurons, a large proportion of which are musculomotor or secretomotor neurons, the slow EPSPs are associated with a decrease in neuronal input resistance that reflects opening of ionic channels and increased membrane ionic conductance. Ionic Mechanisms The ionic mechanism for slow EPSPs in AH neurons includes changes in several ionic conductances. The depolarizing phase occurs when Ca2⫹ channels, normally open at rest, are closed.42,93,110 Closure of the Ca2⫹ channels lowers intraneuronal Ca2⫹, which, in turn, leads to closure of Ca2⫹-activated K⫹ channels. Closure of the K⫹ channels accounts for the increased input resistance seen in the neuron during the depolarization phase of the slow EPSP. Ca2⫹ currents during the spike that in the resting state would lead to postspike increase in Ca2⫹-activated K⫹ conductance are suppressed. This accounts for suppression of the afterhyperpolarization (i.e., the AH) during the EPSP. Enhanced excitability, seen as repetitive spike discharge and anodal break excitation during the EPSP, is probably related to suppression of A-type and delayed rectifier K⫹ currents.141 Unlike AH neurons, the ionic mechanism for the depolarization phase of slow EPSPs in S neurons includes elevated conductance for cations. Opening of Na⫹ and Ca2⫹ conductance channels accounts for the depolarization phase and decreased input resistance seen during the EPSP. Elevated Na⫹ conductance is the major contributor to the depolarization phase in the S neurons.56,58 Mediators of Slow EPSPs Several messenger substances found in neurons, endocrine cells, or immune cells mimic slow EPSPs when applied experimentally to enteric neurons. Receptors for more than one of the messenger substances may be present on the same neuron. Table 12–2 contains a current list of these substances together with receptor subtypes. Substance P (a tachykinin), 5-HT, and acetylcholine fulfill criteria for function as a neurotransmitter in the enteric microcircuits. The other substances are implicated mainly by their presence in enteric neurons and/or by mimicry of the slow EPSP when applied experimentally. Immunohistochemical studies find receptors for these substances, as well as combinations of receptors for the other listed substances, to be co-localized on the same enteric neuronal cell body. Histamine and interleukin-1␤ are examples of slow EPSP mimetics of paracrine origin. Histamine and

266

Function of the Peripheral Nervous System

Table 12–2. Slow EPSP Mimetics Acetylcholine (muscarinic/M1) Cholecystokinin (⫺A) Bombesin Calcitonin gene–related peptide Thyrotropin-releasing hormone 5-Hydroxytryptamine (5-HT1P) Norepinephrine Pituitary adenylate cyclase–activating peptide Interleukin-1␤ Histamine (H2 ) Adenosine (A2)

Vasoactive intestinal peptide Cerulein Gastrin-releasing peptide Tachykinins (NK3 /NK1) Corticotropin-releasing hormone ␥-Aminobutyric acid Motilin Glutamate (group I metabotropic) Interleukin-6 Platelet activating factor Bradykinin (B2)

interleukin-1␤ are released from intestinal mast cells to become a neuromodulatory signal that is decoded by the enteric microcircuits. Mast cells have both detector and signal functions in enteric neuroimmune communication.173 They utilize the specificity and memory of the immune system to detect the presence of threatening antigens in the gut, and after doing so release histamine to signal the ENS to reprogram its output for secretory and propulsive motor behavior that effectively eliminates the antigenic threat from the intestinal lumen.173 The evidence for 5-HT is as complete as for any known neurotransmitter, including synthesis, storage, and release from enteric neurons. Nevertheless, the strongest evidence is that agents such as N-acetyl-5-hydroxytryptophyl-5-hydroxytryptophan amide, renzapride, and anti-idiotypic antibodies to 5-HT block both the slow EPSP-like actions of 5-HT and the slow EPSP in the same AH neurons.39 Aside from its role as a putative enteric neurotransmitter, 5-HT also has a paracrine signaling function. A large fraction of the body’s 5-HT is stored in enterochromaffin cells interspersed in the intestinal epithelial lining. Mechanical stimulation (e.g., shearing forces) at the mucosa or noxious stimulation (e.g., stimulant laxatives such as sennosides) releases the 5-HT, which may then reach receptors on AH neuronal projections in the mucosa, other enteric neural elements, and spinal and vagal sensory afferents.9,40 Bornstein et al.10 reviewed five criteria for transmitter function that were fulfilled by substance P: (1) pharmacologic evidence suggests the intramural release of substance P within intestinal segments; (2) the K⫹ conductance decrease produced by substance P and the slow EPSP is the same; (3) chymotrypsin, which digests substance P, reduces both the response to the peptide and the slow EPSP; (4) the widespread occurrence of slow EPSPs implies that terminals of the responsible axons synapse

with most cell bodies in the ganglion, as is the case for the multipolar (AH/Dogiel type II) neurons that contain substance P; and (5) slow EPSPs were evoked in myenteric neurons with ganglia that contained substance P, but no immunocytochemically demonstrable 5-HT–containing fibers. Slow EPSPs can be initiated by substance P released from AH neurons or from collateral projections of spinal afferents inside the intestinal wall.150 Neurokinin-3 receptors mediate the action of substance P.64,126 The slow EPSP-like action of 5-HT is related to the 5-HP1P serotonergic receptor, so named by Mawe and co-workers.87 This receptor subtype is blocked by the drug renzapride, which is a substituted benzamide compound with stimulatory effects on gastric emptying and intestinal transit.40 Slow EPSP mimetic action of acetylcholine is mediated by the M1 muscarinic receptor subtype,21,107 and the action of histamine occurs at the H2 receptor subtype.99,171 Motilin is a slow EPSP mimetic of particular interest because it is implicated as a messenger substance in the initiation of the intestinal migrating motor complex in the interdigestive period.158 Additional interest emerges from findings that macrolide antibiotics (e.g., erythromycin) act at motilin receptors to facilitate gastric emptying in disorders such as diabetic gastroparesis.148 Slow EPSP-like actions of motilin are prominent in AH neurons of the gastric antrum.178 Signal Transduction in AH Neurons Slow activation and the prolonged time course of slow EPSPs are clues to the mechanism of signal transduction. A 30-ms experimental “puff” of a slow EPSP mimetic or electrical stimulation of synaptic inputs for the slow EPSP evokes changes in excitability that last for several minutes.24,176 This happens in experiments in which the exposure is virtually limited to the 30-ms duration of the puff. The slow EPSP mimetics act transiently at discrete sites on the somal membrane, whereas closure of Ca2⫹-activated K⫹ channels in AH neurons and opening of cation channels in S neurons occurs globally. This global activation in the whole cell suggests involvement of an intraneuronal second messenger system that connects localized surface receptors (i.e., metabotropic receptors) to processes in the neuronal interior. Receptor occupancy by the messenger substance in this case stimulates intraneuronal synthesis or release of a second messenger, which then initiates the neurochemical reactions and molecular conformational changes responsible for transformation of the somal membrane from hypo- to hyperexcitability. Several lines of evidence indicate that receptor-mediated activation of adenylate cyclase, elevation of intraneuronal cAMP, and activation of protein kinase A is the signal transduction mechanism for the slow EPSP in AH neurons. Forskolin, a substance that activates adenylate cyclase, proved to be a useful tool for the study of signal transduction

Neurobiology of the Enteric Nervous System

in AH neurons. Application of forskolin elevates cAMP in the ganglia179 and mimics the slow EPSP in AH- but not in S-type neurons.100 Likewise, other treatments that elevate cAMP, such as intraneuronal injection of cAMP, application of membrane-permeant analogues of cAMP, and treatment with phosphodiesterase inhibitors, each produce slow EPSPlike effects.114 Treatment of myenteric ganglia with 5-HT181 or histamine180 stimulates cAMP formation. Stimulation of cAMP by 5-HT is mediated by 5-HT1P receptors and histamine action is mediated by the H2 receptor subtype. All of the evidence supports the hypothesis that cAMP is an intracellular second messenger in the process of signal transduction in AH neurons. Occupancy of receptors for the slow EPSP activates adenylate cyclase, which in turn leads to synthesis of cAMP, phosphorylation of protein kinases and/or membrane channel proteins, and eventually the dramatic changes in neuronal excitability that occur during the slow EPSP in AH neurons. Receptors for several different kinds of mediators are coupled by G proteins to adenylate cyclase to initiate a single postreceptor cascade of events that culminates in the slow EPSP. Experimental use of forskolin improved insight into the behavior of AH neurons that were found to be inexcitable in their resting state.166 In this state, the neurons show hyperpolarized resting membrane potentials that are close to the potassium equilibrium potential (⬃⫺90 mV) and low input resistance indicative of high resting potassium conductance. Injection of depolarizing current does not evoke action potentials in this state. EPSPs that reflect inputs from other neurons in the circuit may be present but do not evoke action potentials in the inexcitable state. Application of forskolin or one of the endogenous slow EPSP mimetics restores excitability to the AH neurons. When this occurs, the neurons may discharge action potentials with hyperpolarizing afterpotentials if the concentration of the agonist is low, or they may discharge repetitively like an S neuron if exposed to higher concentrations. Neurons responding in this way are interpreted as interneuronal circuit elements that are not used continuously by the system. When the digestive state of the gut does not require operation of the section of circuitry that contains the neurons, the circuit is idled by the low excitability state of its component neurons. The circuit is called to activity by overlays of chemical modulators that act to boost the excitability of its neurons. Signal Transduction in S Neurons Early results suggest that signal transduction for the decreased resistance slow EPSPs in S neurons involves a phospholipase C/Ca2⫹-calmodulin second messenger mechanism. Receptors for the slow EPSP on S neurons are G protein coupled to activation of phospholipase C. Phospholipase C catalyzes the formation of IP3, which acts as a second messenger to release Ca2⫹ from intraneuronal

267

membrane stores. Binding of the released Ca2⫹ to calmodulin activates protein kinase C. Activated protein kinase C phosphorylates cation conductance channels that open when phosphorylated to increase conductance for Na⫹ and Ca2⫹ and thereby depolarize the membrane potential. Opening of the cationic conductance channels accounts for the decreased input resistance observed while recording with microelectrodes during the depolarization phase of the EPSP. Termination of the EPSP results from activation of the intraneuronal phosphatase calcineurin, which catalyzes dephosphorylation of the cationic channels.56 Significance of Slow EPSPs in the Integrated System The functional significance of slow EPSPs in AH neurons is twofold. First is an increased probability of impulse discharge in the postsynaptic neuron that is then transformed into excitation or inhibition at either the next-order neuron or an effector, such as the musculature or secretory epithelium. Intestinal peristaltic propulsion, for example, requires sustained discharge by neuronal elements in the microcircuits to account for the sustained neural drive of several seconds’ duration that occurs in the musculature. The behavior of AH neurons during slow EPSPs fits the requirements for a neuronal element whose functional significance in the circuit would be production of either prolonged excitation or inhibition of the intestinal circular or longitudinal muscle layers or a prolonged glandular secretory response. Application of the putative neurotransmitters for the slow EPSP stimulate the release of the neuromuscular or neuroepithelial neurotransmitters acetylcholine, tachykinins, VIP, and nitric oxide from myenteric and submucosal plexuses of intestinal segments in vivo and in vitro.61,161,183 This is consistent with the hypothesis that AH neurons are interneurons that supply coordinated synaptic drive to the excitatory and inhibitory motor innervation of the musculature and excitatory drive to the secretory epithelium. The extensive ramifications of the processes of the AH/Dogiel type II neurons to innervate large numbers of neurons in the same and neighboring ganglia (see Figs. 12–4B and 12–5) and the localization of substance P in these neurons63,170 is consistent with this function. Slow EPSPs underlie a gating mechanism that controls the spread of action potentials between the neurites arising from opposite poles of the cell bodies of multipolar AH neurons. Figure 12–9 illustrates how slow synaptic gating works in AH/Dogiel morphologic type II neurons. Electrophysiologic recording in AH neurons in the low excitability state shows that action potentials propagating toward the cell body in one of its processes do not fire the membrane of the cell soma. If the invading spike should fire the somal membrane, the action potential in the cell body will be followed by the characteristic afterhyperpolarization, which acts to prevent firing of the cell body by any

268

Function of the Peripheral Nervous System

Closed somal gate

Neurite [1]

Open somal gate

+

Action potential propagation [3]

[1]

Positive feed-forward synaptic circuit

[3]

+

Slow excitatory Synaptic input +

Slow EPSP opens somal gate

[2]

[4]

[2]

[4]

FIGURE 12–9 Slow EPSPs and plasticity of membrane excitability in the cell bodies of AH/Dogiel morphology type II neurons underlie a gating mechanism by which the spread of spike information among neurites, which arise from opposite poles of the cell body, is controlled. Left, When the neuron is in the resting state, excitability of the somal membrane is low, and there is low probability that inbound spike information arriving at the soma in neurite 1 will fire the somal membrane. The discharge is limited to one or a few spikes by the postspike hyperpolarizing afterpotentials (see Fig. 12–7C) in case the somal membrane is fired by an inbound spike. With the somal membrane in a hypoexcitable state, the somal gate is totally or partly closed and no or a restricted amount of information is relayed to the neurites at the opposite poles of the cell body. Right, During a slow EPSP, the probability that the somal membrane will be fired by inbound spikes in neurite 1 and that it will fire repetitively at increased frequency is greatly increased because excitability is enhanced, membrane resistance increased, and hyperpolarizing afterpotentials suppressed. The slow EPSP opens the somal gate so that inbound activity is transferred to neurites 2, 3, and 4 and relayed onto synapses with neighboring neurons (see multipolar morphology of AH neurons in Fig. 12–4B).

additional incoming spikes. The probability for the cell body to be fired by inbound spikes is greatly increased during the slow EPSP when excitability is enhanced, membrane resistance is increased, and afterhyperpolarization is suppressed. The membrane of the cell body behaves like a closed gate to the transfer of spike signals between its neurites when it is in the low-excitability state. In this state, spike information, such as inbound information from the mucosa (see Fig. 12–5), is confined to a single neurite (e.g., neurite 1 in Fig. 12–9). The gate is opened and signals are transferred across the somal membrane to other neurites during the slow EPSP. A closed somal gate isolates the initial segments of each neurite, such that spike discharge in one initial segment does not influence another neurite elsewhere around the cell body (see Figs. 12–4B and 12–5). The increase in membrane resistance during the slow EPSP increases the space constant of the somal membrane and facilitates electrotonic spread of the action potential from the neurites into the cell body. This, together with the enhanced excitability, opens the somal gate, resulting in transfer of signals across the cell body to neurites at

Slow excitatory Synaptic input

AH-Dogiel type II neurons

+ Circuit output Motor neurons Effector systems (musculature, glands, vasculature)

FIGURE 12–10 AH/Dogiel morphologic type II neurons are networked into positive feed-forward synaptic circuits. Feed-forward synaptic excitation is a mechanism for rapid initiation of synchronous discharge in the neural elements of the circuit. Output from the feedforward circuit drives populations of motor neurons, which in turn drive the behavior of the musculature, secretory epithelium, and vasculature. Positive feed-forward synaptic networks synchronize firing of large numbers of motor neurons to ensure simultaneous contraction or inhibition of the muscle around the circumference and along an extended length of bowel. Synchronization of firing of secretomotor neurons ensures coordination of secretion around and along an extended length of bowel.

other poles of the neuron. When this happens, all neurites fire in synchrony with the cell body and the action potentials are conducted away to be distributed in regions of the enteric networks lying further along or around the intestine. Somal gating within a pool of AH neurons that are synaptically interconnected in a feed-forward excitatory circuit (Fig. 12–10) is believed to be the mechanism for rapid buildup of action potential discharge in the neuronal pool. This results in simultaneous spike discharge, characteristic of the slow EPSP, in a population of AH neurons positioned around the circumference and along the length of an intestinal segment. Output from the network of AH neurons drives the discharge of the pools of excitatory and inhibitory musculomotor neurons and secretomotor neurons that innervate the intestine’s effector systems. In a segment of bowel, this mechanism ensures that muscular contraction or inhibition occurs simultaneously around the circumference and along the entire length of the segment.155,170 The same can be postulated for secretion in an intestinal segment. As would be the case also for muscular activity, an effective secretory response requires that the response be activated simultaneously around the circumference and along an extended length of intestine. These functional requirements are achieved by feed-forward

Neurobiology of the Enteric Nervous System

buildup of excitation in AH neuronal driver circuits that connect to secretomotor and musculomotor neuronal pools. The projectional geometry of the neurites identifies the significance of slow excitatory input to the cell body of AH neurons as a gating determinant of neurite-to-neurite communication across the cell soma. Most all myenteric AH neurons project one or more neurites to the mucosa (see Fig. 12–5). This multiplies their functional significance because the projections in the mucosa fire in response to acid pH and in response to serotonin released by mechanical stimulation of mucosal enterochromaffin cells.8,36,70 Figures 12–5, 12–9, and 12–10 illustrate the functional complexity contributed by AH neurons at the circuit level of organization. As coupled elements in the control circuits for the musculature and glandular epithelium, they are stimulated by slow synaptic inputs to fire prolonged trains of impulses. This kind of activity in the cell body is transformed into slow synaptic output to neighboring AH neurons in the circuit and results in feed-forward buildup of neuronal firing in the whole circuit. In the absence of slow synaptic input, the somal gate is closed (see Figs. 12–5 and 12–9). In this state, inbound information from the mucosa is not gated to neurites across the cell body, but may become the afferent arm of an axon reflex. In circumstances in which the cell body is activated by slow synaptic or paracrine input, the somal gate is open and the transformed information from the mucosa is gated in the direction of slow synaptic output from both sides of the cell soma, as illustrated in Figure 12–5.

Inhibitory Postsynaptic Potentials Slow IPSPs are hyperpolarizing synaptic potentials found in both myenteric and submucosal ganglion cell somas of the small and large intestine and in myenteric neurons of the gastric antrum.52,108,149,176,184 IPSPs activate slowly and continue for several seconds after termination of the stimulation (see Fig. 12–8B). Slow IPSPs are hyperpolarizing responses associated with decreased input resistance and suppression of excitability. Reduction in the membrane resistance and excitability together with membrane hyperpolarization decrease the probability of action potential discharge that might occur spontaneously in response to excitatory synaptic inputs or to inbound action potentials in a neurite. IPSPs are more readily demonstrated in the submucosal than in the myenteric plexus of in vitro preparations. Earlier studies found slow IPSPs in less than 10% of small intestinal myenteric neurons,67,176 whereas over 80% of submucosal neurons show IPSPs in response to electrical stimulation.108,142 Interaction with slow EPSPs counteracts the inhibitory inputs and is a factor accounting for the low incidence of IPSPs in myenteric plexus studies.

269

Experimental stimulation of interganglionic connectives activates both excitatory and inhibitory inputs, and the excitatory inputs predominate to a variable degree. Selective blockade of slow synaptic inputs by adenosine A1 receptor agonists uncovers and enhances stimulus-evoked slow IPSPs in the myenteric plexus.23 The decrease in input resistance observed during slow synaptic inhibition reflects increased conductance in potassium channels.49,91,134,144 Of the variety of potassium channels found in the membranes of enteric neurons, inwardly rectifying potassium channels are the ones opened by a variety of inhibitory neurotransmitters to account for the increased conductance and hyperpolarization of the membrane potential during the IPSP.49,50,91,144 The Ca2⫹-activated potassium channels that both contribute to the resting membrane potential and account for postspike hyperpolarization are not involved in the generation of the slow IPSPs.106,134,144,145 GTP-binding proteins are suggested to be involved in direct coupling of the receptors to the potassium channels.90,144 Mediators of Slow IPSPs Several putative neurotransmitters and paracrine signal substances evoke slow IPSP-like responses when experimentally applied to enteric neurons. Some of these substances are peptides, others are purinergic compounds, and another is norepinephrine (Table 12–3). Receptors for two or more of these substances may be localized to the cell body of the same neuron. Enkephalins, dynorphin, and morphine all mimic the IPSPs in enteric neurons. Actions of these substances are mediated by opiate receptors and are limited to subpopulations of neurons. Opiate receptors of the ␮ subtype predominate on myenteric neurons, whereas the receptors on submucosal neurons are the ␦-opioid subtype.111 Morphine addiction and effects of withdrawal in enteric neurons is observed as high-frequency discharge of impulses on the addition of naloxone during chronic exposure to morphine.105,106 IPSP-like actions of nociceptin involve an ORL1 receptor that is distinct from typical naloxone-sensitive opioid receptors.79 Table 12–3. Slow IPSP Mimetics Acetylcholine 5-Hydroxytryptamine (5-HT1A) Neurotensin Somatostatin Adenosine (A1) Nociceptin Opioid peptides Norepinephrine (␣2) Cholecystokinin ATP Galanin

270

Function of the Peripheral Nervous System

Norepinephrine binds to ␣2a adrenoceptors to evoke IPSPs. Noradrenergic IPSPs occur primarily in neurons of the intestinal submucosal plexus.108,144 The noradrenergic inputs come from postganglionic fibers of the sympathetic nervous system that are involved in inhibition of secretomotor activity. Galanin is a 29–amino-acid polypeptide that mimics slow IPSPs in almost all of the neurons of the intestinal myenteric and submucosal plexuses.113 Cholecystokinin does this also, but the effect is only seen in a subpopulation of 10% to 15% of myenteric neurons in the small intestine or gastric antrum.101,132,133 The application of adenosine, ATP, and other purinergic analogues simulates the slow IPSP in intestinal myenteric neurons.115 This is seen in nearly all AH neurons and results from the suppression of adenylate cyclase and reduction in intraneuronal levels of cAMP.182 Both the degradative enzyme adenosine deaminase and a selective adenosine A1 receptor antagonist potentiate slow EPSPs in vitro.23 Potentiation of the slow EPSPs reflects reduction of ongoing inhibitory action of endogenously released adenosine and its accumulation in the intestinal wall. Somatostatin is implicated as the endogenous inhibitory transmitter responsible for the intrinsic inhibitory inputs to submucosal ganglion cells. Hirst and Silinsky53 described IPSPs in the submucosal plexus that persisted after sympathectomy and could therefore not be attributed to activation of norepinephrine release from sympathetic postganglionic fibers. This IPSP, unlike the noradrenergic IPSPs, requires repetitive stimulation and occurs with a more slowly developing time course than noradrenergic IPSPs.80 Somatostatincontaining nerve terminals and the nonadrenergic IPSP disappear in parallel after destruction of the myenteric plexus and its projections to the submucosal plexus.177 Functional Significance of Slow IPSPs A reduction in the membrane resistance and excitability of the somal membranes, together with membrane hyperpolarization during the slow IPSPs, decrease the probability of action potential discharge. The probability is reduced that the somal membrane will be fired during electrotonic invasion by spikes in the initial axonal segment or during excitatory synaptic input (see Figs. 12–5 and 12–9). This influence is the inverse of the slow EPSP and acts to close the gate for transfer of spike information across the multipolar soma of AH neurons. Slow synaptic inhibition probably functions to terminate the excitatory state of slow synaptic excitation and reestablish the low excitability state in the ganglion cell soma of AH neurons. This may be a step in the control of sequentially occurring motor events such as the conversion from inhibition to excitation in the circular muscle of an intestinal segment during propagation of propulsive peristalsis. In the intact animal, there may also be inhibitory substances of endocrine or paracrine origin that function in

particular situations to lock the somal membranes in a low excitability state.

Presynaptic Facilitation Presynaptic facilitation refers to enhancement of synaptic transmission that results from actions of chemical mediators at neurotransmitter release sites on enteric axons. This is known to occur at fast excitatory synapses in the myenteric plexus of the small intestine and gastric antrum and at noradrenergic inhibitory synapses in the submucosal plexus. CCK mediates presynaptic facilitation in gallbladder ganglia.86 Presynaptic facilitation is evident as an increase in amplitude of fast EPSPs at nicotinic synapses, where it reflects enhanced release of acetylcholine from axonal release sites. At noradrenergic inhibitory synapses in the submucosal plexus, presynaptic facilitation appears as enhancement of the hyperpolarizing responses to stimulation of sympathetic postganglionic fibers and is associated with elevation of cAMP in the sympathetic nerve terminals.184

Presynaptic Inhibition Presynaptic inhibition refers to mechanisms that suppress release of neurotransmitters from axons. It involves binding of chemical messengers to inhibitory receptors at transmitter release sites on the axon. Presynaptic inhibition is a significant synaptic event within the enteric microcircuits of the gastric corpus and antrum, as well as the small and large intestine and rectum of the guinea pig.51,104,129,149,153,167 Presynaptic inhibitory receptors are found at fast and slow excitatory synapses, at inhibitory synapses, and at neuromuscular junctions. Presynaptic inhibition in many cases involves axoaxonal transmission in which release of a neurotransmitter from one axon acts at receptors on another axon to suppress release of transmitter from the second axon. Presynaptic inhibition also takes the form of autoinhibition and can occur at both neural synapses and neuroeffector junctions. Autoinhibition occurs when the transmitter released from the same enteric neuron accumulates in the vicinity of the release site and activates presynaptic inhibitory receptors to suppress further release. It functions in this way as a negative feedback mechanism that automatically regulates the concentration of neurotransmitter within the synaptic or junctional space. Substances of non-neural origin, released in a paracrine or endocrine manner into the milieu surrounding the synapse, are known to act presynaptically to suppress neurotransmission. Mast cells and enteroendocrine cells are sources of non-neuronal messages to presynaptic receptors in the ENS. Several messenger substances found in neurons, enteroendocrine cells, or immune cells of the gut produce presynaptic inhibition when applied experimentally to enteric synapses (Table 12–4). Norepinephrine acts at

271

Neurobiology of the Enteric Nervous System

Table 12–4. Mimetics for Presynaptic Inhibition Norepinephrine (␣2a) Histamine (H3) Opioid peptides Neuropeptide Y Peptide YY adenosine (A1) Dopamine 5-Hydroxytryptamine (5-HT4?) Acetylcholine (muscarinic) Pancreatic polypeptide Cytokines

presynaptic ␣2a receptors to suppress transmission at both slow and fast excitatory synapses and at excitatory neuroeffector junctions.144 The presynaptic inhibitory actions of norepinephrine are significant because it is the neurotransmitter released at the synaptic interface between postganglionic sympathetic axons and the ENS.172 Electrical stimulation of sympathetic postganglionic fibers in the periarterial mesenteric nerves suppresses fast EPSPs in myenteric ganglion cells without altering the responses to exogenously applied acetylcholine.51,102 Norepinephrine also fulfills the criteria for a presynaptic action by suppressing stimulus-evoked slow EPSPs without effect on the postsynaptic actions of substance P, serotonin, or acetylcholine.177 As expected for a presynaptic action, norepinephrine does not abort a slow EPSP in progress, but blocks ability to evoke subsequent EPSPs.177 Norepinephrine reduces release of 5-HT and substance P from isolated intestinal segments, consistent with presynaptic suppression of slow synaptic excitation.5,68 Electron micrographs show axoaxonal synapses with vesicular profiles characteristic of noradrenergic synapses in the guinea pig myenteric plexus.83 Observations of suppressive effects of norepinephrine on acetylcholine release from intestinal segments in vitro are supporting evidence for a presynaptic action.31,72,117 Histamine, interleukin-1␤, interleukin-6, and plateletactivating factor receive attention because of their role in neuroimmune communication between mast cells and the ENS program circuits. Histamine, like the other listed immune/inflammatory messengers, acts at presynaptic receptors on cholinergic axons to suppress fast EPSPs at nicotinic synapses and at sympathetic nerve terminals to suppress the release of norepinephrine in the ENS microcircuits.80,168,171 This action of histamine is blocked by selective H3 receptor antagonists and mimicked by selective H3 agonists. The evidence indicates that presynaptic action of histamine is at the H3 receptor subtype.80,152 Histamine released during antigenic degranulation of intestinal mast cells acts in like manner.34,35 Application of 5-HT to enteric synapses reduces the amplitude of the fast EPSPs without effects on nicotinic postsynaptic responses to acetylcholine, thereby fulfilling

criteria for serotonergic presynaptic inhibition.104 The receptor for the presynaptic inhibitory action is not unequivocally identified, but fits the profile of the 5-HT1 receptor in the gastric antrum and small intestine38,143,147 or the putative 5-HT4 receptor subtype in the colon.33 The presynaptic action of acetylcholine on fast EPSPs occurs at muscarinic receptors in the myenteric and submucosal plexuses of the small and large intestine107,153 and at myenteric synapses in the gastric antrum.149 The affinity of the presynaptic receptors for selective agonists and antagonists is suggestive of the M2 class of muscarinic receptor.107 The presynaptic muscarinic receptors at fast nicotinic synapses are components of autoinhibitory mechanisms that function in negative feedback regulation of the amount of acetylcholine released by the arrival of the action potential at the release site. Inhibition of acetylcholinesterase by drugs such as eserine results in suppression of stimulus-evoked fast EPSPs, and this effect is reversed by the application of atropine. In this situation, inhibition of the cholinesterase permits accumulation of acetylcholine, which feeds back on the presynaptic muscarinic receptors to suppress further release of acetylcholine. Atropine blocks this presynaptic influence of the esterase and restores the EPSP.153,167 Members of the pancreatic polypeptide family of messenger peptides (e.g., NPY, PYY, PP) act presynaptically to suppress transmission at nicotinic synapses in the microcircuits of the stomach.128 The presynaptic inhibitory action of NPY is significant because it is co-localized with norepinephrine in sympathetic postganglionic axons as well as being found in intrinsic neurons. This suggests that both the CNS and the ENS may use the long-lasting inhibitory action of this mediator to selectively inactivate synapses within the microcircuits. Presynaptic NPY receptors appear to be present on vagal efferent fibers in the gastric corpus and participate in circuit functions by which the ENS microcircuitry of the corpus may ignore its vagal input.128 Presynaptic inhibition by adenosine occurs at the A1 type of P1 purinoreceptor.4,22,23 This occurs at fast nicotinic synapses, slow excitatory synapses, and noradrenergic inhibitory synapses. Suppression of fast EPSPs occurs in the gastric as well as intestinal microcircuits, whereas presynaptic inhibition of IPSPs is seen in the intestinal submucosal plexus.

CONCLUSIONS Many lines of evidence now implicate dysfunction in the ENS as an important factor underlying symptomatology in patient complaints that fit criteria for functional gastrointestinal disorders. This underscores a need for expanded attention to the neural factors that might explain the

272

Function of the Peripheral Nervous System

functional disorders and justifies accelerated expansion of the basic and clinical investigation that defines the new subspecialty of neurogastroenterology. Consideration that the ENS is an independent integrative nervous system, with most of the neurophysiologic complexities found in the CNS, stimulates premonition that functional motor disorders may reflect neuropathies in the “brain-in-the-gut.”174 Lack of understanding of how subtle malfunctions may occur in the synaptic microcircuits of the ENS is undoubtedly the basis for the use of “functional” as the description for functional disorders such as nonulcer dyspepsia and IBS. The current state of understanding of functional gastrointestinal disorders is reminiscent of neurologic disorders such as Parkinsonian tremors, ballisms, and choreas that were classified as “functional” before understanding of neurotransmission in microcircuits of somatic motor centers in the brain. Parkinson’s disease is a classic example of a functional somatic motor disorder that was ultimately explained as malfunction of dopaminergic neurotransmission in localized brain centers. Like the status of understanding of somatic motor control centers in the brain a half-century ago, now as we enter the 21st century, the ENS remains a virtual “black box” that will need to be opened scientifically to understand gastrointestinal disorders more fully.

REFERENCES

9.

10.

11.

12. 13. 14.

15.

16.

17. 18. 19.

1. Akasu, T., and Tokimasa, T.: Potassium currents in submucous neurones of guinea-pig caecum and their synaptic modification. J. Physiol. (Lond.) 416:571, 1989. 2. Altermatt, H., Rodriguez, M., and Scheithauer, B.: Paraneoplastic anti-Purkinje and type 1 antineuronal nuclear autoantibodies bind selectively to central, peripheral autonomic nervous system cells. Lab. Invest. 65:412, 1991. 3. Baidan, L. V., Zholos, A. V., and Wood, J. D.: Modulation of calcium currents by G-proteins and adenosine receptors in myenteric neurones cultured from guinea-pig small intestine. Br. J. Pharmacol. 116:1882, 1995. 4. Barajas-Lopez, C., Surprenant, A., and North, R. A.: Adenosine A1 and A2 receptors mediate presynaptic inhibition and postsynaptic excitation in guinea pig submucosal neurons. J. Pharmacol. Exp. Ther. 258:490, 1991. 5. Bartho, L., Holzer, P., and Lembeck, F.: Sympathetic control of substance P releasing enteric neurones in the guinea pig ileum. Neurosci. Lett. 38:291, 1983. 6. Bearcroft, C. P., Perrett, D., and Farthing, M. J. G.: Postprandial plasma 5-hydroxytryptamine in diarrhoea predominant irritable bowel syndrome: a pilot study. Gut 42:42, 1998. 7. Berthoud, H. R., and Powley, T. L.: Vagal afferent innervation of the rat fundic stomach—morphological characterization of the gastric tension receptor. J. Comp. Neurol. 319:261, 1992. 8. Bertrand, P. P., Kunze, W. A. A., Bornstein, J. C., et al.: Analysis of the responses of myenteric neurons in the small

20.

21.

22.

23.

24.

25.

26. 27.

intestine to chemical stimulation of the mucosa. Am. J. Physiol. 36:G422, 1997. Beubler, E., and Schirgi-Degen, A.: Serotonin antagonists inhibit sennoside-induced fluid secretion and diarrhea. Pharmacology 47(Suppl. 1):64, 1993. Bornstein, J. C., North, R. A., Costa, M., and Furness, J. B.: Excitatory synaptic potentials due to activation of neurons with short projections in the myenteric plexus. Neuroscience 11:723, 1984. Bose, M., Nickols, C., Feakins, R., and Farthing, M. J.: 5-Hyroxytryptamine and enterochromaffin cells in the irritable bowel syndrome. Gastroenterology 118(Suppl. 2):A563, 2000. Brookes, S. J. H.: Classes of enteric nerve cells in the guinea-pig small intestine. Anat. Rec. 262:58, 2001. Brookes, S. J. H.: Retrograde tracing of enteric neuronal pathways. Neurogastroenterol. Motil. 13:1, 2001. Camilleri, M., Chey, W. Y., Mayer, E. A., et al.: A randomized controlled clinical trial of the serotonin type 3 receptor antagonist alosetron in women with diarrhea-predominant irritable bowel syndrome. Arch. Intern. Med. 23:1733, 2001. Camilleri, M., Northcutt, A. R., Kong, S., et al.: Efficacy and safety of alosetron in women with irritable bowel syndrome: a randomized, placebo-controlled trial. Lancet 355:1035, 2000. Camilleri, M., and Phillips, S.: Disorders of small intestinal motility. Gastroenterol. Clin. North Am. 18:405, 1989. Cervero, F.: Afferent activity evoked by natural stimulation of the biliary system in the ferret. Pain 13:137, 1982. Cervero, F., and Janig, W.: Visceral nocireceptor: a new world order? Trends Neurosci. 15:374, 1992. Chey, W. D., Kim, M., Hasler, W. L., and Owyang, C.: Hyperglycemia alters perception of rectal distention and blunts the rectoanal inhibitory reflex in healthy volunteers. Gastroenterology 108:1700, 1995. Chey, W. Y., Jin, H. O., Lee, M. H., et al.: Colonic motility abnormality in patients with irritable bowel syndrome exhibiting abdominal pain and diarrhea. Am. J. Gastroenterol. 96:1499, 2001. Christofi, F. L., Palmer, J. M., and Wood, J. D.: Neuropharmacology of the muscarinic antagonist telenzepine in myenteric ganglia of the guinea-pig small intestine. Eur. J. Pharmacol. 195:333, 1991. Christofi, F. L., Tack, J., and Wood, J. D.: Suppression of nicotinic synaptic transmission by adenosine in myenteric ganglia of the guinea-pig gastric antrum. Eur. J. Pharmacol. 216:17, 1992. Christofi, F. L., and Wood, J. D.: Presynaptic inhibition by adenosine A1 receptors on guinea-pig small intestinal myenteric neurons. Gastroenterology 104:1420, 1993. Clerc, N., Furness, J. B., Kunze, W. A., et al.: Long-term effects of synaptic activation at low frequency on excitability of myenteric AH neurons. Neuroscience 90:279, 2000. Coelho, A. M., Fioramonti, J., and Bueno, L.: Mast cell degranulation induces delayed rectal allodynia in rats: role of histamine and 5-HT. Dig. Dis. Sci. 43:727, 1998. Collins, S. M.: Is the irritable gut an inflamed gut? Scand. J. Gastroenterol. 27(Suppl. 192):102, 1992. Cooke, H. J.: Complexities of nervous control of the intestinal epithelium. Gastroenterology 94:1087, 1988.

Neurobiology of the Enteric Nervous System 28. Daniel, E. E., Parrish, M. B., Watson, E. G., et al.: The tachykinin receptors inducing contractile responses of canine ileum circular muscle. Am. J. Physiol. 31:G161, 1995. 29. Davison, J. S., and Clarke, G. D.: Mechanical properties and sensitivity to CCK of vagal gastric slowly adapting mechanoreceptors. Am. J. Physiol. 255:G55, 1988. 30. De Giorgio, R., Bassotti, G., Stanghellini, V., et al.: Clinical, morpho-functional and immunological features of idiopathic myenteric ganglionitis. Gastroenterology 110:A655, 1996. 31. Del Tacca, M., Soldani, G., Selli, M., and Crema, A.: Action of catecholamines on release of acetylcholine from human taenia coli. Eur. J. Pharmacol. 9:80, 1970. 32. Derkach, V., Surprenant, A., and North, R. A.: 5-HT3 receptors are membrane ion channels. Nature 339:706, 1989. 33. Frieling, T., Cooke, H. J., and Wood, J. D.: Serotonin receptors on submucous neurons in the guinea-pig colon. Am. J. Physiol. 261:G1017, 1991. 34. Frieling, T., Cooke, H. J., and Wood, J. D.: Neuroimmune communication in the submucous plexus of guinea-pig colon after sensitization to milk antigen. Am. J. Physiol. 267:G1087, 1994. 35. Frieling, T., Palmer, J. M., Cooke, H. J., and Wood, J. D.: Neuroimmune communication in the submucous plexus of guinea-pig colon after infection with Trichinella spiralis. Gastroenterology 107:1602, 1994. 36. Furness, J. B., Kunze, W. A. A., Bertrand, P. P., et al.: Intrinsic primary afferent neurons of the intestine. Prog. Neurobiol. 54:1, 1998. 37. Galligan, J. J., LePard, K. J., Schneider, D. A., and Zhou, X. P.: Multiple mechanisms of fast excitatory synaptic transmission in the enteric nervous system. J. Auton. Nerv. Syst. 81:97, 2000. 38. Galligan, J. J., Surprenant, A., Tonini, M., and North, R. A.: Differential localization of 5-HT1 receptors on myenteric and submucosal neurons. Am. J. Physiol. 255:G603, 1988. 39. Gershon, M. D.: Review article: Roles played by 5-hydroxytryptamine in the physiology of the bowel. Aliment. Pharmacol. Ther. 13:15, 1998. 40. Gershon, M. D.: The Second Brain. New York, HarperCollins, 1998. 41. Gershon, M. D., and Erde, S. M.: The nervous system of the gut. Gastroenterology 80:1571, 1981. 42. Grafe, P., Mayer, C. J., and Wood, J. D.: Synaptic modulation of calcium-dependent potassium conductance in myenteric neurons. J. Physiol. (Lond.) 305:235, 1980. 43. Grider, J. R., Cable, M. B., and Said, S. I.: Vasoactive intestinal peptide as a neural mediator of gastric relaxation. Am. J. Physiol. 248:G73, 1985. 44. Grider, J. R., and Makhlouf, G. M.: Colonic peristaltic reflex: identification of vasoactive intestinal peptide as mediator of descending relaxation. Am. J. Physiol. 251:G40, 1986. 45. Gwee, K. A., Leong, Y. L., Graham, C., et al.: The role of psychological and biological factors in post-infective gut function. Gut 44:400, 1999. 46. Heidenreich, F., Schober, R., Brinck, U., and Hartung, H. P.: Multiple paraneoplastic syndromes in a patient with antibodies to neuronal nucleoproteins (anti-Hu). J. Neurol. 242:210, 1995.

273

47. Hellstrom, P. M., Murthy, K. S., Grider, J. R., and Makhlouf, G. M.: Coexistence of three tachykinin receptors coupled to Ca⫹⫹ signaling pathways in intestinal muscle cells. J. Pharmacol. Exp. Ther. 270:236, 1994. 48. Hirst, G. D. S., Holman, M. E., and Spence, I.: Two types of neurones in the myenteric plexus of duodenum in the guinea-pig. J. Physiol. (Lond.) 236:303, 1974. 49. Hirst, G. D. S., Johnson, S. M., and van Helden, D. F.: The calcium current in a myenteric neurone of the guinea pig ileum. J. Physiol. (Lond.) 361:297, 1985. 50. Hirst, G. D. S., Johnson, S. M., and van Helden, D. F.: The slow calcium-dependent current in a myenteric neurone of the guinea pig ileum. J. Physiol. (Lond.) 361:315, 1985. 51. Hirst, G. D. S., and McKirdy, H. C.: Presynaptic inhibition at a mammalian peripheral synapse. Nature (Lond.) 250:430, 1974. 52. Hirst, G. D. S., and McKirdy, H. C.: Synaptic potentials recorded from neurones of the submucous plexus of guinea-pig small intestine. J. Physiol. (Lond.) 249:369, 1975. 53. Hirst, G. D. S., and Silinsky, E. M.: Some effects of 5-hydroxytryptamine, dopamine and noradrenaline on neurones in the submucous plexus of guinea-pig small intestine. J. Physiol. (Lond.) 251:817, 1975. 54. Hirst, G. D. S., and Spence, I.: Calcium action potentials in mammalian peripheral neurones. Nature (Lond.) 243:54, 1973. 55. Hogan, W. J., Greenen, J. E., and Dodds, W. J.: Paradoxical motor response to cholecystokinin-octapeptide (CCK-OP) in patients with suspected sphincter-of-Oddi dysfunction. Gastroenterology 82:A1085, 1982. 56. Hu, H.-Z., Gao, N., Gao, C., et al.: Calmodulin (CaM) and CaM kinase signaling in the enteric nervous system. Neurosci. Abstr. 27:839.11, 2001. 57. Hu, H.-Z., Gao, N., Lin, Z., et al.: P2X7 receptors in the enteric nervous system of guinea-pig small intestine. J. Comp. Neurol. 440:299, 2001. 58. Hu, H.-Z., Gao, N., Ren, J., et al.: A novel metabotropic P2Y1 receptor in submucous plexus of guinea-pig small intestine. Gastroenterology 120:A509, 2001 59. Hukuhara, T., Kotania, S., and Sato, G.: Effect of destruction of intramural ganglion cells on colon motility: possible genesis of congenital megacolon. J. Physiol. (Japan) 11:635, 1961. 60. Iyer, V., Bornstein, J. C., Costa, M., et al.: Electrophysiology of guinea-pig myenteric neurons correlated with immunoreactivity for calcium binding proteins. J. Auton. Nerv. Syst. 22:141, 1988. 61. Javed, N. H., and Cooke, H. J.: Acetylcholine release from colonic submucous neurons in the guinea pig associated with Cl⫺ secretion. Am. J. Physiol. 262:G1331, 1992. 62. Jean, W., Dalmau, J., Ho, A., and Posner, J. B.: Analysis of the IgG subclass distribution and inflammatory infiltrates in patients with anti-hu-associated paraneoplastic encephalomyelitis. Neurology 44:144, 1994. 63. Jenkinson, K. M., Mann, P. T., Southwell, B. R., and Furness, J. B.: Independent endocytosis of the NK(1) and NK(3) tachykinin receptors in neurons of rat myenteric plexus. Neuroscience 100:191, 2000. 64. Jenkinson, K. M., Morgan, J. M., Furness, J. B., and Southwell, B. R.: Neurons bearing NK(3) tachykinin

274

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75. 76.

77.

78.

79.

80.

81.

Function of the Peripheral Nervous System receptors in the guinea-pig ileum revealed by specific binding of fluorescently labelled agonists. Histochem. Cell Biol. 112:233, 1999. Jin, J. G., Murthy, K. S., Grider, J. R., and Makhlouf, G. M.: Stoichiometry of neurally induced VIP release, NO formation, and relaxation in rabbit and rat gastric muscle. Am. J. Physiol. 34:G357, 1996. Johnson, S. M., Katayama, Y., and North, R. A.: Multiple actions of 5-hydroxytryptamine on myenteric neurones of the guinea-pig ileum. J. Physiol. (Lond.) 304:459, 1980. Johnson, S. M., and North, R. A.: Slow synaptic potentials in neurones of the myenteric plexus. J. Physiol. (Lond.) 301:505, 1980. Jonakait, G. M., Tamir, H., Gintzler, A. R., and Gershon, M. D.: Release of [3H] serotonin and its binding protein from enteric neurons. Brain Res. 174:55, 1979. Katayama, Y., North, R. A., and Williams, J. T.: The action of substance P on neurons of the myenteric plexus of the guinea-pig small intestine. Proc. R. Soc. Lond. 206:191, 1979. Kirchgessner, A. L., Tamir, H., and Gershon, M. D.: Identification and stimulation by serotonin of intrinsic sensory neurons of the submucosal plexus of the guinea pig gut—activity-induced expression of Fos immunoreactivity. J. Neurosci. 12:235, 1992. Kirkup, A. J., Brunsden, A. M., and Grundy, D.: Receptors and transmission in the brain-gut axis: I. Receptors on visceral afferents. Am. J. Physiol. 280:G787, 2001. Knoll, J., and Vizi, E. S.: Presynaptic inhibition of acetylcholine release by endogenous and exogenous noradrenaline at high rate of stimulation. Br. J. Pharmacol. 40:400, 1970. Krishnamurthy, S., and Schuffler, M. D.: Pathology of neuromuscular disorders of the small intestine and colon. Gastroenterology 93:610, 1997. Lang, I. M., and Sarna, S. K.: Motor and myoelectric activity associated with vomiting, regurgitation, and nausea. In Wood, J. D. (ed.): Handbook of Physiology: The Gastrointestinal System, Motility and Circulation. New York, Oxford University Press, p. 1179, 1989. Langley, J. N.: The Autonomic Nervous System, Part I. Cambridge, UK, W. Heffer and Sons, 1921. Langley, J. N., and Magnus, R.: Some observations of movements of the intestine before and after degenerative section of the mesenteric nerves. J. Physiol. (Lond.) 33:34, 1905. Lembo, T., Munakata, J., Mertz, H., et al.: Evidence for the hypersensitivity of lumbar splanchnic afferents in irritable bowel syndrome. Gastroenterology 107:1686, 1994. Lennon, V. A., San, D. F., Busk, M. F., et al.: Enteric neuronal autoantibodies in pseudoobstruction with small-cell lung carcinoma. Gastroenterology 100:137, 1991. Liu, S., Hu, H.-Z., Ren, J., et al.: Pre- and postsynaptic inhibition by nociceptin in guinea pig myenteric plexus in vitro. Am. J. Physiol. 281:G237, 2001. Liu, S., Xia, Y., Hu, H.-Z., et al.: Histamine H3 receptormediated suppression of inhibitory synaptic transmission in the guinea-pig submucous plexus. Eur. J. Pharmacol. 397:49, 2000. Loening-Baucke, V.: Sensitivity of the sigmoid colon and rectum in children treated for chronic constipation. J. Pediatr. Gastroenterol. Nutr. 3:454, 1984.

82. Manaka, H., Manaka ,Y., Kostolanska, F., et al.: Release of VIP and substance-P from isolated perfused canine ileum. Am. J. Physiol. 257:G182, 1989. 83. Manber, L., and Gershon, M. D.: A reciprocal adrenergiccholinergic axoaxonic synapse in the mammalian gut. Am. J. Physiol. 5:E738, 1979. 84. Mao, Y. K., Wang, Y. F., and Daniel, E. E.: Characterization of neurokinin type 1 receptor in canine small intestinal muscle. Peptides 17:839, 1996. 85. Mawe, G. M.: Intracellular recording from gall-bladder neurones of the guinea-pig gall bladder. J. Physiol. (Lond.) 429:323, 1990. 86. Mawe, G. M.: The role of cholecystokinin in ganglionic transmission in the guinea-pig gall-bladder. J. Physiol. (Lond.) 439:89, 1991. 87. Mawe, G. M., Branchek, T. A., and Gershon, M. D.: Peripheral neural serotonin receptors: identification and characterization with specific antagonists and agonists. Proc. Natl. Acad. Sci. U. S. A. 83:9799, 1986. 88. Mayer, E., Schuffler, M., Rotter, J., et al.: Familial visceral neuropathy with autosomal dominant transmission. Gastroenterology 91:1528, 1986. 89. Mertz, H., Morgan, V., Tanner, G., et al.: Regional cerebral activation in irritable bowel syndrome and control subjects with painful and nonpainful rectal distention. Gastroenterology 118:842, 2000. 90. Mihara, S., Hirai, K., Katayama, Y., and Nishi, S.: Mechanisms underlying intracellular signal transduction of the slow IPSP in submucous neurones of the guinea-pig caecum. J. Physiol. (Lond.) 436:621, 1991. 91. Mihara, S., and North, R. A.: Opioids increase potassium conductance in submucous neurones of guinea-pig caecum by activating delta receptors. Br. J. Pharmacol. 88:315, 1986. 92. Monoz-Garcia, D., and Ludwin, S.: Adult-onset neuronal intranuclear hyaline inclusion disease. Neurology 36:785, 1986. 93. Morita, K., and Katayama, Y.: Substance P inhibits activation of calcium-dependent potassium conductances in guinea-pig myenteric neurones. J. Physiol. (Lond.) 447:203, 1992. 94. Morita, K., and North, R. A.: Clonidine activates membrane potassium conductance in myenteric neurones. Br. J. Pharmacol. 74:419, 1981. 95. Morita, K., and North, R. A.: Opiate activation of potassium conductance in myenteric neurons: inhibition by calcium ion. Brain Res. 242:145, 1982. 96. Morita, K., North, R. A., and Tokimasa, T.: The calciumactivated potassium conductance in guinea-pig myenteric neurones. J. Physiol. (Lond.) 329:341, 1982. 97. Murthy, K. S., Grider, J. R., Jin, J. G., and Makhlouf, G. M.: Interplay of VIP and nitric oxide in the regulation of neuromuscular activity in the gut. Arch. Int. Pharmacodyn. Ther. 329:27, 1995. 98. Murthy, K. S., and Makhlouf, G. M.: Vasoactive intestinal peptide pituitary adenylate cyclase-activating peptidedependent activation of membrane-bound NO synthase in smooth muscle mediated by pertussis toxin-sensitive G(i1–2). J. Biol. Chem. 269:15977, 1994. 99. Nemeth, P. R., Ort, C. A., and Wood, J. D.: Intracellular study of effects of histamine on electrical behavior of

Neurobiology of the Enteric Nervous System

100.

101.

102.

103.

104.

105.

106. 107.

108.

109.

110.

111.

112.

113.

114.

115.

116.

myenteric neurons in guinea-pig small intestine. J. Physiol. (Lond.) 355:411, 1984. Nemeth, P. R., Palmer, J. M., Wood, J. D., and Zafirov, D. H.: Effects of forskolin on electrical behavior of myenteric neurones in guinea-pig small intestine. J. Physiol. (Lond.) 376:439, 1986. Nemeth, P. R., Zafirov, D. H., and Wood, J. D.: Effects of cholecystokinin, caerulein and pentagastrin on electrical behavior of myenteric neurons. Eur. J. Pharmacol. 116:263, 1985. Nishi, S., and North, R. A.: Intracellular recording from the myenteric plexus of the guinea-pig ileum. J. Physiol. (Lond.) 231:471, 1973. North, R. A.: The calcium-dependent slow after-hyperpolarization in myenteric plexus neurons with tetrodotoxinresistant action potentials. Br. J. Pharmacol. 49:709, 1973. North, R. A., Henderson, G., Katayama, Y., and Johnson, S.M.: Electrophysiological evidence for presynaptic inhibition of acetylcholine release by 5-hydroxytryptamine in the enteric nervous system. Neuroscience 5:581, 1980. North, R. A., and Karras, P. J.: Opiate tolerance and dependence induced in vitro in single myenteric neurones. Nature 272:73, 1978. North, R. A., and Sieglgansberger, W.: Opiate withdrawal signs in single myenteric neurones. Brain Res. 144:208, 1978. North, R. A., Slack, B. E., and Surprenant, A.: Muscarinic M1 and M2 receptors mediate depolarization and presynaptic inhibition in guinea-pig enteric nervous system. J. Physiol. (Lond.) 368:435, 1985. North, R. A., and Surprenant, A.: Inhibitory synaptic potentials resulting from alpha2 adrenoceptor activation in guinea-pig submucous plexus neurones. J. Physiol. (Lond.) 358:17, 1985. North, R. A., and Tokimasa, T.: Depression of calciumdependent potassium conductance of guinea-pig myenteric neurones by muscarinic agonists. J. Physiol. (Lond.) 342:253, 1983. North, R. A., and Tokimasa, T.: Persistent calciumsensitive potassium current and the resting properties of guinea-pig myenteric neurones. J. Physiol. (Lond.) 386:333, 1987. North, R. A., and Tonini, M.: The mechanism of action of narcotic analgesics in the guinea-pig ileum. Br. J. Pharmacol. 61:541, 1977. O’Sullivan, M., Clayton, N., Breslin, N. P., et al.: Increased mast cells in the irritable bowel syndrome. Neurogastroenterol. Motil. 12:449, 2000. Palmer, J. M., Schemann, M., Tamura, K., and Wood, J. D.: Galanin mimics slow synaptic inhibition in myenteric neurons. Eur. J. Pharmacol. 124:379, 1986. Palmer, J. M., Wood, J. D., and Zafirov, D. H.: Elevation of cyclic adenosine monophosphate mimics slow synaptic excitation in myenteric neurones of the guinea-pig. J. Physiol. (Lond.) 376:451, 1986. Palmer, J. M., Wood, J. D., and Zafirov, D. H.: Purinergic inhibition in the small intestinal myenteric plexus of the guinea-pig. J. Physiol. (Lond.) 387:357, 1987. Palo, J., Haltia, M., Carpenter, S., et al.: Neurofilament subunit-related proteins in neuronal intranuclear inclusion. Ann. Neurol. 15:316, 1984.

275

117. Paton, W. D. M., and Vizi, E. S.: The inhibitory action of noradrenaline and adrenaline on acetylcholine output by guinea-pig ileum longitudinal muscle strip. Br. J. Pharmacol. 35:10, 1969. 118. Read, N. W.: Rectal distention: from sensation to feeling. Gastroenterology 118:972, 2000. 119. Reiche, D., and Schemann, M.: Ascending choline acetyltransferase and descending nitric oxide synthase immunoreactive neurones of the myenteric plexus project to the mucosa of the guinea pig gastric corpus. Neurosci. Lett. 241:61, 1998. 120. Richter, J. E., Barish, C. F., and Castell, D. O.: Abnormal sensory perception in patients with esophageal chest pain. Gastroenterology 91:845, 1986. 121. Rolny, P., Arleback, A., Funch-Jensen, P., et al.: Paradoxical response of sphincter of Oddi to intravenous injection of cholecystokinin or ceruletide: manometric findings and results of treatment in biliary dyskinesia. Gut 27:1507, 1986. 122. Rosenblum, M.: Paraneoplastic and autoimmunologic injury of the nervous system: the anti-hu syndrome. Brain Pathol. 3:199, 1993. 123. Roy, A., Bharucha, H., Nevin, N., and Odling-Smee, G. W.: Idiopathic intestinal pseudo-obstruction: a familial visceral neuropathy. Clin. Genet. 18:292, 1980. 124. Sanders, K. M., Ordog, T., Koh, S. D., and Ward, S. M.: A novel pacemaker mechanism drives gastrointestinal rhythmicity. News Physiol. Sci. 15:291, 2000. 125. Schemann, M., and Grundy, D.: Electrophysiological identification of vagally innervated enteric neurons in the guinea pig stomach. Am. J. Physiol. 263:G709, 1992. 126. Schemann, M., and Kayser, H.: Effects of tachykinins on myenteric neurones of the guinea-pig gastric corpus: involvement of NK-3 receptors. Pflugers Arch. 419:566, 1991. 127. Schemann, M., and Schaaf, C.: Differential projection of cholinergic and nitroxidergic neurons in the myenteric plexus of guinea pig stomach. Am. J. Physiol. 32:G186, 1995. 128. Schemann, M., and Tamura, K.: Presynaptic inhibitory effects of the peptides NPY, PYY, and PP on nicotinic EPSPs in guinea-pig gastric myenteric neurones. J. Physiol. (Lond.) 451:79, 1992. 129. Schemann, M., and Wood, J. D.: Synaptic behavior of myenteric neurones in the gastric corpus of the guinea-pig. J. Physiol. (Lond.) 417:519, 1989. 130. Schiller, W., Suriyapa, C., Mutchler, J., and Anderson, M. C.: Surgical alteration of intestinal motility. Am. J. Surg. 125:122, 1973. 131. Schuffler, M., Baird, H., and Fleming, C.: Intestinal pseudo-obstruction as the presenting manifestation of small cell carcinoma of the lung: a paraneoplastic neuropathy of the gastrointestinal tract. Ann. Intern. Med. 98:129, 1983. 132. Schutte, I. W., Akkermans, L. M., and Kroese, A. B.: CCKA and CCKB receptor subtypes both mediate the effects of CCK-8 on myenteric neurons in the guinea-pig ileum. J. Auton. Nerv. Syst. 67:51, 1997. 133. Schutte, I. W., Hollestein, K. B., Akkermans, L. M., and Kroese, A. B.: Evidence for a role of cholecystokinin as

276

134.

135.

136.

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

148.

149.

Function of the Peripheral Nervous System neurotransmitter in the guinea-pig enteric nervous system. Neurosci. Lett. 236:155, 1997. Shen, K.-Z., North, R. A., and Surprenant, A.: Potassium channels opened by noradrenaline and other transmitters in excised membrane patches of guinea-pig submucosal neurones. J. Physiol. (Lond.) 445:581, 1992. Silverman, D. H. S., Munakata, J. A., Ennes, H., et al.: Regional cerebral activity in normal and pathological perception of visceral pain. Gastroenterology 112:64, 1997. Smith, G. P., Jerome, C., and Norgren, R.: Afferent axons in abdominal vagus mediate satiety effect of cholecystokinin in rats. Am. J. Physiol. 249:R638, 1985. Smith, V., Gregson, N., and Foggensteiner, L.: Acquired intestinal aganglionosis and circulating autoantibodies without neoplasia or other neural involvement. Gastroenterology 112:1366, 1997. Song, Z. M., Brookes, S. J. H., and Costa, M.: Identification of myenteric neurons which project to the mucosa of the guinea-pig small intestine. Neurosci Lett. 129:294, 1991. Song, Z. M., Brookes, S. J. H., and Costa, M.: All calbindinimmunoreactive myenteric neurons project to the mucosa of the guinea-pig small intestine. Neurosci. Lett. 180:219, 1994. Stanghellini, V., Camilleri, M., and Malagelada, J.-R.: Chronic idiopathic intestinal pseudo-obstruction: clinical and intestinal manometric findings. Gut 23:824, 1987. Starodub, A. M., and Wood, J. D.: A-type potassium current in myenteric neurons of guinea-pig small intestine. Neuroscience 99:389, 2000. Surprenant, A.: Synaptic transmission in neurons of the submucous plexus. In Nerves and the Gastrointestinal Tract (Falk Symposium 50). Boston, MTP Press, p. 253, 1989. Surprenant, A., and Crist, J.: Electrophysiological characterization of functionally distinct 5-hydroxytryptamine receptors on guinea-pig submucous plexus. Neuroscience 24:283, 1988. Surprenant, A., and North, R. A.: Mechanism of synaptic inhibition by noradrenaline acting at alpha 2-adrenoceptors. Proc. R. Soc. Lond. 234:85, 1988. Surprenant, A., and North, R. A.: Inhibitory receptors and signal transduction in submucosal neurones. In Holle, G. E., and Wood, J. D. (eds.): Advances in the Innervation of the Gastrointestinal Tract. Amsterdam, Elsevier Scientific Press, p. 239, 1992. Tack, J., Caenepeel, P., Fischler, B., et al.: Symptoms associated with hypersensitivity to gastric distention in functional dyspepsia. Gastroenterology 121:526, 2001. Tack, J., Janssens, J., Vantrappen, G., and Wood, J. D.: Actions of 5-hydroxytryptamine on myenteric neurons in the guinea-pig gastric antrum. Am. J. Physiol. 263:G838, 1992. Tack, J., Janssens, J., Vantrappen, G., et al.: Effect of erythromycin on gastric motility in controls and diabetic gastroparesis. Gastroenterology 103:72, 1992. Tack, J. D., and Wood, J. D.: Synaptic behavior in the myenteric plexus of the guinea-pig gastric antrum. J. Physiol. (Lond.) 445:389, 1992.

150. Takaki, M., and Nakayama, S.: Effects of mesenteric nerve stimulation on the electrical activity of myenteric neurons in the guinea pig ileum. Brain Res. 442:351, 1988. 151. Talley, N. J.: Review article: 5-Hydroxytryptamine agonists and antagonists in the modulation of gastrointestinal motility and sensation—clinical implications. Aliment. Pharmacol. Ther. 6:273, 1992. 152. Tamura, K., Palmer, J. M., and Wood, J. D.: Presynaptic inhibition produced by histamine at nicotinic synapses in enteric ganglia. Neuroscience 25:171, 1987. 153. Tamura, K., and Wood, J. D.: Electrical and synaptic properties of myenteric plexus neurones in the terminal large intestine of the guinea-pig. J. Physiol. (Lond.) 415:275, 1989. 154. Tatsumi, H., Hirai, K., and Katayama, Y.: Measurement of the intracellular calcium concentration in guinea-pig myenteric neurons by using fura-2. Brain Res. 451:371, 1988. 155. Thomas, E. A., Bertrand, P. P., and Bornstein, J. C.: Genesis and role of coordinated firing in a feed forward network: a model study of the enteric nervous system. Neuroscience 93:1525, 1999. 156. Thompson, W. G., Longstreth, G. L., Drossman, D. A., et al.: Functional bowel disorders and functional abdominal pain. In Drossman, D. A., Talley, N. J., Thompson, W. G., et al. (eds.): The Functional Gastrointestinal Disorders. McLean, VA, Degnon Associates, p. 351, 2000. 157. Timmermans, J.-P., Adriaensen, D., Cornelissen, W., and Scheuermann, D. W.: Structural organization and neuropeptide distribution in the mammalian enteric nervous system, with special attention to those components involved in mucosal reflexes. Comp Biochem. Physiol. 118:331, 1997. 158. Vantrappen, G., Peeters, T. L., Bloom, S. R., et al.: Motilin and the interdigestive migrating motor complex in man. Am. J. Dig. Dis. 24:497, 1979. 159. Vergara, P., Woskowska, Z., Cipris, S., et al.: Somatostatin excites canine ileum ex vivo: role for nitric oxide? Am. J. Physiol. 269:G12, 1995. 160. Verne, G., Sallustio, J., and Eaker, E.: Anti-myenteric neuronal antibodies in patients with achalasia: a prospective study. Dig. Dis. Sci. 42:307, 1997. 161. Vizi, V. A., and Vizi, E. S.: Direct evidence for acetylcholine releasing effect of serotonin in the Auerbach’s plexus. J. Neural Transm. 42:127, 1978. 162. Whitehead, W. E., Holtkotter, B., Enck, P., et al.: Tolerance for rectosigmoid distention in irritable bowel syndrome. Gastroenterology 98:1187, 1990. 163. Wood, J., Hudson, L., Jessel, T., and Yamamoto, M.: A monoclonal antibody defining antigenic determinants on subpopulations of mammalian neurones and Trypanosoma cruzi parasites. Nature 296:34, 1982. 164. Wood, J. D.: Excitation of intestinal muscle by atropine, tetrodotoxin and Xylocaine. Am. J. Physiol. 222:118, 1972. 165. Wood, J. D.: Electrical activity of the intestine of mice with hereditary megacolon and absence of myenteric ganglion cells. Am. J. Dig. Dis. 18:477, 1973. 166. Wood, J. D.: Neuronal interactions within ganglia of Auerbach’s plexus of the small intestine. In Bülbring, E., and Shuba, M. F. (eds.): Physiology of Smooth Muscle. New York, Raven Press, p. 321, 1976.

Neurobiology of the Enteric Nervous System 167. Wood, J. D.: Electrical and synaptic behavior of enteric neurons. In Wood, J. D. (ed.): Handbook of Physiology: The Gastrointestinal System, Motility and Circulation. New York, Oxford University Press, p. 465, 1989. 168. Wood, J. D.: Histamine signals in enteric neuroimmune interactions. Ann. N. Y. Acad. Sci. 664:275, 1992. 169. Wood, J. D.: Application of classification schemes to the enteric nervous system. J. Auton. Nerv. Syst. 48:179, 1994. 170. Wood, J. D.: Physiology of the enteric nervous system. In Johnson, L. R., Alpers, D. H., Christensen, J., et al. (eds.): Physiology of the Gastrointestinal Tract, 3rd ed. New York, Raven Press, p. 423, 1994. 171. Wood, J. D.: Physiological and pathophysiological paracrine functions of intestinal mast cells in enteric neuro-immune signaling. In Singer, M. F., Zigler, R., and Rohr, G. (eds.): Gastrointestinal Tract and Endocrine System. London, Kluwer Academic Publishers, p. 254, 1995. 172. Wood, J. D.: Neurotransmission at the interface of sympathetic and enteric divisions of the autonomic nervous system. Chin. J. Physiol. 42:201, 1999. 173. Wood, J. D.: Allergies and the brain-in-the-gut. Clin. Perspect. Gastroenterol. 3:343, 2000. 174. Wood, J. D.: Neuropathy in the brain-in-the-gut. Eur. J. Gastroenterol. Hepatol. 12:597, 2000. 175. Wood, J. D.: Neuropathophysiology of IBS. J. Clin. Gastroenterol. 35(Suppl.):11, 2002. 176. Wood, J. D., and Mayer, C. J.: Intracellular study of electrical activity of Auerbach’s plexus in guinea-pig small intestine. Pflugers Arch. 374:265, 1978. 177. Wood, J. D., and Mayer, C. J.: Adrenergic inhibition of serotonin release from neurons in guinea-pig Auerbach’s plexus. J. Neurophysiol. 42:594, 1979.

277

178. Wood, J. D., and Tack, J. F.: Motilin excites myenteric neurons in the gastric antrum of the guinea-pig. Gastroenterology 99:A1236, 1990. 179. Xia, Y., Baidan, L. V., Fertel, R. H., and Wood, J. D.: Determination of levels of cyclic AMP in the myenteric plexus of guinea-pig small intestine. Eur. J. Pharmacol. 206:231, 1991. 180. Xia, Y., Fertel, R., and Wood, J.: Stimulation of formation of adenosine 3⬘,5⬘- phosphate by histamine in myenteric ganglia isolated from guinea-pig small intestine. Eur. J. Pharmacol. 316:81, 1996. 181. Xia, Y., Fertel, R. H., and Wood, J. D.: Stimulation of formation of cAMP by 5-hydroxytryptamine in myenteric ganglia isolated from guinea pig small intestine. Life Sci. 55:685, 1994. 182. Xia, Y., Fertel, R. H., and Wood, J. D.: Suppression of cAMP formation by adenosine in myenteric ganglia of guinea-pig small intestine. Eur. J. Pharmacol. 32:95, 1997. 183. Yau, W. M., and Youther, M. L.: Direct evidence for a release of acetylcholine from myenteric plexus of guinea pig small intestine by substance P. Eur. J. Pharmacol. 81:665, 1982. 184. Zafirov, D. H., Cooke, H. J., and Wood, J. D.: Elevation of cAMP facilitates noradrenergic transmission in submucous neurons of guinea-pig ileum. Am. J. Physiol. 262:G442, 1993. 185. Zholos, A. V., Baidan, L. V., Starodub, A. M., and Wood, J. D.: Potassium channels of myenteric neurons in guinea-pig small intestine. Neuroscience 89:603, 1999. 186. Zholos, A. V., Baidan, L. V., and Wood, J. D.: Sodium conductance in cultured myenteric AH-type neurons from guinea-pig small intestine. Auton. Neurosci. 96:93, 2002.

13 Autonomic and Somatic Systems to the Anorectum and Pelvic Floor ADIL E. BHARUCHA AND CHRISTOPHER J. KLINGELE

Embryology Anatomy Pelvic Floor Rectum and Anal Canal Nerve Supply to the Pelvic Floor Sympathetic Nerves Parasympathetic Nerves Central Nervous System Pathways Regulating the Pelvic Floor Afferent Innervation

Anal Sphincter Tone and Reflexes Internal Anal Sphincter External Anal Sphincter Puborectalis Sacral Reflexes Pharmacology Mechanisms of Continence and Defecation Anorectal Dysfunction in Neurologic Disorders

The anorectum is the caudal end of the gastrointestinal tract, responsible for fecal continence and defecation. In humans, defecation is a viscerosomatic reflex that is often preceded by several attempts to preserve continence. Walter C. Bornemeier captured the intricacies of anorectal function in the following words: There is not a muscle or structure in the body that has a more keenly developed sense of alertness and ability to accommodate itself to varying situations . . . . Yet the sphincter ani can do it! The sphincter can apparently differentiate between solid, fluid and gas. It apparently can tell whether its owner is alone or with someone; whether standing up or sitting down, whether the owner is dressed or undressed. No other muscle in the body is such a protector of the dignity of man, yet so ready to come to his relief. A muscle like this is worth protecting. The anal canal is normally closed by an internal sphincter made of smooth muscle and an external sphincter made of tonically active striated muscle. The internal and external anal sphincters are supplied by autonomic and somatic nerves, respectively. Rectal smooth muscle is primarily regulated by the intrinsic or enteric nervous system while

Obstructed Defecation Fecal Incontinence Management of Fecal Incontinence Bowel Habit Modification Pharmacologic Approaches to Increase Sphincter Pressure Biofeedback Therapy Surgical Approaches

extrinsic (i.e., autonomic) nerves modulate the enteric nervous system. Within the gastrointestinal tract, extrinsic nerves have the most pronounced effects on portions lined by skeletal muscle (i.e., the upper esophageal sphincter and the external anal sphincter). Extrinsic nerves are critical for anorectal functions of storing and evacuating feces; these functions require coordination between visceral components (e.g., rectal contraction, rectal sensation, and internal sphincter relaxation) and somatic components (e.g., external sphincter contraction and relaxation). Consequently, extrinsic neural dysfunction often causes or contributes to defecatory disorders, including obstructed defecation and fecal incontinence.

EMBRYOLOGY The anorectal region in humans derives from four separate embryologic structures: the hindgut, cloaca, proctodeal pit, and anal tubercles.31 The rectum above the pubococcygeal line develops from the primitive hindgut, while the rectum below this line develops from the cloaca. The cloaca is initially a single tube that is subsequently separated by caudad migration of the urorectal septum into urogenital 279

280

Function of the Peripheral Nervous System

ANATOMY

uterus, and rectum. These defects are closed by connective tissue anterior to the urethra, anterior to the rectum (i.e., the perineal body), and posterior to the rectum (i.e., the postanal plate). The levator ani is subdivided into four muscles: pubococcygeus, ileococcygeus, coccygeus, and puborectalis. These muscles are attached peripherally to the pubic body, the ischial spine, and the arcus tendinus, a condensation of the obturator fascia in between these areas (Fig. 13–1). It is unclear whether the puborectalis should be regarded as a component of the levator ani complex or the external anal sphincter. Based on developmental evidence, innervation, and histologic studies, the puborectalis appears distinct from the majority of the levator ani.31 However, the puborectalis and external sphincter complex are innervated by separate nerves originating from S2-S4 (see below), suggesting phylogenetic differences between these two muscles.122

Pelvic Floor

Rectum and Anal Canal

The levator ani, or pelvic diaphragm, is a dome-shaped muscular sheet68 that predominantly contains striated muscle and has midline defects enclosing the bladder,

The rectum is 15 to 20 cm long, and extends from the rectosigmoid junction at the level of third sacral vertebra to the anal orifice (Fig. 13–2). The upper and lower rectum are

and intestinal passages. Mesenchymal elements from the lateral wall of the terminal cloaca fold inward, fusing with the urorectal septum above and the cloacal membrane below. Thereafter the cloacal membrane disintegrates, forming the upper anal canal, which is lined by ectoderm (i.e., columnar epithelium) and endoderm (i.e., squamous epithelium) in the upper and lower portions, respectively. The transition from ectoderm to exoderm may be at, or distal to, the pectinate line. The proctodeal portion of the cloacal membrane disintegrates to form the anal tubercles that join posteriorly and migrate ventrally to encircle a depression, the proctodeal pit. The anal tubercles join the urorectal septum and genital tubercles to form the perineal body, completing the separation between the rectum and urogenital tract.

Symphysis pubis

Deep dorsal vein of clitoris Puborectalis

Urethra Pubococcygeus

Vagina Illiococcygeus Rectum Coccygeus

Piriformis

Sacrum

FIGURE 13–1 Pelvic view of the levator ani demonstrating its four main components: puborectalis, pubococcygeus, iliococcygeus, and coccygeus.

281

Autonomic and Somatic Systems to the Anorectum and Pelvic Floor

Pelvi- Longitudinal Circular muscle Obturator rectal muscle layer space space internus

Para-vertebral sympathectic chain

L1 L2

IllioLevator cocygeus ani Pubococygeus

Transverse folds Rectal ampulla Anal columns

Ischiorectal fossa Gluteus maximus Semitendinosus

L3

L4

L5 Deep Superficial Skin Subcutaneous Sphincter Parts of ani internus sphincter ani externus

Anal sinuses

FIGURE 13–2 Diagram of a coronal section of the rectum, anal canal, and adjacent structures. The pelvic barrier includes the anal sphincters and pelvic floor muscles.

separated by a horizontal fold. The upper rectum is derived from the embryologic hindgut, generally contains feces, and can distend toward the peritoneal cavity. The lower part, derived from the cloaca, is confined by a tube of condensed extraperitoneal connective tissue, and is generally empty in normal subjects, except during defecation. In humans, there are fewer enteric ganglia in the rectum compared to the colon and very few ganglia in the anal sphincter.69,168

NERVE SUPPLY TO THE PELVIC FLOOR Sympathetic Nerves The anorectum and pelvic floor are supplied by sympathetic, parasympathetic, and somatic fibers. Sympathetic preganglionic fibers originate from the lowest thoracic ganglion in the paravertebral sympathetic chain. After traveling in the lowest of four (thoracic) splanchnic nerves, these preganglionic fibers join the third lumbar splanchnic nerve and branches from the aortic plexus to form the superior hypogastric plexus (Fig. 13–3). The superior hypogastric plexus is situated in front of the aortic bifurcation at the L4-S1 level. The alternative term for this plexus (i.e., presacral nerve) is misleading because it is seldom condensed, and does not resemble a single nerve. The superior hypogastric plexus provides branches to the uteric and ovarian (or testicular) plexus, and divides into right and left hypogastric nerves that descend, uniting with parasympathetic fibers in the pelvic splanchnic nerves to form corresponding inferior hypogastric plexuses. Branches from each hypogastric nerve supply the testicular or ovarian plexus, the ureteric

Pelvic splanchnic nerves S1

Aortic plexus

Lumbar splanchnic nerves Superior hypogastric plexus

Left hypogastric nerve Right hypogastric nerve

S2 S3 S4 S5

Nerve to levator ani Pudendal nerve Middle rectal plexus Inferior rectal nerve

Inferior hypogastric plexus Uterovaginal plexus Vesical plexus

Labial branches of Perineal nerve perineal nerve

FIGURE 13–3 Sympathetic, parasympathetic, and pudendal nerve supply to the anorectum.

plexus, and the plexus on the internal iliac artery and to the sigmoid colon. Each hypogastric nerve is also joined by the lowest lumbar splanchnic nerve, that from the last lumbar sympathetic ganglion. The inferior hypogastric plexus is located posterior to the urinary bladder, beside the rectum and prostate in men; in women this plexus is also adjacent to the uterus. The inferior hypogastric plexus contains numerous small ganglia with sympathetic fibers conveyed predominantly by the hypogastric nerves, and also by branches from the lumbar splanchnic nerves. The origin and distribution of parasympathetic fibers in the inferior hypogastric plexus are discussed below. The inferior hypogastric plexus gives rise to the middle rectal plexus, vesical plexus, prostatic plexus, and uterovaginal plexus. The nerve supply to the rectum and anal canal is derived from the superior, middle, and inferior rectal plexus. Parasympathetic fibers in the superior and middle rectal plexuses synapse with postganglionic neurons in the myenteric plexus in the rectal wall. The inferior rectal nerve is a branch of the pudendal nerve, conveying motor

282

Function of the Peripheral Nervous System

fibers to the external anal sphincter and sensory input from the lower anal canal.

Parasympathetic Nerves Preganglionic parasympathetic fibers originating from ventral rami of the second, third, and often fourth sacral nerves travel in the pelvic splanchnic nerves, synapsing with neurons in the inferior hypogastric plexus or more distally, in the walls of viscera supplied by the plexus. In addition, ascending fibers from the inferior hypogastric plexus travel via superior hypogastric and aortic plexuses to reach the inferior mesenteric plexus, ultimately innervating the descending and sigmoid colon. After entering the colon, these fibers form the ascending colonic nerves, traveling cephalad in the plane of the myenteric plexus to supply a variable portion of the left colon. Sacral parasympathetic pathways to the colon have excitatory and inhibitory components.58 Excitatory pathways play an important role in colonic propulsive activity, especially during defecation. Consequently, defecation may be affected after surgical section of pelvic nerves in humans.38,145 In other species (e.g., guinea pig), feces transport may be entirely organized by the enteric nervous system; spinal and supraspinal reflexes are also involved in the process.110 Inhibitory pathways allow colonic volume to adapt to its contents, and also mediate descending inhibition, which initiates colonic relaxation ahead of a fecal bolus.

Central Nervous System Pathways Regulating the Pelvic Floor Cortical mapping with transcranial magnetic stimulation suggests that rectal and anal responses are bilaterally represented on the superior motor cortex (Brodmann’s area 4).162 There are subtle differences in the degree of bilateral hemispheric representation between subjects. The external sphincter is innervated by motor neurons in Onuf’s nucleus, located in sacral segments, predominantly in the second sacral segment in cats.94,141 Though they supply striated muscles under voluntary control, these motor neurons are smaller than usual alpha motor neurons and resemble autonomic motor neurons72,140; however, the conduction velocity in pudendal nerve fibers is comparable to that of peripheral nerves.26,135 Unlike other somatic motor neurons in the spinal cord, these neurons are relatively spared in amyotrophic lateral sclerosis, but affected in Shy-Drager syndrome.25,159 Holstege and others have demonstrated that motor neurons in the dorsomedial and ventrolateral parts of Onuf’s nucleus supply the anal and urethral sphincters, respectively. A cellular bridge connects Onuf’s nucleus with the sacral intermediolateral cell column, which innervates the detrusor muscle of the bladder,72 facilitating coordination between the sphincter and detrusor muscles during micturition. Moreover, projections from

the pontine reticular formation project to Onuf’s nucleus and to sacral intermediolateral cell groups, facilitating coordination of processes involved in micturition.72 These dorsolateral pontine reticular formation neurons may also project to the nucleus retroambiguous (NRA) in the caudal medulla oblongata; the NRA is the only area projecting specifically to abdominal muscle motor neurons, and also projects to pelvic floor motor neurons.72 Stimulation of the caudal NRA induced simultaneous contraction of the abdominal and pelvic floor muscles or straining, perhaps explaining why pelvic floor sphincters contract during all maneuvers associated with raised intra-abdominal procedures with the exception of defecation.47 The paraventricular hypothalamic nucleus projects to virtually all autonomic motor neurons in the caudal brainstem and spinal cord, including Onuf’s nucleus and the sacral intermediolateral cell group.70 Diffuse projections from noradrenergic neurons in the locus ceruleus and the ventromedial medullary segmental field to Onuf’s nucleus also exist; their functions are unclear.71 Motor neurons in Onuf’s nucleus lack Renshaw inhibition94 and crossed disynaptic inhibition. In contrast to a joint, the sphincter is not composed of agonist and antagonist muscles, obviating the need for reciprocal inhibition. It is also conceivable that reduced inhibitory input may facilitate maintenance of resting anal sphincter tone.80 Somatic branches originating from Onuf’s nucleus travel in the pudendal nerve, the muscular branches, and the coccygeal plexus. The pudendal nerve branches into inferior rectal, perineal, and posterior scrotal nerves. The inferior rectal nerve supplies the lower part of the external anal sphincter, the anal canal lining, and the skin around the anus. The perineal nerve divides into posterior scrotal (or labial) branches and muscular branches. The posterior scrotal branches innervate the skin, while muscular branches are distributed to the transverse perinei, bulbospongiosus, ischiocavernosus, urethral sphincter, anterior part of the external anal sphincter, and levator ani. Motor fibers of the right and left pudendal nerves have overlapping distributions within the external anal sphincter. Sherrington observed that stimulation of the right pudendal nerve caused circumferential contraction of the external anal sphincter.150 Conversely, tonic external sphincter activity, sphincter inhibition during colonic distention, and the cutaneoanal reflex were not affected by sectioning either pudendal nerve.14

Afferent Innervation Colonic balloon distention in humans evokes a generally ill-defined sensation of gas, perhaps discomfort, and eventually pain at higher intensities. Rectal distention is perceived as a more localized sensation of rectal fullness, interpreted by the patient as a desire for motion or to pass wind. The anal canal responds to distention and to

Autonomic and Somatic Systems to the Anorectum and Pelvic Floor

innocuous mucosal proximodistal mechanical shearing stimuli.78 Sensory traffic is conveyed by unmyelinated small C fibers and larger A fibers. Although there are very few rectal mucosal nerve endings, the anal canal is lined by numerous free and organized nerve endings (i.e., Meissner’s corpuscles, Krause’s end bulbs, Golgi-Mazzoni bodies, and genital corpuscles), perhaps explaining why it is exquisitely sensitive to light touch, pain, and temperature. Animal models and clinicopathologic findings in humans suggest that pelvic nerves traveling to the sacral segments are more important for conveying non-noxious and noxious colonic sensations than lumbar colonic (sympathetic) nerves.36,57,58,79,112,113,137,171 There are more afferent neurons supplying the colon in the sacral, compared to lumbar, segments in the cat (7500 vs. 4500 neurons).5,108 However, the number of spinal visceral afferent neurons is relatively small, only 2.5% or less of the total number of spinal afferent neurons supplying skin and deep somatic structures.79 During colonic distention in cats, lumbar and sacral afferents can accurately encode intraluminal pressure ranging from 20 to 100 mm Hg by increasing discharge frequencies.77 Sacral afferents may be better suited for conveying afferent information than lumbar afferents, because they are more likely to lack resting activity and respond to pressure increments with a wider range of discharge frequency.3,15 Janig and Morrison identified three different classes of mechanosensitive visceral afferents in the cat colon. 79 Tonic afferents fired throughout colonic distention and accurately encoded the intensity of distention between 20 and 100 mm Hg. Phasic colonic afferents generally discharged at the onset, and occasionally at the cessation, of a distention stimulus. Tonic afferents were predominantly unmyelinated, slowly conducting C fibers, while most phasic afferents were faster conducting myelinated A fibers. The afferents innervating the anal canal responded to shearing stimuli, but not colonic or anal distention. Two different theories have been proposed to explain visceral pain perception. Proponents of the specificity theory suggested that pain was a distinct sensory modality, mediated by sequential activation of visceral nociceptors and central pain-specific neurons in the spinal dorsal horn.100 However, in the cat colon, Janig and Koltzenburg found no afferent fibers that were selectively activated by noxious stimuli, arguing against the specificity theory. The alternative hypothesis for pain perception—pattern or intensity theory—attributes pain perception to spatial and temporal patterns of impulses generated in nonspecific visceral afferent neurons.78 However, electrophysiologic studies of visceral afferent fibers in other organs, including the colon, have documented high-threshold visceral afferent fibers that only respond to noxious mechanical stimuli.100,147 Subsequently, Cervero and Janig reconciled these opposing concepts in a convergence model wherein input from low- and high-threshold mechanoreceptors

283

converges on spinothalamic and other ascending tract cells.24 Physiologic processes are generally accompanied by low-level activity, mediation of regulatory reflexes, and transmission of nonpainful sensations. High-intensity stimuli increase firing of low-threshold afferents and also activate high-threshold afferents, thereby activating nociceptive pathways and triggering pain.24 More than 90% of all unmyelinated pelvic afferents are silent, being activated by electrical stimulation, but not by even extreme noxious stimuli.79 Silent afferents can respond to chemical stimuli or tissue irritation, however, becoming responsive to even innocuous mechanical stimuli after sensitization.100 These neurophysiologic changes are detectable within minutes after tissue irritation and likely to persist for the duration of irritation. Putative mediators of sensitization include bradykinin, histamine, serotonin, catecholamines, prostanoids, hydrogen and potassium ions, adenosine, substance P, cytokines, and other neuropeptides. Upregulation of silent afferents by inflammation has been implicated to explain visceral hypersensitivity, as is observed in patients with functional gastrointestinal disorders after an acute diarrheal infection.60

ANAL SPHINCTER TONE AND REFLEXES Internal Anal Sphincter The internal anal sphincter is a thickened extension of the circular smooth muscle layer surrounding the colon and is generally considered responsible for maintaining approximately 70% of resting tone, ensuring that the anal canal is closed at rest.51,58 Internal sphincter tone is primarily attributable to smooth muscle activity; the extent to which extrinsic nerves contribute to resting anal sphincter tone is unclear. While resting anal sphincter tone is not affected by spinal transection between C6 and L1,36,49 high spinal anesthesia reduced resting tone, suggesting extrinsic nerves may contribute to resting sphincter tone.51,87 Moreover, resting sphincter tone declined by an average of 30% after bilateral hypogastric nerve sectioning and an average of 63% after lumbar colonic nerve sectioning in dogs,106 returning to baseline 1 month thereafter. Conversely, sympathetic stimulation evoked either internal anal sphincter relaxation,73,93,121 or contraction followed by relaxation.22 Gowers demonstrated that rectal insufflation with air induced anal relaxation; anal pressure returned to normal thereafter.59 The rectoanal inhibitory reflex (RAIR) is mediated by intrinsic nerves and absent in Hirschsprung’s disease. The extrinsic nerves are not essential for the reflex, since it is preserved in patients with cauda equina lesions or after spinal cord transection.36 However, extrinsic nerves may modulate the reflex, since relaxation is more pronounced and prolonged in children with sacral agenesis.107 The RAIR is

284

Function of the Peripheral Nervous System

probably mediated by nitric oxide; morphologic studies reveal an efferent descending nitrergic rectoanal pathway.153 Other nonadrenergic-noncholinergic neurotransmitters (i.e., vasoactive intestinal peptide and ATP) may also participate in the RAIR.12,116

External Anal Sphincter Though resting sphincter tone is predominantly attributed to the internal anal sphincter, studies under general anesthesia or after pudendal nerve block suggest the external anal sphincter generally accounts for approximately 25%, and up to 50%, of resting anal tone.9,33,40,50,144,170 When continence is threatened, the external sphincter contracts to augment anal tone, preserving continence. This “squeeze” response may be voluntary or induced by increased intra-abdominal pressure,47 or by merely moving a finger across the anal canal lining.161 Conversely, the external sphincter relaxes during defecation. The only other striated muscles that display resting activity are the puborectalis, external urethral sphincter, cricopharyngeus, and laryngeal abductors. Resting or tonic activity depends on monosynaptic reflex drive, perhaps explaining why resting anal sphincter tone is reduced, but voluntary contraction of the external sphincter is preserved, in tabes dorsalis.102 However, the presence of muscle spindles in the human external anal sphincter is debatable.16,39 In cats, the tonic reflex is maintained by afferents entering the spinal cord at S2, while the phasic contractile “guarding”

responses to perineal scratch and anal penetration disappear after bilateral section of S3 dorsal roots.13 Thus the afferent axons involved in tonic and phasic reflexes are different and penetrate the spinal cord at different levels. Moreover, afferent discharge in the pudendal nerve is directly related to muscular activity of the external sphincter.13 The fiber distribution also favors tonic activity; type I fibers predominate in the human anal sphincter,142 while that of cats and rabbits predominantly contains type II or fast-twitch muscle fibers.85 External sphincter fibers are circumferentially oriented and very small, separated by profuse connective tissue.142 Although the active lengthtension relationship in the cat external anal sphincter is similar to that in other mammalian skeletal muscles, the muscle generates passive tension even below the optimal length.86 Above optimal length, the passive length-tension relationship is considerably steeper compared to skeletal muscles. Krier and Adams speculated that higher passive stiffness might prevent the external sphincter from being stretched beyond optimal length for force generation during defecation.84

Puborectalis The tonically active puborectalis muscle maintains the resting anorectal angle. Moreover, puborectalis contraction during a sudden rise in abdominal pressure reduces the anorectal angle, preserving continence (Fig. 13–4). While cadaveric studies suggested the puborectalis was supplied

FIGURE 13–4 Magnetic resonance fluoroscopic images of the pelvis at rest (Left), during squeeze (middle), and during simulated defecation (right) in a subject. The asymptomatic rectum was filled with ultrasound gel. At rest, the pelvic floor was well supported; the anorectal angle was relatively obtuse (126 degrees). Pelvic floor contraction during the squeeze maneuver was accompanied by normal upward and anterior motion of the anorectal junction; the angle declined to 95 degrees. During rectal evacuation, the bladder base dropped by 2.5 cm below the pubococcygeal line; a 2.8-cm anterior rectocele emptied completely, and was probably not clinically significant; perineal descent (5 cm) was outside the normal range for evacuation proctography.

Autonomic and Somatic Systems to the Anorectum and Pelvic Floor

285

by the pudendal nerve, electrophysiologic stimulation studies in humans suggest this muscle is supplied, strictly ipsilaterally, by branches originating from the sacral plexus above the pelvic floor.122 This spinal reflex response may be mediated by stretch receptors in the levator ani.163 While no or minor incontinence results after surgical division of the internal or external anal sphincter, disruption of the puborectalis inevitably causes significant incontinence, underscoring the importance of this muscle in maintaining continence.104

and pudendal evoked response, can be recorded after electrical stimulation of the dorsal nerves of the glans penis or clitoris.163 Shafik has described several pelvic sphincter responses elicited by stimulating the urethra or anal canal.148,149 The role of these reflexes in normal physiology or as investigative tools is unclear.

Sacral Reflexes

Sympathetic nerves excite while parasympathetic nerves inhibit the sphincters, which is the reverse of their effects on nonsphincteric regions. The internal anal sphincter in humans and monkeys has a dense adrenergic innervation. Stimulation of adrenergic  and  receptors contracted and relaxed human internal sphincter strips, respectively.119 The internal anal sphincter is more sensitive to adrenergic compared to cholinergic agonists; cholinergic agonists either contracted or relaxed internal anal sphincter strips in humans.119 In the vervet monkey, muscarinic receptor stimulation contracted internal sphincter strips, but relaxed them after muscarinic receptors were blocked, probably via nonadrenergic-noncholinergic mechanisms.127 Nicotinic agonists also relaxed internal sphincter strips, probably via nonadrenergic-noncholinergic mechanisms. Table 13–1 summarizes the effects of neurotransmitters and pharmacologic agents that modulate internal anal sphincter tone.

The pelvic floor striated muscles contract reflexly in response to stimulation of perineal skin (i.e., a somatosomatic reflex) or anal mucosa (i.e., a viscerosomatic reflex). The cutaneoanal reflex is elicited by scratching or pricking the perianal skin and involves the pudendal nerves and S4 roots. Electrophysiologic recordings reveal both short latency and long latency responses; the latter corresponds to the visible anal sphincter contraction.160 The short latency response has a latency of 5 to 6 ms, which is too short for a spinal reflex response because the latency from the conus medullaris to the anal sphincter is approximately 7 ms.98 Thus the short latency response is probably attributable to direct stimulation of nerve fibers in or near the sphincter, causing an antidromic volley up to a branch point, followed by orthodromic conduction back to muscle. The longer, stimulus-dependent response is probably polysynaptic, and has an extremely variable latency, limiting its utility as a diagnostic tool.163 In humans, electrical activity of the internal anal sphincter increases during urinary bladder emptying, returning to normal after micturition.138 Vesical distention in cats also increased anal pressure.54 Conversely, the external anal sphincter relaxed during micturition in humans,143 cats,14 and dogs.146 The pudendoanal reflex, also known as the sacral evoked potential, sacral reflex, pudendal sexual reflex,

PHARMACOLOGY

MECHANISMS OF CONTINENCE AND DEFECATION In addition to the pelvic barrier, the rectal curvatures and transverse rectal folds are other anatomic factors impeding fecal evacuation; rectoanal sensation and rectal compliance

Table 13–1. Effects of Neurotransmitters on the Anal Sphincters Neurotransmitter/Pharmacologic Agent

Model

Effects

Adrenergic Phenylephrine

In vitro—human and monkey Healthy subjects and incontinent patients

Acetylcholine52,119 Nicotine52,119 Bethanechol Loperamide Others: glucagon, ketanserin, diltiazem, enkephalin, somatostatin121

In vitro—human and monkey In vitro—human and monkey Healthy human subjects Incontinent patients Healthy human subjects

: contraction; : relaxation q resting pressure in healthy subjects; did not reduce incontinence Either contraction, relaxation, or 4 Relaxation (upper); 4 (lower) p resting pressure q resting pressure p resting pressure

119

q  increased; p  decreased; 4  no change.

286

Function of the Peripheral Nervous System

also help maintain continence. Defecation can often be postponed until convenient because the rectum relaxes or accommodates to retain stool, and the external sphincter and/or puborectalis contract voluntarily; this “squeeze” response is critically dependent on normal rectoanal sensation. Stool is often transferred into the rectum by colonic high-amplitude propagated contractions, which often occur after awakening or meals.8 Denny-Brown and Robertson observed that rectal distention evoked rectal contraction and anal sphincter relaxation, facilitating evacuation.37 Rectal distention by stool is associated with several processes that serve to preserve continence or, if circumstances are appropriate, proceed to defecation. Rectal distention induces reflex relaxation of the internal anal sphincter. The anal sphincter may also relax independently of rectal distention, allowing anal epithelium to periodically “sample” and ascertain whether rectal contents are gas, liquid, or stool.103 Perception of rectal distention prompts voluntary contraction of the external sphincter if defecation is inconvenient, or relaxation if defecation is convenient.156 The voluntary components of defecation include assumption of an appropriate posture, abdominal contraction, and pelvic floor relaxation (see Fig. 13–4). The central nervous system plays a greater role in regulating anorectal sensimotor functions compared to other regions of the gastrointestinal tract. Garry observed that colonic stimulation in cats induced colonic contraction and anal relaxation even after destruction of the lumbosacral cord, concluding that the gut “seems not to have wholly surrendered its independence.”55 However, the elaborate somatic defecation response depends on centers above the lumbosacral cord, and probably craniad to the spinal cord itself.

ANORECTAL DYSFUNCTION IN NEUROLOGIC DISORDERS Common neurologic disorders associated with anorectal manifestations of pelvic floor dysfunction include stroke, autonomic neuropathies, multiple sclerosis, Parkinson’s disease, and some forms of multiple system atrophy. Parkinson’s disease and multiple sclerosis are associated with paradoxical puborectalis contraction, causing obstructed defecation. 30,99 Anorectal sensorimotor disturbances in diabetes mellitus may be symptomatic or asymptomatic.133 Onuf’s nucleus is relatively spared in diseases that affect the somatic motor system (i.e., motor neuron disease),83 but severely affected in diseases that involve visceromotor neurons (e.g., Shy-Drager syndrome,97 mannosidosis, Hurler’s syndrome,158 and Fabry’s disease157).

Obstructed Defecation Obstructed defecation is a disorder of function attributable to either an inadequate “pushing” force or an inability to overcome resistance to expulsion caused by inadequate relaxation of the external anal sphnicter (i.e., anismus) and/or puborectalis sling (i.e., puborectalis dyssynergia).10 In addition to excessive straining during defecation, other features of obstructed defecation include the need for digital removal of stool from the rectum, and a sense of incomplete evacuation after defecation.173 Most patients have “idiopathic” obstructed defecation. Multiple sclerosis and Parkinson’s disease are perhaps the most common neurologic diseases associated with obstructed defecation. The diagnosis is predominantly based on the history and anorectal examination. Anorectal inspection may disclose prominent external hemorrhoids or an anal fissure. Resting anal tone, as gauged by the resistance to inserting a finger in the anal canal, is increased in anismus. On palpation, the puborectalis should contract and generally relax when patients are asked to squeeze and expel the examining finger, respectively. Paradoxical puborectalis contraction during simulated defecation is suggestive, but not necessarily diagnostic, of obstructed defecation. The perineum normally descends by 1 to 3 cm during simulated defecation. Perineal descent is often reduced in anismus. Conversely, perineal descent may be increased in other patients (see Fig. 13–4). One or more tests may be necessary to confirm a clinical suspicion of obstructed defecation. Anal manometry may reveal a high anal resting pressure; though normal ranges are age, gender, and technique dependent, an average resting pressure greater than 100 mm Hg is probably abnormal and suggestive of anismus. Anal pressure may be unchanged, or paradoxically increase, instead of declining during simulated evacuation; the ratio of rectal to anal pressures during expulsion, termed the defecation index, can quantify this process.126 The importance of considering test results in the overall clinical context cannot be overemphasized; paradoxical contraction has been observed in up to 20% of asymptomatic subjects and may not always indicate anismus.166 The balloon expulsion test is more than 80% sensitive and more than 80% specific for identifying pelvic floor dysfunction.123,126 Subjects are asked to expel a rectal balloon, connected over a pulley to a series of weights (Fig. 13–5). Patients without pelvic floor dysfunction can expel the balloon with no or limited external rectal traction. Patients with pelvic floor dysfunction require additional external traction to facilitate expulsion of the rectal balloon. Notwithstanding technical limitations,4 evacuation proctography is often useful for confirming clinically suspected rectocele or pelvic floor dysfunction not confirmed by other diagnostic tests. More recently, rapid magnetic resonance imaging (MRI) sequences have been developed to visualize pelvic floor motion in

287

Autonomic and Somatic Systems to the Anorectum and Pelvic Floor

Balloon

Weight

FIGURE 13–5 Rectal balloon expulsion test. Patients are asked to expel a latex balloon filled with 50 mL warm water and connected over a pulley to a series of weights. Graded external traction is provided by adding weights as necessary to facilitate balloon expulsion.

real-time without radiation exposure, as discussed below.46 Finally, since colonic transit is often delayed in obstructed defecation, it is necessary to exclude obstructed defecation before making a primary diagnosis of slow-transit constipation in patients with delayed colonic transit (Fig. 13–6).126

Fecal Incontinence Fecal incontinence is a multifactorial symptom attributable to reduced stool consistency, and/or impairment in one or more factors responsible for maintaining continence (i.e., pelvic floor barrier, rectoanal sensation, and rectal

compliance).11 The most common predisposing factor for incontinence in women is obstetric injury to the pelvic floor and pudendal nerve,154 often compounded by diarrhea. Disorders of the peripheral nervous system causing incontinence include spinal cord lesions, lumbosacral nerve root lesions, and pudendal neuropathy. Lesions of the conus medullaris or cauda equina causing fecal incontinence are also associated with weakness and/or sensory disturbances in the lower extremities and/or paraspinal muscles. Lumbosacral plexopathy is a rare complication of metastatic cancer, pelvic radiotherapy,75 or pregnancy.81 Other neurologic disorders associated with fecal incontinence include dementia,109,111,134 multiple sclerosis,23,66,67 and multiple system atrophy. The evidence for denervation in “idiopathic” fecal incontinence was based on histologic changes, more pronounced in the external sphincter compared to the puborectalis.120 Subsequent studies demonstrated prolonged pudendal nerve latencies in constipated patients with excessive perineal descent,82 and even after 1 minute of acute straining in symptomatic patients.42 It has been suggested that, over the long term, prolonged straining during defecation causes perineal descent, thereby stretching and injuring the pudendal nerves; pudendal nerve injury may lead to denervation atrophy of pelvic floor muscles supplied by these nerves. Perhaps this explains why some patients with pelvic floor dysfunction may progress from constipation-predominant symptoms to excessive perineal descent, ultimately developing fecal incontinence.

2 1 hr

4 hr

24 hr

48 hr

1

3 0 4 5 Geometric Center

Normal Range

1.3

1.6

1.7

0.7-1.7

1.6 - 3.8

3.0 - 4.8

FIGURE 13–6 Scintigraphic assessment of colonic transit in a patient with obstructed defecation. Transit was assessed by a capsule containing 111In. The pH-sensitive methacrylate coating of the capsule dissolved in the alkaline pH of the terminal ileum, releasing the isotope in the cecum at time 0. Scans taken over the next 48 hours monitored isotope progression throughout the colon. At 24 and 48 hours, the isotope was predominantly distributed in the transverse colon, as reflected by a geometric center of 1.7, which corresponds to the mid-transverse colon. The geometric center is the weighted distribution of counts throughout the colon.

288

Function of the Peripheral Nervous System

Clinical Features The clinical assessment is critical for making an accurate diagnosis, establishing rapport with the patient, and formulating a logical strategy for diagnostic testing and treatment. Patients with chronic diarrhea, fecal urgency, constipation, prolapsed hemorrhoids, urinary incontinence, dementia, and diabetes mellitus must be asked whether they have fecal incontinence because they may be embarrassed to disclose the symptom.89 Staining, soiling, seepage, and leakage are terms used to reflect the nature and severity of incontinence. Soiling indicates more leakage than staining of underwear and can be qualified further to indicate severity (i.e., of underwear, outer clothing, or furnishing/bedding). Seepage refers to leakage of small amounts of stool. Severity can also be rated by scales that quantify amount and frequency of stool loss; some scales also incorporate number of pads used, severity of urgency, and the impact of fecal incontinence on coping mechanisms and/or lifestyle-behavioral changes.11 For incontinent patients, the correlation between symptom severity and quality of life may not be perfect. Physicians and patients generally agree that infrequent incontinence for flatus reflects mild, while frequent leakage of liquid and stool Rest Posterion mmHg 100 2.0

reflects severe, incontinence.131 In other areas, physicians and patients may disagree; patients attach greater significance to frequent leakage of gas compared to physicians, while physicians attach greater significance to leakage of solid stool. Consequently, it is vital to ascertain quality of life, reflected not only by items connected with coping, behavior, self-perception, and embarrassment,132 but also practical day-to-day limitations (e.g., the ability to socialize and get out of the house).19 Patients are affected even by the possibility and unpredictability of incontinence episodes;19 conversely, lifestyle or behavioral adjustments may also influence the type and frequency of incontinent episodes. Diagnostic Tests Anal Manometry. Anal manometry provides an overall functional assessment of the anal sphincters. Pressures are measured by withdrawing a catheter with perfused or solid-state transducers by the station pull-through method; resting and squeeze pressures are recorded at serial 1-cm intervals from the rectum to the anal verge (Fig. 13–7). Though anal manometry is technically undemanding and widely available, variations in catheter design, definitions,

Squeeze

Resting P

1

50 mmHg 100

Resting P

Right 2.0

2

50 100

Anterior

Resting P

2.0

3

50 mmHg

Left

Resting Pr

2.0

4

50 mmHg 50 0 mmHg 50

Right pos Resting P 4.0

5

Right Ante Resting P 4.0

6

0 mmHg 50

Left Anter Resting P 4.0

7 0 mmHg

Left Poste Resting P 4.0

8 0

10 second

30 second

FIGURE 13–7 Assessment of squeeze response by manometry. Tracing shows resting and squeeze pressures recorded by circumferentially oriented water-perfused sensors located 2 cm (sensors 1–4) and 4 cm (sensors 5–8) from the anal verge. The squeeze pressure was sustained for approximately 30 seconds in this healthy subject.

Autonomic and Somatic Systems to the Anorectum and Pelvic Floor

and methods to calculate average and maximum resting and squeeze pressures between centers should be borne in mind when interpreting pressures. Also, normal values are strongly influenced by technique, age, and gender; both resting and squeeze pressures decline with increasing age, even in asymptomatic subjects.76,101 Therefore, anal pressures should be compared to normal values obtained using the same technique. Average resting and squeeze pressures are generally low in incontinent patients. Other factors causing incontinence (e.g., disordered stool consistency, rectoanal sensation, and/or rectal compliance) should be considered, particularly in incontinent patients with normal sphincter pressures.152 Pudendal Nerve Latencies. Delayed pudendal nerve terminal motor latencies (PNTMLs) are regarded as surrogate markers of pudendal nerve injury, and used to ascertain if anal sphincter weakness is attributable to pudendal nerve injury, sphincter defect, or both.4 The pudendal nerve is stimulated by an electrode on a gloved index finger as the finger courses around the pelvic brim. Measuring the shortest latency between stimulus delivery and recording is critically dependent on close proximity of the examining finger to the nerve. Delayed pudendal nerve latencies have been extensively used to identify pudendal nerve injury caused by parturition,154 and perineal descent related to excessive straining,42 and prior to surgical repair of sphincter defects.96 However, the utility of PNTMLs as a marker for pudendal nerve injury is questionable since pudendal neuropathy is attributed to an axonopathy rather than demyelination. Because PNTML is dependent only on the fastest conducting fibers in the pudendal nerve, nerve latencies may be normal if even a few normally conducting fibers remain. There are several methodologic limitations: test reproducibility is unknown, normative data are limited, and multiple factors (i.e., age, parturition, body mass index) influencing nerve latencies are not generally taken into account. The “normal” ranges for nerve latencies are extremely stringent. Thus latencies that would be regarded as within normal limits for other peripheral nerve injuries are regarded as abnormal for the pudendal nerve. The sensitivity and specificity of the test are uncertain. In one study, approximately 50% of patients with prolonged PNTML had normal anal canal squeeze pressures.169 Finally, in contrast to initial studies, recent studies suggest the test does not predict improvement, or lack thereof, after surgical repair of anal sphincter defects.96 A recent consensus statement concluded that PNTML “cannot be recommended for evaluation of patients with fecal incontinence.”4 Anal Sphincter Electromyography. Endoanal ultrasound (US) has replaced anal sphincter electromyography (EMG) for identifying external anal sphincter defects. Given the limitations of PNTML, concentric needle anal sphincter

289

EMG may be useful for identifying myopathic (small polyphasic motor unit potentials), neurogenic (large polyphasic motor unit potentials), or mixed injury.29,35 The external anal sphincter is examined on each side with one or two needle insertions. The puborectalis muscle is examined by inserting a needle in the midline between the anus and tip of the coccyx, passing the needle through the external anal sphincter and into the deeper puborectalis. Insertional activity at rest and motor unit potential (MUP) amplitude, duration, percent polyphasia, and recruitment following mild to moderate voluntary muscle contraction are assessed. These parameters are measured in a standardized semiquantitative manner that has been demonstrated to have minimal intraobserver and interobserver variability.34,35 Assessment of MUP recruitment is also a reliable measure of muscle weakness in neurogenic lesions. Needle EMG of the puborectalis muscle can be used to distinguish disorders that affect this muscle and the external sphincter muscle selectively or in combination.6 In experienced hands, anal sphincter EMG is not accompanied by severe discomfort. EMG recording by an anal sponge or hard plug electrode correlates well with sphincter pressures, but cannot distinguish among the causes of sphincter weakness specified above. Podnar and colleagues quantified EMG activity in each of four sites in the subcutaneous and/or deeper portion of the external anal sphincter; the taped EMG signal was analyzed off-line.124 Because the level of continuous motor unit activity in the relaxed external sphincter was low, most parameters for the EMG interference pattern were zero. Therefore, the authors counted the number of MUPs, and estimated the lower limits of normal at each of four sites in the subcutaneous and deeper portions of the external anal sphincter. MUP counts were larger in men and in the subcutaneous external sphincter; MUP counts were not affected by age, parity, or fecal continence status.124 Assessment of Corticoanal Pathways. Corticoanal pathways to the external anal sphincter have been investigated by recording motor evoked potentials and pressure profiles from the anal canal in healthy subjects after cortical electrical or magnetic stimulation in research studies62,63; these assessments are not routinely used in clinical practice. Magnetic stimulation is less painful and therefore preferred to electrical stimulation. A more intense stimulus was required to elicit motor responses in the bulbocavernosus muscle and anal sphincter compared to the upper or lower extremities.117 Voluntary facilitation of the pelvic floor muscles reduced the latency for the electrical response in some,117 but not all, studies63; the latency for the pressure response to electrical stimulation was shorter during voluntary facilitation.63 Moreover, contraction times in the external sphincter were similar across a wide range of amplitudes during nonfacilitated,

290

Function of the Peripheral Nervous System

facilitated, and voluntary contractions, indicative of isochronism (i.e., the external sphincter adjusts the rate of rise in pressure during a contraction to attain a wide range of pressure amplitudes within a constant time).63 Lumbosacral and pudendal nerve stimulation facilitated cortical pathways to the external anal sphincter. Magnetic stimulation of the lumbosacral nerve roots or electrical stimulation of the pudendal nerve augmented the amplitude and shortened the latency of the anal sphincteric response to the cortical stimulation.61 Facilitation of sphincteric responses to cortical activation is probably attributable to direct antidromic excitation of sacral motor neurons innervating the external anal sphincter by lumbosacral or pudendal nerve stimulation.92 Moreover, lumbosacral, but not pudendal, nerve stimulation induced a second delayed rise in corticoanal response amplitude, which was maximal when the duration between lumbosacral nerve and cortical stimulation was 100 ms.61 Perhaps this late facilitation was attributable to cortical excitation by activation of ascending afferent pathways, indicative of sensorimotor, peripheral-central neural interactions. The relationship between stimulus intensity and response latency for the anal sphincter differs from that for the tibialis anterior.44 For the tibialis anterior, low-intensity cortical electrical stimulation induced a fast response with a latency of 30 ms, consistent with monosynaptic conduction, while a late response (latency  100 ms) occurred during stimulation at a higher intensity. Controversially, in the external anal sphincter, the late response occurred at a very low stimulus intensity, suggesting activation of highly excitable central motor neurons connected to polysynaptic pathways. The response latency was shorter at higher stimulus intensities;

only one response could be recorded at a time and no inhibition was observed. These observations suggest that, despite different latencies, all responses are transmitted by the same pathway; higher stimulus intensities probably circumvent some interneurons in the pathways. The indications, strengths, and limitations of other tests are detailed below and summarized in Table 13–2. Endoanal Ultrasound. Ultrasound can identify internal sphincter thinning or defects,7,164 and external sphincter scars or defects; these defects are often clinically unrecognized,154 and, in a small proportion of patients, external sphincter defects are amenable to surgical repair (Fig. 13–8).114 Interpretation of external sphincter images is more subjective, operator dependent, and confounded by normal anatomic variations in the external sphincter7,64; there is substantial interobserver variability.41,56 The etiology of internal sphincter thinning is unknown. Since the external sphincter is often asymmetrical in the upper anal canal, particularly in women, it may be difficult to distinguish a normal variant from a sphincter defect.7,74 Moreover, the significance of clinically occult isolated external anal sphincter defects to fecal incontinence is unclear, since these defects have been observed in 1 in 3 women after a vaginal delivery.154 Finally, the external sphincter and perirectal fat are both echogenic and frequently indistinguishable, precluding accurate characterization of external sphincter thickness and identification of external sphincter atrophy. Evacuation Proctography (Defecography). During evacuation proctography, anorectal anatomy and pelvic floor motion are recorded on video with the patient at rest,

Table 13–2. Investigation of Fecal Incontinence Test

Purpose/Common Findings

Limitations

Anorectal manometry

p resting and squeeze pressure reflect internal and external anal sphincter weakness, respectively p or q rectal sensation occur commonly in fecal incontinence and may be remediable by biofeedback therapy Internal and external sphincter defects or scars are most frequently attributable to obstetric or surgical trauma Useful for documenting rectocele, perineal descent, enterocele, and internal rectal intusussception

Technically undemanding, but methods vary between centers

Rectal sensation

Endoanal ultrasound Evacuation proctography (barium defecography) Pelvic MRI

Anal sphincter EMG Pudendal nerve terminal motor latency

• Static imaging visualizes morphology of anal sphincters and entire pelvic floor (i.e., bladder, uterus, and rectum) • Dynamic imaging of pelvic floor motion visualizes prolapse Identify neurogenic, myogenic, and mixed lesions affecting external anal sphincter Delayed nerve latencies reflect pudendal nerve injury

q  increased; p  decreased; EMG  electromyography; MRI  magnetic resonance imaging.

Variations in normal anatomy may be mistaken for an external sphincter defect • Measurements not standardized • Imperfect correlation between symptoms and findings Not widely available

Potentially painful Several methodologic limitations significantly restrict utility of measurements

291

Autonomic and Somatic Systems to the Anorectum and Pelvic Floor

FIGURE 13–8 Endoanal ultrasound (US) image of anal sphincters in a patient with fecal incontinence. The internal anal sphincter is hypoechoic. Thick and thin arrows indicate normal internal sphincter and tear, respectively (located approximately between 10 and 5 o’clock) on US images. Large and small arrowheads indicate normal-appearing and partially torn external sphincter (between 10 and 2 o’clock), respectively.

coughing, squeezing, and straining to expel barium from the rectum; the anorectal angle and position of the anorectal junction are tracked during these maneuvers. Prior to dynamic MRI, evacuation proctography was the only modality for identifying excessive perineal descent, internal rectal intussusception, rectoceles, sigmoidoceles or enteroceles; puborectalis dysfunction during squeeze and evacuation can also be characterized.1 However, the methods for conducting and interpreting evacuation proctography are incompletely standardized4; for incontinent patients, a thick barium paste (Anatrast; E-Z-EM, Westbury, NY) is probably preferable to liquid barium. Despite these limitations, evacuation proctography in incontinent patients may be useful, particularly prior to surgery, when there is a high index of suspicion for excessive perineal descent, a significant rectocele (e.g., 2 cm in size, with the need to splint the vagina to facilitate rectal emptying), enterocele, or internal rectal intussusception. These findings may also help educate patients about the nature of the disorder and reinforce the need for treatment (e.g., pelvic floor retraining). Pelvic MRI. With rapid MRI sequences, MRI can visualize both anal sphincter anatomy and global pelvic floor motion in real time, without radiation exposure.46 The anal sphincters are visualized by axial T2-weighted fast spin echo images and corresponding T1-weighted

spin echo images with a disposable endorectal colon coil.46 Disagreement exists over the best technique for the evaluation of the internal sphincter; however, MRI performed the same as or better than US for the assessment of the external sphincter.95,130 In contrast to US, MRI can also identify external sphincter atrophy.18 External sphincter atrophy is a good prognosticator for poor continence after repair of external sphincter defects.17 Dynamic images are acquired as patients squeeze their sphincters or try to expel US gel from the rectum, providing a unique appreciation of global pelvic floor motion—in addition to the anorectum, the bladder and genital organs are also visualized (see Fig. 13–4).46,129 As with any new diagnostic imaging modality, comparisons to age- and gender-matched asymptomatic subjects are necessary to assess the role of static and dynamic MRI as a diagnostic tool in clinical practice. Rectoanal Sensation. Rectal sensation is assessed by progressively distending a latex balloon manually, or a polyethylene balloon by a barostat, while measuring thresholds for first perception, desire to defecate, and severe discomfort (Fig. 13–9).172 Alternatively, perception is recorded by a visual analogue scale during phasic distentions of graded intensity by a barostat.88 Rectal sensation is reduced in diabetes, in multiple sclerosis, and after spinal cord injury23,66,67,133,155; within limits, rectal sensation can be modulated by biofeedback therapy, improving continence.105 Anal sensation is assessed by determining the perception threshold to an electrical

1

Manometric sensor Colonic ballon

2 Barostat 3

Rectal ballon To second barostat

6

5 Volume Pressure

FIGURE 13–9 Barostat-manometer assembly. Highly compliant polyethylene balloons positioned in the colon and/or rectum can be distended to assess motor activity, sensation, and compliance. Motor activity is generally recorded by a balloon inflated to a fixed pressure and opposed to the colonic mucosa; under these circumstances, reduced or increased colonic volume reflects colonic contraction or relaxation, respectively.

292

Function of the Peripheral Nervous System

stimulus, or temperature change in the anal canal. Anal sensitivity to temperature change is reduced in fecal incontinence.139 Rectal Compliance. Compliance is a measure of rectal distensibility, reflecting rectal ability to hold a larger volume of stool at a given pressure, thereby postponing defecation to a convenient time37; the rectum becomes less compliant with aging, perhaps predisposing to incontinence.48 Compliance is measured by assessing rectal pressure-volume relationships with a balloon, either by manually inflating a latex balloon with air or water or, more accurately, by distending a highly compliant polyethylene balloon with a barostat (see Fig. 13–9). Rectal compliance is reduced in ulcerative and radiation proctitis, perhaps contributing to fecal urgency and incontinence.20,32,125

MANAGEMENT OF FECAL INCONTINENCE The management must be tailored to clinical manifestations, and includes treatment of underlying diseases, and other approaches detailed in Table 13–3.

Bowel Habit Modification Modifying irregular bowel habits is the cornerstone to effectively managing incontinence, particularly when pelvic floor weakness is attributable to an irremediable neuromuscular disorder. Several simple approaches are underutilized because of a misperceived lack of benefit. Accurate characterization of bowel habits is necessary to tailor therapy. Diarrhea may be attributable to loss of the sympathetic inhibitory “brake,” or coexistent bacterial overgrowth in patients with an autonomic neuropathy;

Table 13–3. Management of Fecal Incontinence Intervention

Side Effects

Comments

Mechanism of Action

Skin irritation

Disposable products provide skin protection superior to nondisposable products Underpad products were slightly cheaper than body-worn products

Provide skin protection and prevent soiling of linen; polymers conduct moisture away from the skin

Constipation

Titrate dose for each patient; administer before meals and social events

q fecal consistency, purgency; q anal sphincter tone

151

Incontinence pads* Disposable body-worns, reusable body-worns, disposable underpads, and reusable underpads Disposable body-worns are the largest category Antidiarrheal agents* Loperamide (Imodium): up to 16 mg/day in divided doses Diphenoxylate: 5 mg qid Enemas†

Inconvenient; side effects of specific preparations

Biofeedback therapy using anal canal pressure or surface EMG sensors† 115 Rectal balloon for modulating sensation Sphincteroplasty for sphincter defects† 2 Sacral nerve stimulation† Artificial sphincter Gracilis transposition†

Wound infection Recurrent incontinence (delayed) Infection Device erosion, failure, and infection

Rectal evacuation decreases likelihood of fecal incontinence Prerequisites for success include motivation, intact cognition, absence of depression, and some rectal sensation

Improved rectal sensation and coordinated external sphincter contraction;  q anal sphincter tone

Restricted to isolated sphincter defects without denervation

Restore sphincter integrity

Preliminary uncontrolled trials promising Either artificial device or gracilis transposition with/without electrical stimulation

Unclear; q anal sphincter tone; may modulate rectal sensation Restore anal barrier

*Grade A and †grade B therapeutic recommendations are supported by at least one randomized controlled trial, or one high-quality study of nonrandomized cohorts. q  increased; p  reduced;   possible. Data from Bharucha, A. E., and Camilleri, M. H.: GI dysmotility and sphincter dysfunction. In Noseworthy, J. H. (ed.): Neurological Therapeutics: Principles and Practice. London, Martin Dunitz, p. 2255, 2003.

Autonomic and Somatic Systems to the Anorectum and Pelvic Floor

evaluation and management are covered in detail in Chapter 122. The sympathetic brake may be restored by clonidine, reducing diarrhea.45 In addition to reducing diarrhea, loperamide, given at an adequate dose (i.e., 2 to 4 mg 30 minutes before meals, up to 16 mg/day), slightly increased internal sphincter tone and reduced incontinence128; the importance of taking loperamide before meals and in adequate doses cannot be overemphasized. The dose must be titrated to reduce diarrhea but avoid constipation. By taking loperamide before social occasions, or meals outside the home, incontinent patients may avoid having an accident outside the home, and gain confidence in their ability to participate in social activities. Diphenoxylate is an alternative option for diarrhea118; the serotonin 5-HT3 antagonist alosetron (Lotronex), currently available under a restricted use program, may benefit patients with severe diarrhea whose symptoms do not respond to other agents. Patients with constipation, fecal impaction, and overflow incontinence may benefit from a regularized evacuation program, incorporating timed evacuation by digital stimulation and/or bisacodyl/glycerol suppositories, fiber supplementation, and selective use of oral laxatives as detailed in recent reviews.90,91

Pharmacologic Approaches to Increase Sphincter Pressure There are no proven approaches. Phenylephrine suppositories increased anal resting pressure by 33% in healthy subjects and incontinent patients.28 However, in a randomized, double-blind, placebo-controlled crossover study, phenylephrine did not significantly improve incontinence scores or resting anal pressure fecal incontinence.21

Biofeedback Therapy Biofeedback therapy is based on the principle of operant conditioning.43 Using a rectal balloon–anal manometry device, patients are taught to contract the external anal sphincter when they perceive balloon distention; perception may be reinforced by visual tracings of balloon volume and anal pressure, and the procedure is repeated with progressively smaller volumes. Several uncontrolled studies suggest continence improves in approximately 70% of patients65; controlled studies are in progress. Sensory assessments (i.e., preserved baseline sensation and improved sensory discrimination after biofeedback therapy) are more likely to be associated with improved continence after biofeedback therapy than are sphincter pressures.105,167

293

Surgical Approaches A colostomy is the last resort for patients with severe incontinence.27 The results of overlapping anterior sphincteroplasty for anal sphincter defects are disappointing96; the procedure may be particularly inadvisable when fecal incontinence is attributable to neuromuscular weakness. The more complicated surgical procedures (e.g., dynamic graciloplasty and artificial anal sphincter) are fraught with significant morbidity, precluding widespread use.27 Sacral nerve stimulation is approved by the Food and Drug Administration for treating urge urinary incontinence in the United States. Sacral nerve stimulation may also augment anal pressures, modulate rectal sensation, and significantly improved continence.53,136,165 Sacral stimulation is conducted as a staged procedure; patients whose symptoms respond to temporary stimulation over approximately 2 weeks proceed to permanent subcutaneous implantation of the device. The procedure for device placement is technically straightforward, and device-related complications are less frequent or significant relative to more invasive artificial sphincter devices discussed above. Further studies of safety and efficacy are awaited.

REFERENCES 1. Agachan, F., Pfeifer, J., and Wexner, S. D.: Defecography and proctography: results of 744 patients. Dis. Colon Rectum 39:899, 1996. 2. Bachoo, P., Brazzelli, M., and Grant, A.: Surgery for faecal incontinence in adults. Cochrane Database Syst. Rev. 4: CD001757, 2000. 3. Bahns, E., Halsband, U., and Janig, W.: Responses of sacral visceral afferents from the lower urinary tract, colon and anus to mechanical stimulation. Pflugers Arch. 410:296, 1987. 4. Barnett, J. L., Hasler, W. L., and Camilleri, M.: American Gastroenterological Association medical position statement on anorectal testing techniques. Gastroenterology 116:732, 1999. 5. Baron, R., Janig, W., and McLachlan, E. M.: The afferent and sympathetic components of the lumbar spinal outflow to the colon and pelvic organs in the cat. III. The colonic nerves, incorporating an analysis of all components of the lumbar prevertebral outflow. J. Comp. Neurol. 238:158, 1985. 6. Bartolo, D. C., Jarratt, J. A., Read, M. G., et al.: The role of partial denervation of the puborectalis in idiopathic faecal incontinence. Br. J. Surg. 70:664, 1983. 7. Bartram, C. I., and Sultan, A. H.: Anal endosonography in faecal incontinence. Gut 37:4, 1995. 8. Bassotti, G., Crowell, M. D., and Whitehead, W. E.: Contractile activity of the human colon: lessons from 24 hour studies. Gut 34:129, 1993. 9. Bennett, R. C., and Duthie, H. L.: The functional importance of the internal sphincter. Br. J. Surg. 51:355, 1964.

294

Function of the Peripheral Nervous System

10. Bharucha, A. E.: Obstructed defecation: don’t strain in vain! [comment]. Am. J. Gastroenterol. 93:1019, 1998. 11. Bharucha, A. E.: Fecal incontinence. Gastroenterology 124:1672, 2003. 12. Biancani, P., Walsh, J., and Behar, J.: Vasoactive intestinal peptide: a neurotransmitter for relaxation of the rabbit internal anal sphincter. Gastroenterology 89:867, 1985. 13. Bishop, B.: Reflex activity of the external anal sphincter of cat. J. Neurophysiol. 22:679, 1959. 14. Bishop, B., Garry, R., Roberts, T., and Todd, J.: Control of the external sphincter of the anus in the cat. J. Physiol. (Lond.) 134:229, 1956. 15. Blumberg, H., Haupt, P., Janig, W., and Kohler, W.: Encoding of visceral noxious stimuli in the discharge patterns of visceral afferent fibres from the colon. Pflugers Arch. 398:33, 1983. 16. Borghi, F., DiMolfetta, L., Garavoglia, M., et al.: Questions about the uncertain presence of muscle spindles in the human external anal sphincter. Panminerva Med. 33:170, 1991. 17. Briel, J. W., Stoker, J., Rociu, E., et al.: External anal sphincter atrophy on endoanal magnetic resonance imaging adversely affects continence after sphincteroplasty. Br. J. Surg. 86:1322, 1999. 18. Briel, J. W., Zimmerman, D. D., Stoker, J., et al.: Relationship between sphincter morphology on endoanal MRI and histopathological aspects of the external anal sphincter. Int. J. Colorectal Dis. 15:87, 2000. 19. Byrne, C. M., Pager, C., Rex, J., et al.: Assessment of quality of life in the treatment of patients with neuropathic fecal incontinence. Dis. Colon Rectum 45:1431, 2002. 20. Camilleri, M., Thompson, W. G., Fleshman, J. W., and Pemberton, J. H.: Clinical management of intractable constipation. Ann. Intern. Med. 121:520, 1994. 21. Carapeti, E. A., Kamm, M. A., and Phillips, R. K.: Randomized controlled trial of topical phenylephrine in the treatment of faecal incontinence. Br. J. Surg. 87:38, 2000. 22. Carlstedt, A., Nordgren, S., Fasth, S., et al.: Sympathetic nervous influence on the internal anal sphincter and rectum in man [comment]. Int. J. Colorectal Dis. 3:90, 1988. 23. Caruana, B. J., Wald, A., Hinds, J. P., and Eidelman, B. H.: Anorectal sensory and motor function in neurogenic fecal incontinence: comparison between multiple sclerosis and diabetes mellitus. Gastroenterology 100:465, 1991. 24. Cervero, F., and Janig, W.: Visceral nociceptors: a new world order? [comment]. Trends Neurosci. 15:374, 1992. 25. Chalmers, D., and Swash, M.: Selective vulnerability of urinary Onuf motoneurons in Shy-Drager syndrome. J. Neurol. 234:259, 1987. 26. Chantraine, A., De Leval, J., and Onkelinx, A.: Motor conduction velocity in the internal pudendal nerves. In Desmedt, J. E. (ed.): New Developments in Electromyography and Clinical Neurophysiology. Basel, Karger, p. 433, 1973. 27. Chapman, A. E., Geerdes, B., Hewett, P., et al.: Systematic review of dynamic graciloplasty in the treatment of faecal incontinence. Br. J. Surg. 89:138, 2002. 28. Cheetham, M. J., Kamm, M. A., and Phillips, R. K.: Topical phenylephrine increases anal canal resting pressure in patients with faecal incontinence. Gut 48:356, 2001. 29. Cheong, D. M., Vaccaro, C. A., Salanga, V. D., et al.: Electrodiagnostic evaluation of fecal incontinence. Muscle Nerve 18:612, 1995.

30. Chia, Y. W., Gill, K. P., Jameson, J. S., et al.: Paradoxical puborectalis contraction is a feature of constipation in patients with multiple sclerosis. J. Neurol. Neurosurg. Psychiatry 60:31, 1996. 31. Cook, T. A., Mortensen, N. J.: Colon, rectum, anus, anal sphincters and the pelvic floor. In Pemberton, J. H., Swash, M., and Henry, M. M. (eds.): The Pelvic Floor: Its Function and Disorders. London, Harcourt Publishers, p. 61, 2002. 32. Corestti, M., Bhoori, S., and Basilisco, G.: Perceptual sensitivity and response bias during rectal distention in patients with irritable bowel syndrome. Neurogastroenterol. Motil. 11:255, 1999. 33. Culver, P. J., and Rattan, S.: Genesis of anal canal pressures in the opossum. Am. J. Physiol. 251(6 Pt. 1):G765, 1986. 34. Daube, J. R.: The description of motor unit potentials in electomyography. Neurology 28:623, 1978. 35. Daube, J. R.: Assessing the motor unit with needle electromyography. In Daube, J. (ed.): Clinical Neurophysiology (Contemporary Neurology Series). Philadelphia, F. A. Davis, p. 257, 1996. 36. Denny-Brown, D., and Robertson, E.: An investigation of the nervous control of defecation. Brain 58:256, 1935. 37. Devroede, G.: Functions of the anorectum: defecation and continence. In Phillips, S., Pemberton, J., and Shorter, R. (eds.): The Large Intestine: Physiology and Disease. New York, Raven Press, p. 115, 1991. 38. Devroede, G., and Lamarche, J.: Functional importance of extrinsic parasympathetic innervation to the distal colon and rectum in man. Gastroenterology 66:273, 1974. 39. Dubrovsky, B., and Filipini, D.: Neurobiological aspects of the pelvic floor muscles involved in defecation. Neurosci. Biobehav. Rev. 14:157, 1990. 40. Duthie, H. L., and Watts, J. M.: Contribution of the external anal sphincter to the pressure zone in the anal canal. Gut 6:64, 1965. 41. Enck, P., Heyer, T., Gantke, B., et al.: How reproducible are measures of the anal sphincter muscle diameter by endoanal ultrasound? Am. J. Gastroenterol. 92:293, 1997. 42. Engel, A. F., and Kamm, M. A.: The acute effect of straining on pelvic floor neurological function. Int. J. Colorectal Dis. 9:8, 1994. 43. Engel, B. T., Nikoomanesh, P., and Schuster, M. M.: Operant conditioning of rectosphincteric responses in the treatment of fecal incontinence. N. Engl. J. Med. 290:646, 1974. 44. Ertekin, C., Hansen, M. V., Larsson, L. E., and Sjodahl, R.: Examination of the descending pathway to the external anal sphincter and pelvic floor muscles by transcranial cortical stimulation. Electroencephalogr. Clin. Neurophysiol. 75:500, 1990. 45. Fedorak, R. N., Field, M., and Chang, E. B.: Treatment of diabetic diarrhea with clonidine. Ann. Intern. Med. 102:197, 1985. 46. Fletcher, J. G., Busse, R. F., Riederer, S. J., et al.: Magnetic resonance imaging of the anatomic and dynamic defects of the pelvic floor in defecatory disorders. Am. J. Gastroenterol. 98:399, 2003. 47. Floyd, W., and Walls, E.: Electromyography of the sphincter ani externus in man. J. Physiol. (Lond.) 122:599, 1953. 48. Fox, J. C., Rath-Harvey, D., Helwig, P. S., et al.: Anal sphincter pressures and rectal compliance decline with aging in

Autonomic and Somatic Systems to the Anorectum and Pelvic Floor

49. 50. 51. 52.

53.

54.

55. 56.

57.

58.

59. 60.

61.

62.

63.

64.

65.

66.

67.

68.

asymptomatic women. Gastroenterology 122(Suppl.): A-69, 2002. Frenckner, B.: Function of the anal sphincters in spinal man. Gut 16:638, 1975. Frenckner, B., and Euler, C. V.: Influence of pudendal block on the function of the anal sphincters. Gut 16:482, 1975. Frenckner, B., and Ihre, T.: Influence of autonomic nerves on the internal and sphincter in man. Gut 17:306, 1976. Friedmann, C. A.: The action of nicotine and catecholamines on the human internal anal sphincter. Am. J. Dig. Dis. 13:428, 1968. Ganio, E., Ratto, C., Masin, A., et al.: Neuromodulation for fecal incontinence: outcome in 16 patients with definitive implant. The initial Italian Sacral Neurostimulation Group (GINS) experience. Dis Colon Rectum 44:965, 2001. Garrett, J. R., Howard, E. R., and Jones, W.: The internal anal sphincter in the cat: a study of nervous mechanisms affecting tone and reflex activity. J. Physiol. (Lond.) 243:153, 1974. Garry, R.: The responses to stimulation of the caudal end of the large bowel in the cat. J. Physiol. (Lond.) 78:208, 1933. Gold, D. M., Halligan, S., Kmiot, W. A., and Bartram, C. I.: Intraobserver and interobserver agreement in anal endosonography. Br. J. Surg. 86:371, 1999. Goligher, J. C., and Hughes, E. S. R.: Sensibility of rectum and colon: its role in the mechanism of anal continence. Lancet 1:543, 1951. Gonella, J., Bouvier, M., and Blanquet, F.: Extrinsic nervous control of motility of small and large intestines and related sphincters. Physiol. Rev. 67:902, 1987. Gowers, W.: The automatic action of the sphincter ani. Proc. R. Soc. Lond. 26:77, 1877. Gwee, K. A., Leong, Y. L., Graham, C., et al.: The role of psychological and biological factors in postinfective gut dysfunction. [comment]. Gut 44:400, 44. Hamdy, S., Enck, P., Aziz, Q., et al.: Spinal and pudendal nerve modulation of human corticoanal motor pathways. Am. J. Physiol. Gastrointest. Liver Physiol. 274(2 Pt. 1): G419, 1998. Herdmann, J., Bielefeldt, K., and Enck, P.: Quantification of motor pathways to the pelvic floor in humans. Am. J. Physiol. Gastrointest. Liver Physiol. 260(5 Pt. 1):G720, 1991. Herdmann, J., Enck, P., Zacchi-Deutschbein, P., and Ostermann, U.: Speed and pressure characteristics of external anal sphincter contractions. Am. J. Physiol. Gastrointest. Liver Physiol. 269(2 Pt. 1):G225, 1995. Heyer, T., Enck, P., Gantke, B., et al.: Anal endosonography: are morphometric measurements of the anal sphincter reproducible? Gastroenterology 108:A613, 1995. Heymen, S., Jones, K. R., Ringel, Y., et al.: Biofeedback treatment of fecal incontinence: a critical review. Dis. Colon Rectum 44:728, Hinds, J. P., Eidelman, B. H., and Wald, A.: Prevalence of bowel dysfunction in multiple sclerosis: a population survey, Gastroenterology 98:1538, 1990. Hinds, J. P., and Wald, A.: Colonic and anorectal dysfunction associated with multiple sclerosis. Am. J. Gastroenterol. 84:587, 1989. Hjartardottir, S., Nilsson, J., Petersen, C., and Lingman, G.: The female pelvic floor: a dome—not a basin. Acta Obstet. Gynecol. Scand. 76:567, 1997.

295

69. Hoffman, S., and Orestano, F.: Histology of the myenteric plexus in relation to rectal biopsy in congenital megacolon. J. Pediatr. Surg. 2:575, 1967. 70. Holstege, G.: Some anatomical observations on the projections from the hypothalamus to brainstem and spinal cord: an HRP and autoradiographic tracing study in the cat. J. Comp. Neurol. 260:98, 1987. 71. Holstege, G., Kuypers, H. G., and Boer, R. C.: Anatomical evidence for direct brain stem projections to the somatic motoneuronal cell groups and autonomic preganglionic cell groups in cat spinal cord. Brain Res. 171:329, 1979. 72. Holstege, G., and Tan, J.: Supraspinal control of motoneurons innervating the striated muscles of the pelvic floor including urethral and anal sphincters in the cat. Brain 110(Pt.5):1323, 1987. 73. Horgan, P. G., O’Connell, P. R., Shinkwin, C. A., and Kirwan, W. O.: Effect of anterior resection on anal sphincter function. Br. J. Surg. 76:783, 1989. 74. Hussain, S. M., Stoker, J., and Lameris, J. S.: Anal sphincter complex: endoanal MR imaging of normal anatomy. Radiology 197:671, 1995. 75. Iglicki, F., Coffin, B., Ille, O., et al.: Fecal incontinence after pelvic radiotherapy: evidences for a lumbosacral plexopathy. Report of a case. Dis. Colon Rectum 39:465, 1996. 76. Jameson, J. S., Chia, Y. W., Kamm, M. A. et al.: Effect of age, sex and parity on anorectal function. Br. J. Surg. 81:1689, 1994. 77. Janig, W., and Koltzenburg, M.: On the function of spinal primary afferent fibres supplying colon and urinary bladder. J. Auton. Nerv. Syst. 30(Suppl.):S89, 1990. 78. Janig, W., and Koltzenburg, M.: Receptive properties of sacral primary afferent neurons supplying the colon. J. Neurophysiol. 65:1067, 1991. 79. Janig, W., and Morrison, J. F. B.: Functional properties of spinal visceral afferents supplying abdominal and pelvic organs, with special emphasis on visceral nociception. Prog. Brain Res. 67:87, 1986. 80. Jankowska, E., Padel, Y., and Zarzecki, P.: Crossed disynaptic inhibition of sacral motoneurones. J. Physiol. (Lond.) 285:425, 1978. 81. Katirji, B., Wilbourn, A. J., Scarberry, S. L., and Preston, D. C.: Intrapartum maternal lumbosacral plexopathy. Muscle Nerve 26:340, 2002. 82. Kiff, E. S., Barnes, P. R., and Swash, M.: Evidence of pudendal neuropathy in patients with perineal descent and chronic straining at stool. Gut 25:1279, 1984. 83. Kihira, T., Yoshida, S., Yoshimasu, F., et al.: Involvement of Onuf’s nucleus in amyotrophic lateral sclerosis. J. Neurol. Sci. 147:81, 1997. 84. Krier, J., and Adams, T.: Autonomic and somatic systems to the anorectum and pelvic floor. In Dyck, P. J., Thomas, P. K., Griffin, J. W., et al. (eds.): Peripheral Neuropathy, 3rd ed. Philadelphia, W. B. Saunders, p. 187, 1993. 85. Krier, J., Adams, T., and Meyer, R. A.: Physiological, morphological, and histochemical properties of cat external anal sphincter. Am. J. Physiol. Gastrointest. Liver Physiol. 255 (6 Pt. 1):G772, 1988. 86. Krier, J., Meyer, R. A., and Percy, W. H.: Length-tension relationship of striated muscle of cat external anal sphincter. Am. J. Physiol. Gastrointest. Liver Physiol. 256 (4 Pt. 1):G773, 1989.

296

Function of the Peripheral Nervous System

87. Langley, J. N., and Anderson, H. K.: On the innervation of the pelvic and adjoining viscera. Part 1. The lower portion of the intestine. J. Physiol. (Lond) 18:67, 1895. 88. Law, N.-M., Bharucha, A. E., Undale, A. S., and Zinsmeister, A. R.: Cholinergic stimulation enhances colonic motor activity, transit and sensation in humans. Am. J. Physiol. 281:G1228, 2001. 89. Leigh, R. J., and Turnberg, L. A.: Faecal incontinence: the unvoiced symptom. Lancet 1:1349, 1982. 90. Locke, G. R. 3rd, Pemberton, J. H., and Phillips, S. F.: AGA technical review on constipation. American Gastroenterological Association. Gastroenterology 119: 1766, 2000. 91. Locke, G. R. 3rd, Pemberton, J. H., and Phillips, S. F.: American Gastroenterological Association Medical Position Statement: guidelines on constipation. Gastroenterology 119:1761, 2000. 92. Loening-Baucke, V., Read, N. W., Yamada, T., and Barker, A. T.: Evaluation of the motor and sensory components of the pudendal nerve. Electroencephalogr. Clin. Neurophysiol. 93:35, 1994. 93. Lubowski, D. Z., Nicholls, R. J., Swash, M., and Jordan, M. J.: Neural control of internal anal sphincter function. Br. J. Surg. 74:668, 94. Mackel, R.: Segmental and descending control of the external urethral and anal sphincters in the cat. J. Physiol. (Lond.) 294: 105, 1979. 95. Malouf, A. J., Halligan, S., Williams, A. B., et al.: Prospective assessment of interobserver agreement for endoanal MRI in fecal incontinence. Abdom. Imaging 26:76, 2001. 96. Malouf, A. J., Norton, C. S., Engel, A. F., et al.: Long-term results of overlapping anterior anal-sphincter repair for obstetric trauma. Lancet 355:260, 2000. 97. Mannen, T., Iwata, M., Toyokura, Y., and Nagashima, K.: The Onuf’s nucleus and the external anal sphincter muscles in amyotrophic lateral sclerosis and Shy-Drager syndrome. Acta Neuropathol. 58:255, 1982. 98. Marsden, C. D., Merton, P. A., and Morton, H. B.: The latency of the anal reflex. J. Neurol. Neurosurg. Psychiatry 45:857, 1982. 99. Mathers, S. E., Kempster, P. A., Law, P. J., et al.: Anal sphincter dysfunction in Parkinson’s disease. Arch. Neurol. 46:1061, 1989. 100. Mayer, E. A., and Gebhart, G. F.: Basic and clinical aspects of visceral hyperalgesia [comment]. Gastroenterology 107:271, 1994. 101. McHugh, S. M., and Diamant, N. E.: Effect of age, gender, and parity on anal canal pressures: contribution of impaired anal sphincter function to fecal incontinence. Dig. Dis. Sci. 32:726, 1987. 102. Melzak, J., and Porter, N. H.: Studies of the reflex activity of the external sphincter ani in spinal man. Paraplegia 1:277, 103. Miller, R., Bartolo, D. C., Cervero, F., and Mortensen, N. J.: Anorectal sampling: a comparison of normal and incontinent patients. Br. J. Surg. 75:44, 1988. 104. Milligan, E. T. C., and Morgan, C. N.: Surgical anatomy of the anal canal with special reference to anorectal fistulae. Lancet 2:1150, 1934.

105. Miner, P. B., Donnelly, T. C., and Read, N. W.: Investigation of mode of action of biofeedback in treatment of fecal incontinence. Dig. Dis. Sci. 35:1291, 1990. 106. Mizutani, M. Neya, T., Ono, K., et al.: Histochemical study of the lumbar colonic nerve supply to the internal anal sphincter and its physiological role in dogs. Brain Res. 598: 45, 1992. 107. Morera, C., and Nurko, S.: Rectal manometry in patients with isolated sacral agenesis. J. Pediatr. Gastroenterol. Nutr. 37:47, 2003. 108. Morgan, C., Nadelhaft, I., and de Groat, W. C.: The distribution of visceral primary afferents from the pelvic nerve to Lissauer’s tract and the spinal gray matter and its relationship to the sacral parasympathetic nucleus. J. Comp. Neurol. 201: 415, 1981. 109. Nakanishi, N., Tatara, K., Naramura, H., et al.: Urinary and fecal incontinence in a community-residing older population in Japan. J. Am. Geriatr. Soc. 45:215, 1997. 110. Nakayama, S., Neya, T., Yamasato, T., et al.: Activity of the spinal defaecation centre in the guinea pig. Ital. J. Gastroenterol. 11:168, 1979. 111. Nelson, R., Furner, S., and Jesudason, V.: Fecal incontinence in Wisconsin nursing homes: prevalence and associations. Dis. Colon Rectum 41:1226, 1998. 112. Ness, T. J., and Gebhart, G. F.: Colorectal distension as a noxious visceral stimulus: physiologic and pharmacologic characterization of pseudaffective reflexes in the rat. Brain Res. 450:153, 1988. 113. Ness, T. J., and Gebhart, G. F.: Visceral pain: a review of experimental studies. Pain 41:167, 1990. 114. Nielsen, M. B., Hauge, C., Pedersen, J. F., and Christiansen, J.: Endosonographic evaluation of patients with anal incontinence: findings and influence on surgical management. AJR Am. J. Roentgenol. 160:771, 1993. 115. Norton, C., Hosker, G., and Brazzelli, M.: Biofeedback and/or sphincter exercises for the treatment of faecal incontinence in adults. Cochrane Database Syst. Rev. 4: CD002111, 2003. 116. Nurko, S., and Rattan, S.: Role of vasoactive intestinal polypeptide in the internal anal sphincter relaxation of the opossum. J. Clin. Invest. 81:1146, 1988. 117. Opsomer, R. J., Caramia, M. D., Zarola, F., et al.: Neurophysiological evaluation of central-peripheral sensory and motor pudendal fibres. Electroencephalogr. Clin. Neurophysiol. 74:260, 1989. 118. Palmer, K. R., Corbett, C. L., and Holdsworth, C. D.: Double-blind cross-over study comparing loperamide, codeine and diphenoxylate in the treatment of chronic diarrhea. Gastroenterology 79:1272, 1980. 119. Parks, A. G., Fishlock, D. J., Cameron, J. D., and May, H.: Preliminary investigation of the pharmacology of the human internal anal sphincter. Gut 10:674, 1969. 120. Parks, A. G., Swash, M., and Urich, H.: Sphincter denervation in anorectal incontinence and rectal prolapse. Gut 18:656, 1977. 121. Penninckx, F., Lestar, B., and Kerremans, R.: The internal anal sphincter: mechanisms of control and its role in maintaining anal continence. Baillieres Clin. Gastroenterol. 6: 193, 1992. 122. Percy, J. P., Neill, M. E., Swash, M., and Parks, A. G.: Electrophysiological study of motor nerve supply of pelvic floor. Lancet 1:16, 1981.

Autonomic and Somatic Systems to the Anorectum and Pelvic Floor 123. Pezim, M. E., Pemberton, J. H., Levin, K. E., et al.: Parameters of anorectal and colonic motility in health and in severe constipation. Dis. Colon Rectum 36:484, 1993. 124. Podnar, S., Mrkaic, M., and Vodusek, D. B.: Standardization of anal sphincter electromyography: quantification of continuous activity during relaxation. Neurourol. Urodyn. 21: 540, 2002. 125. Rao, S. S., Read, N. W., Davison, P. A., et al.: Anorectal sensitivity and responses to rectal distention in patients with ulcerative colitis. Gastroenterology 93:1270, 1987. 126. Rao, S. S., Welcher, K. D., and Leistikow, J. S.: Obstructive defecation: a failure of rectoanal coordination. [comment]. Am. J. Gastroenterol. 93:1042, 1998. 127. Rayner, V.: Observations on the functional internal anal sphincter of the vervet monkey. J. Physiol. (Lond.) 213:27P, 1971. 128. Read, M., Read, N. W., Barber, D. C., and Duthie, H. L.: Effects of loperamide on anal sphincter function in patients complaining of chronic diarrhea with fecal incontinence and urgency. Dig. Dis. Sci. 27:807, 1982. 129. Rentsch, M., Paetzel, C., Lenhart, M., et al.: Dynamic magnetic resonance imaging defecography: a diagnostic alternative in the assessment of pelvic floor disorders in proctology. Dis. Colon Rectum 44:999, 2001. 130. Rociu, E., Stoker, J., Eijkemans, M. J., et al.: Fecal incontinence: endoanal US versus endoanal MR imaging. Radiology 212:453, 1999. 131. Rockwood, T. H., Church, J. M., Fleshman, J. W., et al.: Patient and surgeon ranking of the severity of symptoms associated with fecal incontinence: the fecal incontinence severity index. Dis. Colon Rectum 42:1525, 1999. 132. Rockwood, T. H., Church, J. M., Fleshman, J. W., et al.: Fecal incontinence quality of life scale: quality of life instrument for patients with fecal incontinence. Dis. Colon Rectum 43:9; discussion 16, 2000. 133. Rogers, J., Levy, D. M., Henry, M. M., and Misiewicz, J. J.: Pelvic floor neuropathy: a comparative study of diabetes mellitus and idiopathic faecal incontinence. Gut 29:756, 1988. 134. Romero, Y., Evans, J. M., Fleming, K. C., and Phillips, S. F.: Constipation and fecal incontinence in the elderly population. Mayo Clin. Proc. 71:81, 1996. 135. Roppolo, J. R., Nadelhaft, I., and de Groat, W. C.: The organization of pudendal motoneurons and primary afferent projections in the spinal cord of the rhesus monkey revealed by horseradish peroxidase. J. Comp. Neurol. 234:475, 1985. 136. Rosen, H. R., Urbarz, C., Holzer, B., et al.: Sacral nerve stimulation as a treatment for fecal incontinence. Gastroenterology 121:536, 2001. 137. Ruch, T. C.: Pathophysiology of pain. In Ruch, T. C., and Patton, H. D. (eds.): Physiology and Biophysics: I. The Brain and Neural Function. Philadelphia, W. B. Saunders, p. 272, 1979. 138. Salducci, J., Planche, D., and Naudy, B.: Physiological role of the internal anal sphincter and the external anal sphincter during micturition. In Wienbeck, M. (ed.): Motility of the Digestive Tract. New York, Raven Press, p. 513, 1982. 139. Salvioli, B., Bharucha, A. E., Rath-Harvey, D., et al.: Rectal compliance, capacity and rectoanal sensation in fecal incontinence. Am. J. Gastroenterol. 96:2158, 2001.

297

140. Sato, M., Mizuno, N., and Konishi, A.: Localization of motoneurons innervating perineal muscles: a HRP study in cat. Brain Res. 140: 1978. 141. Schroder, H. D.: Onuf’s nucleus X: a morphological study of a human spinal nucleus. Anat. Embyrol. 162:443, 1981. 142. Schroder, H. D., and Reske-Nielsen, E.: Fiber types in the striated urethral and anal sphincters. Acta Neuropathol. 60: 278, 1983. 143. Schuster, M. M.: Motor action of rectum and anal sphincters in continence and defecation. In Code, C. F. (ed.): Handbook of Physiology: Section 6. Alimentary Canal. Washington, DC, American Physiol Society, p. 2121, 1968. 144. Schweiger, M.: Method for determining individual contributions of voluntary and involuntary anal sphincters to resting tone. Dis. Colon Rectum 22:415, 1979. 145. Scott, H. W. J., and Cantrell, J. R.: Colonmetrographic studies of the effects of section of the parasympathetic nerves of the colon. Bull. Johns Hopkins Hosp. 85:310, 1949. 146. Semba, T., Mishima, H., and Date, T.: Studies on a vesico-anal inhibitory reflex. Jpn. J. Physiol. 6:108, 1956. 147. Sengupta, J. N., Saha, J. K., and Goyal, R. K.: Stimulusresponse function studies of esophageal mechanosensitive nociceptors in sympathetic afferents of opossum. J. Neurophysiol. 64:796, 1990. 148. Shafik, A.: Straining puborectalis reflex: description and significance of a “new” reflex. Anat. Rec. 229:281, 1991. 149. Shafik, A.: A new concept of the anatomy of the anal sphincter mechanism and the physiology of defecation: mass contraction of the pelvic floor muscles. Int. Urogynecol. J. 9:28, 1998. 150. Sherrington, C. S.: Notes on the arrangement of some motor fibres in the lumbosacral plexus. J. Physiol. (Lond.) 13:672, 1892. 151. Shirran, E., and Brazzelli, M.: Absorbent products for containing urinary and/or faecal incontinence in adults. Cochrane Database Syst. Rev. 4: CD001406, 2003. 152. Siproudhis, L., Bellissant, E., Juguet, F., et al.: Perception of and adaptation to rectal isobaric distension in patients with faecal incontinence. Gut 44:687, 1999. 153. Stebbing, J. F., Brading, A. F., and Mortensen, N. J.: Nitric oxide and the rectoanal inhibitory reflex: retrograde neuronal tracing reveals a descending nitrergic rectoanal pathway in a guinea-pig model. Br. J. Surg. 83:493, 1996. 154. Sultan, A. H., Kamm, M. A., Hudson, C. N., et al.: Anal-sphincter disruption during vaginal delivery [see comments]. N. Engl. J. Med. 329:1905, 1993. 155. Sun, W. M., Read, N. W., and Donnelly, T. C.: Anorectal function in incontinent patients with cerebrospinal disease. Gastroenterology 99:1372, 1990. 156. Sun, W. M., Read, N. W., and Miner, P. B.: Relation between rectal sensation and anal function in normal subjects and patients with faecal incontinence. Gut 31: 1056, 1990. 157. Sung, J. H.: Autonomic neurons affected by lipid storage in the spinal cord in Fabry’s disease: distribution of autonomic neurons in the sacral cord. J. Neuropathol. Exp. Neurol. 38: 87, 1979.

298

Function of the Peripheral Nervous System

158. Sung, J. H. and Mastri, A. R.: Spinal autonomic neurons in Werdnig-Hoffmann disease, mannosidosis, and Hurler’s syndrome: distribution of autonomic neurons in the sacral spinal cord. J. Neuropathol. Exp. Neurol. 39: 441, 1980. 159. Sung, J. H., Mastri, A. R., and Segal, E.: Pathology of Shy-Drager syndrome. J. Neuropathol. Exp. Neurol. 38: 353, 1979. 160. Swash, M.: Early and late components in the human anal reflex. J. Neurol. Neurosurg. Psychiatry 45:767, 1982. 161. Taverner, D., and Smiddy, F. G.: An electromyographic study of the normal function of the external anal sphincter and pelvic diaphragm. Dis. Colon Rectum 2:153, 1959. 162. Turnbull, G. K., Hamdy, S., Aziz, Q., et al.: The cortical topography of human anorectal musculature. Gastroenterology 117:32, 1999. 163. Uher, E. M., and Swash, M.: Sacral reflexes: physiology and clinical application. Dis. Colon Rectum 41:1165, 1998. 164. Vaizey, C. J., Kamm, M. A., and Bartram, C. I.: Primary degeneration of the internal anal sphincter as a cause of passive faecal incontinence. Lancet 349:612, 1997. 165. Vaizey, C. J., Kamm, M. A., Turner, I. C., et al.; Effects of short term sacral nerve stimulation on anal and rectal function in patients with anal incontinence. Gut 44:407, 1999.

166. Voderholzer, W. A., Neuhaus, D. A., Klauser, A. G., et al.: Paradoxical sphincter contraction is rarely indicative of anismus. Gut 41:258, 1997. 167. Wald, A., and Tunuguntla, A. K.: Anorectal sensorimotor dysfunction in fecal incontinence and diabetes mellitus: modification with biofeedback therapy. N. Engl. J. Med. 310:1282, 1984. 168. Weinberg, A. G.: The anorectal myenteric plexus: its relation to hypoganglionosis of the colon. Am. J. Clin. Pathol. 54:637, 1970. 169. Wexner, S. D., Marchetti, F., and Salanga, V. D. et al.: Neurophysiologic assessment of the anal sphincters. Dis. Colon Rectum 34:606, 1991. 170. Wheatley, I. C., Hardy, K. J., and Dent, J.: Anal pressure studies in spinal patients. Gut 18:488, 1977. 171. White, J. C., and Sweet, W. H.: Pain and the Neurosurgeon: A Forty Year Experience. Springfield, IL, Charles C Thomas, 1969. 172. Whitehead, W. E., and Palsson, O. S.: Is rectal pain sensitivity a biological marker for irritable bowel syndrome? Psychological influences on pain perception. Gastroenterology 115:1263, 1998. 173. Whitehead, W. E., and Wald, A.: Functional disorders of the anus and rectum. Gut 45 (Suppl. II): II55, 1999.

14 Autonomic Systems to the Urinary Bladder and Sexual Organs WILLIAM C. DE GROAT AND AUGUST M. BOOTH

Lower Urinary Tract Anatomy Innervation Parasympathetic Pathways Sympathetic Pathways Neural Modulatory Mechanisms Somatic Pathways Afferent Pathways Urothelial-Afferent Interactions Reflex Mechanisms Controlling the Lower Urinary Tract Anatomy of Central Nervous Pathways Controlling the Lower Urinary Tract Pathways in the Spinal Cord

Pathways in the Brain Organization of Urine Storage and Voiding Reflexes Sympathetic Pathways Somatic Pathways to the Urethral Sphincter Voiding Reflexes Neurotransmitters in Micturition Reflex Pathways Excitatory Neurotransmitters Inhibitory Neurotransmitters Neurogenic Dysfunction of the Lower Urinary Tract Spinal Injury

The urogenital system is affected by a variety of neuropathologic conditions and often is the site of the first symptoms produced by diffuse neurologic diseases, such as diabetic neuropathy or multiple sclerosis.12,13,41,62,82,98,111,112,137 This sensitivity to neurogenic disorders is not surprising in view of the complexity and importance of neural mechanisms in the regulation of urogenital function.33,45,48,98,111,131 For example, the activity of the lower urinary tract and certain sex organs is almost totally dependent on central nervous control, whereas many other visceral systems, in particular the gastrointestinal tract and cardiovascular system, continue to function in the absence of extrinsic neural input. In addition, many responses of the urogenital system require the coordination of autonomic and somatic efferent mechanisms at various levels of the lumbosacral spinal cord. These spinal mechanisms are in turn modulated by inputs from higher centers in the brain and in some instances are subject to voluntary control. This chapter reviews the organization of the neural pathways regulating the urogenital system with

Sex Organs Innervation Parasympathetic Pathways Sympathetic Pathways Somatic Pathways Erection Peripheral Mechanisms Central Mechanisms Glandular Secretion Emission-Ejaculation Central Reflex Pathways

the aim of providing the anatomic and physiologic background for understanding urinary and sexual dysfunction resulting from neurologic disorders.

Lower Urinary Tract ANATOMY The storage and periodic elimination of urine is controlled by the activity of two functional units in the lower urinary tract: (1) a reservoir (the bladder) and (2) an outlet (consisting of the bladder neck, urethra, and striated muscles of the pelvic floor). Two regions of the bladder have been distinguished based on embryonic origin. The detrusor, which comprises the major portion of the bladder, is derived from endoderm, whereas the trigone, a small region located posteriorly at the neck of the bladder, is derived from mesoderm and represents the extension of 299

300

Function of the Peripheral Nervous System

the longitudinal muscle layers of the ureters. A more physiologic designation of regions of the bladder has been proposed based on neural innervation and responses to pharmacologic agents.56,98,131 According to this approach, the bladder can be divided circumferentially at the level of the ureteral orifices into cranial and caudal segments termed the body and base, respectively. The trigone occupies the posterior area of the base. The distinguishing characteristics of these two regions are discussed later in this chapter. The luminal surface of the bladder is covered by a multilayered urothelium that functions as a highly efficient barrier to movement of water and ionized substances across the bladder wall.33,56,79,81 The primary urineplasma barrier is attributed to the layer of superficial umbrella cells, which are interconnected by tight junctions and have an unusual asymmetrical unit membrane on their apical surface that consists of an outer leaflet of plaques composed of proteins called uroplakins. Umbrella cells have large numbers of discoid vesicles with asymmetrical membrane structure located just under the apical surface. It is believed that trafficking of these vesicles to the apical surface during stretch allows the umbrella cells to increase their surface area during bladder distention.134 Urothelial cells also have neuronallike properties, including the ability to release neurotransmitters (ATP and nitric oxide [NO]) in response to stretch or chemical stimulation.14,15,18,60,136 Transmitter release may be mediated by exocytosis associated with vesicle trafficking. Transmitters released from the urothelial cells may act in an autocrine/paracrine manner within the urothelium or on subepithelial myofibroblasts, nerves, or blood vessels to influence various functions, including the urothelial barrier, local blood flow, and sensory mechanisms.15,40,58,60,127 According to the traditional view, the bladder smooth muscle is arranged in three layers (inner and outer longitudinal and middle circular), although there is considerable interdigitation between these layers, giving the bladder wall a meshlike appearance.33,56,131 The inner longitudinal layer is continuous with the longitudinal smooth muscle of the urethra, which constitutes the greater portion of the urethral wall. The urethra has a very thin layer of circular muscle but is surrounded by striated muscle that in males is very prominent and extends from the urogenital diaphragm to the neck of the bladder. An anatomic sphincter between the bladder and the urethra has not been identified; however, radiographic studies and measurements of urethral pressure indicate the existence of a physiologic internal sphincter that maintains urinary continence by closure of the bladder neck and proximal urethra.33 The sphincter-like properties of this region have been attributed to the abundance of elastic tissue in the submucosa and the tone of the urethral smooth muscle. Continence is thought to be dependent on

a combination of factors, including urethral wall tension, the caliber of the urethral lumen, and the functional length of the urethra. The striated muscles surrounding the urethra are not essential for urinary continence, but are important in the voluntary termination of urine flow (Fig. 14–1) and in the prevention of stress incontinence.33,129

FIGURE 14–1 Combined cystometrograms and sphincter electromyograms (EMGs) comparing reflex voiding responses in an infant (A) and in a paraplegic patient (C) with a voluntary voiding response in an adult (B). The abscissa in all records represents bladder volume in milliliters and the ordinates represent bladder pressure in centimeters of H2O and electrical activity of the EMG recording. On the left side of each trace, the arrows indicate the start of a slow infusion of fluid into the bladder (bladder filling). Vertical dashed lines indicate the start of sphincter relaxation, which precedes by a few seconds the bladder contraction in A and B. B, Note that a voluntary cessation of voiding (stop) is associated with an initial increase in sphincter EMG followed by a reciprocal relaxation of the bladder. A resumption of voiding is again associated with sphincter relaxation and a delayed increase in bladder pressure. C, Conversely, in the paraplegic patient the reciprocal relationship between bladder and sphincter is abolished. During bladder filling, transient uninhibited bladder contractions occur in association with sphincter activity. Further filling leads to more prolonged and simultaneous contractions of the bladder and sphincter (bladder-sphincter dyssynergia). Loss of the reciprocal relationship between bladder and sphincter in paraplegic patients interferes with bladder emptying.

Autonomic Systems to the Urinary Bladder and Sexual Organs

Under normal conditions, the urinary bladder and outlet exhibit a reciprocal relationship in effecting the storage and elimination of urine (see Fig. 14–1). During storage, the bladder neck and proximal urethra are closed, the sphincter electromyogram (EMG) gradually increases during bladder filling, and maximal intraurethral pressures range from 20 to 50 cm H2O. The detrusor muscle, in contrast, is quiescent, allowing intravesical pressure to remain low (5 to 15 cm H2O) over a wide range of bladder volumes (100 to 400 mL). During voluntary micturition the initial event is a reduction of intraurethral pressure, which reflects a relaxation of the pelvic floor and the paraurethral striated muscles (see Fig. 14–1).129 This coincides with a mechanical shortening of the urethra and the appearance of a cone-shaped opening of the bladder neck. These changes in the urethra are followed in a few seconds by a detrusor contraction and a rise in intravesical pressure that is maintained until the bladder empties. The detrusor contraction produces a further opening of the proximal urethra by shortening of the inner longitudinal muscle bundles, which extend from the bladder into the urethra, and by contractions of outer longitudinal muscle bundles, which loop around the bladder neck. Reflex inhibition of the smooth and striated muscles of the urethra also contributes to the reduction of outlet resistance during micturition. These changes are coordinated by neural pathways at the thoracolumbar and sacral levels of the spinal cord.

INNERVATION The innervation of the lower urinary tract is derived from three sets of peripheral nerves: sacral parasympathetic (pelvic nerves), thoracolumbar sympathetic (hypogastric nerves and sympathetic chain), and sacral somatic nerves (primarily the pudendal nerves) (Fig. 14–2).48,131

Parasympathetic Pathways The sacral parasympathetic outflow, which in humans originates from the S2 to S4 segments of the spinal cord, provides the major excitatory input to the bladder.33,56 Cholinergic preganglionic neurons (PGNs) located in the intermediolateral region of the sacral spinal cord (Fig. 14–3) send axons to cholinergic ganglion cells in the pelvic plexus and in the bladder wall.122 Transmission in bladder ganglia is mediated by a nicotinic cholinergic mechanism, which is sensitive to modulation by various transmitter systems, including muscarinic, adrenergic, purinergic, and peptidergic (Table 14–1).46,50,53,75 The ganglion cells in turn excite the bladder smooth muscle. Histochemical studies of the ganglia and nerves supplying the human lower urinary tract have shown that a large proportion of ganglion cells contain acetylcholinesterase (AChE) as well as vesicular acetyl-

301

choline transporter (VAChT) and therefore are presumably cholinergic. AChE- and VAChT-positive nerves are abundant in all parts of the bladder but are less extensive in the urethra.46,56,98 Neuropeptide Y (NPY) and nitric oxide synthase (NOS) have also been identified in a large percentage (40% to 95%) of intramural ganglia of the human bladder. Several populations of axonal varicosities have been detected in close proximity to intramural ganglion cells, including (1) substance P and calcitonin gene–related peptide (CGRP) positive axons, which are presumably collaterals of extrinsic sensory nerves; (2) tyrosine hydroxylase and NPY axons, which are likely to be sympathetic axons; and (3) vasoactive intestinal polypeptide (VIP), galanin, and NPYcontaining axons, which are presumably preganglionic nerve terminals.56 Parasympathetic neuroeffector transmission in the bladder is mediated by acetylcholine (ACh) acting on postjunctional muscarinic receptors.98 Both M2 and M3 muscarinic receptor subtypes are expressed by bladder smooth muscle; however, examination of subtype-selective muscarinic receptor antagonists and studies of muscarinic receptor knockout mice have revealed that the M3 subtype is the principal receptor involved in excitatory transmission.88,89 In bladders of various animals, stimulation of parasympathetic nerves also produces a noncholinergic contraction that is resistant to atropine and other muscarinic receptor– blocking agents. Activation of M3 receptors triggers intracellular Ca2⫹ release, whereas activation of M2 receptors inhibits adenylate cyclase. The latter may contribute to bladder contractions by suppressing adrenergic inhibitory mechanisms, which are mediated by ␤3-adrenergic receptors and stimulation of adenylate cyclase. ATP (see Table 14–1) has been identified as the excitatory transmitter mediating the noncholinergic contractions.32,106 ATP excites the bladder smooth muscle by acting on P2X purinergic receptors, which are ligand-gated ion channels. Among the seven types of P2X receptors that have been identified in the rat bladder, P2X1 is the major subtype expressed in the rat and also in the human bladder smooth muscle. Although purinergic excitatory transmission is not important in the normal human bladder, it appears to be involved in bladders from patients with pathologic conditions such as chronic urethral outlet obstruction or interstitial cystitis.32,80,98,103,104 Parasympathetic pathways to the urethra induce relaxation during voiding. In various species the relaxation is not affected by muscarinic antagonists and therefore is not mediated by ACh. However inhibitors of NOS block the relaxation in vivo during reflex voiding or block the relaxation of urethral smooth muscle strips induced in vitro by electrical stimulation of intramural nerves, indicating that NO is the inhibitory transmitter involved in relaxation.28,68,98 In some species neurally evoked contractions of the urethra are reduced by muscarinic receptor antagonists or by desensitization of P2X receptors, indicating that

302

Function of the Peripheral Nervous System

FIGURE 14–2 The innervation of the lower urinary tract and male genitalia. (Prepared by Tim Whitney, M.D.)

ACh or ATP is involved in excitatory transmission to urethral smooth muscle.147

Sympathetic Pathways Sympathetic preganglionic pathways that arise from the T11 to L2 spinal segments pass to the sympathetic chain ganglia and then to prevertebral ganglia in the superior hypogastric and pelvic plexuses (see Fig. 14–2), and also

to short adrenergic neurons in the bladder and urethra.48,56 Sympathetic postganglionic nerves that release norepinephrine provide an excitatory input to smooth muscle of the urethra and bladder base, an inhibitory input to smooth muscle in the body of the bladder, and inhibitory and facilitatory input to vesical parasympathetic ganglia.46,56 Histofluorescence microscopy in animals and humans has shown that the smooth muscle of the bladder base is richly innervated by adrenergic terminals, but the

303

Autonomic Systems to the Urinary Bladder and Sexual Organs

FIGURE 14–3 Transneuronal virus tracing of the central pathways controlling the urinary bladder of the rat. Injection of pseudorabies virus into the wall of the urinary bladder leads to retrograde transport of virus (dashed arrows) and sequential infection of postganglionic neurons, preganglionic neurons, and then various central neural circuits synaptically linked to the preganglionic neurons. Normal synaptic connections are indicated by solid arrows. At long survival times virus can be detected with immunocytochemical techniques in neurons at specific sites throughout the spinal cord and brain, extending to the pontine micturition center in the pons (i.e., Barrington’s nucleus) and to the cerebral cortex. Other sites in the brain labeled by virus are (1) the paraventricular nucleus (PVN), medial preoptic area (MPOA), and periventricular nucleus (Peri V.N.) of the hypothalamus; (2) periaqueductal gray (PAG); (3) locus ceruleus (LC) and subceruleus; (4) red nucleus; (5) medullary raphe nuclei; and (6) the noradrenergic cell group designated A5. An L6 spinal cord section shows, on the left side, the distribution of virus-labeled parasympathetic preganglionic neurons (ⵧ) and interneurons (䊉) in the region of the parasympathetic nucleus, the dorsal commissure (DCM), and the superficial laminae of the dorsal horn (DH) 72 hours after injection of the virus into the bladder. The right side shows the entire population of preganglionic neurons (PGN) (ⵧ) labeled by axonal tracing with fluorescent dye (fluorogold) injected into the pelvic ganglia, and the distribution of virus-labeled bladder PGN (䊏). Composite diagram of neurons in 12 spinal sections (42 ␮m).

Table 14–1. Reflexes to the Lower Urinary Tract Afferent Pathway

Efferent Pathway

Central Pathway

Urine Storage Low-level vesical afferent activity (pelvic nerve)

1. External sphincter contraction (somatic nerves) 2. Internal sphincter contraction (sympathetic nerves) 3. Detrusor inhibition (sympathetic nerves) 4. Ganglionic inhibition (sympathetic nerves) 5. Sacral parasympathetic outflow inactive 1. Inhibition of external sphincter activity 2. Inhibition of sympathetic outflow 3. Activation of parasympathetic outflow to the bladder 4. Activation of parasympathetic outflow to the urethra

Spinal reflexes

Micturition High-level vesical afferent activity (pelvic nerve)

Spinobulbospinal reflexes

Spinal reflex

304

Function of the Peripheral Nervous System

bladder body has a considerably weaker adrenergic innervation.98,131 Vesical parasympathetic ganglion cells and the smooth muscle of the proximal urethra also receive an extensive adrenergic innervation.46,48 Radioligand receptor-binding studies showed that ␣-adrenergic receptors are concentrated in the bladder base and proximal urethra, whereas ␤-adrenergic receptors are most prominent in the bladder body.1 These observations are consistent with pharmacologic studies showing that sympathetic nerve stimulation or exogenous catecholamines produce ␤-adrenergic receptor–mediated inhibition of the body and ␣-adrenoceptor–mediated contraction of the base, dome, and urethra. Molecular and contractility studies have shown that ␤3-adrenergic receptors elicit inhibition and ␣1-adrenergic receptors elicit contractions. The ␣1Aadrenergic receptor subtype is most prominent in normal bladders and the ␣1D subtype is upregulated in bladders from patients with outlet obstruction, raising the possibility that ␣1-adrenergic receptor excitatory mechanisms in the bladder might contribute to irritative lower urinary tract symptoms in patients with benign prostatic hyperplasia.55,98

Neural Modulatory Mechanisms Presynaptic modulatory mechanisms and synaptic communication between parasympathetic and sympathetic pathways to the bladder has been demonstrated in bladder ganglia and at postganglionic nerve terminals. In the cat, stimulation of the hypogastric (sympathetic) nerves elicits an initial inhibitory and a delayed facilitatory modulation of cholinergic transmission in parasympathetic bladder ganglia mediated by ␣2- and ␣1-adrenergic receptors, respectively,46,53,75 indicating that sympathetic pathways can influence neural input to the bladder as well as directly affect the bladder smooth muscle (see Table 14–1). Transmission in cat bladder ganglia is also modulated by enkephalins released as cotransmitters along with ACh from preganglionic nerve terminals.49,73 The inhibitory effect of enkephalins can be blocked by the opioid antagonist naloxone and occurs by a presynaptic inhibitory mechanism. Adenosine, presumably derived from ATP released in the bladder ganglia, also exerts an inhibitory action on nicotinic cholinergic transmission in cat bladder ganglia.46 Adrenergic and enkephalinergic inhibitory mechanisms in cat bladder ganglia are frequency dependent, being prominent at low frequencies (0.25 to 0.5 Hz) and markedly reduced at higher frequencies (1 to 10 Hz). This phenomenon is presumably related to the marked temporal facilitation that occurs in cat bladder ganglia at frequencies of preganglionic nerve stimulation above 0.5 Hz.23,54 It has been speculated that cat bladder ganglia function as high-pass filters, eliminating parasympathetic input to the bladder when preganglionic activity is low during urine storage, but significantly amplifying the

input to the bladder when firing is increased during voiding.46 Thus the ganglion acts like a gating circuit. Frequency-dependent adrenergic and enkephalinergic inhibitory mechanisms complement this gating function by effectively inhibiting transmission during urine storage but turning off during voiding to allow complete bladder emptying. In the rat, transmission at synapses in the major pelvic ganglion occurs with a high safety factor, exhibits very little frequency-dependent modulation, and is relatively resistant to the actions of inhibitory transmitters.46 However, the autonomic pathways to the rat urinary bladder are susceptible to modulation at postganglionic sites.46,118–121,132,147 For example, NPY, which is a cotransmitter in cholinergic and adrenergic nerves in the rat, can act prejunctionally to suppress the release of both ACh and norepinephrine in the bladder and urethra.132,147 The NPY inhibition is also frequency dependent, being most prominent at low frequencies of nerve stimulation. In addition, in the rat urinary bladder ACh released from parasympathetic nerves can induce a heterosynaptic facilitation of norepinephrine release from sympathetic nerves by acting on prejunctional M1 muscarinic receptors on the adrenergic terminals. Prejunctional M1 facilitatory and M2/4 inhibitory muscarinic modulatory mechanisms have also been identified on parasympathetic nerve terminals in the rat bladder.119–121 These receptors mediate positive and negative cholinergic feedback mechanisms, respectively, and regulate the release of ACh. Inhibitory M2/4 mechanisms are dominant at low frequencies of nerve activity and therefore could contribute to urine storage, whereas M1 facilitatory mechanisms are dominant at high frequencies of nerve stimulation and could contribute to an enhancement of neurally evoked bladder contractions during micturition to induce complete bladder emptying.

Somatic Pathways The external urethral sphincter (EUS), which is composed of striated muscle, receives a somatic innervation via the pudendal nerve from anterior horn cells in the third and fourth sacral segments (see Fig. 14–2). Branches of the pudendal nerve and other sacral somatic nerves also carry efferent impulses to muscles of the pelvic floor and proprioceptive afferent signals from these muscles, as well as sensory information from the urethra. Analysis of urethral closure mechanisms in the female rat during sneezeinduced stress conditions revealed that the major rise in urethral pressure occurred in the mid-urethra and was mediated by efferent pathways in the pudendal nerve to the EUS as well as pathways in nerves to the iliococcygeus and pubococcygeus muscles, but not by pathways in the sympathetic or parasympathetic nerves.72 Studies of the biomechanical properties of the intact female rat urethra in vitro have confirmed the large contribution of striated

Autonomic Systems to the Urinary Bladder and Sexual Organs

muscle activity and nicotinic receptor mechanisms to the contractions of the mid-urethra.70

Afferent Pathways Afferent axons innervating the urinary tract are present in the three sets of nerves.11,24,42,48,52 The most important afferents for initiating micturition are those passing in the pelvic nerve to the sacral spinal cord. These afferents are small myelinated (A␦) and unmyelinated (C) fibers that convey information from receptors in the bladder wall to second-order neurons in the spinal cord. A␦ bladder afferents in the cat respond in a graded manner to passive distention as well as active contraction of the bladder and exhibit pressure thresholds in the range of 5 to 15 mm Hg, which are similar to those pressures at which humans report the first sensation of bladder filling.33,98,99 These fibers also code for noxious stimuli in the bladder. Conversely, C-fiber bladder afferents in the cat have very high thresholds and commonly do not respond to even high levels of intravesical pressure.65,98 However, activity in some of these afferents is unmasked or enhanced by chemical irritation of the bladder mucosa. These findings indicate that C-fiber afferents in the cat have specialized functions, such as the signaling of inflammatory or noxious events in the lower urinary tract. Nociceptive and mechanoceptive information is also carried in the hypogastric nerves to the thoracolumbar segments of the spinal cord.11 In the rat, A-fiber and C-fiber bladder afferents are not distinguishable on the basis of stimulus modality; thus both types of afferents consist of mechanosensitive and chemosensitive populations.99,108,116,117 C-fiber afferents that respond only to bladder filling have also been identified in the rat bladder and appear to be volume receptors possibly sensitive to stretch of the mucosa. C-fiber afferents are sensitive to the neurotoxins capsaicin and resiniferatoxin as well as to other substances such as tachykinins, NO, ATP, prostaglandins, and neurotrophic factors released in the bladder by afferent nerves, urothelial cells, and inflammatory cells.39,80,98,108,136 These substances can sensitize the afferent nerves and change their response to mechanical stimuli. The properties of lumbosacral dorsal root ganglion cells innervating the bladder, urethra, and EUS in the rat have been studied with patch-clamp recording techniques in combination with axonal tracing methods to identify the different populations of neurons.140,141,144–146 Based on responsiveness to capsaicin, it is estimated that approximately 70% of bladder afferent neurons in the rat are of the C-fiber type. These neurons exhibit high-threshold tetrodotoxin (TTX)–resistant sodium channels and action potentials and phasic firing (one to two spikes) in response to prolonged depolarizing current pulses. Approximately 20% of the bladder C-fiber afferent neurons also are excited by ATP, which induces a depolarization and firing

305

of afferent neurons by activating P2X3 or P2X2/3 receptors. These neurons express isolectin B4 binding, which is commonly used as a marker for ATP-responsive sensory neurons. In contrast, A-fiber afferent neurons are resistant to capsaicin and ATP, exhibiting low-threshold TTX-sensitive sodium channels and action potentials and tonic firing (multiple spikes) in response to depolarizing current pulses. C-fiber bladder afferent neurons also express a slowly decaying A-type K⫹ channel that controls spike threshold and firing frequency.141,144,146 Suppression of this K⫹ channel induces hyperexcitability of bladder afferent nerves. These properties of dorsal root ganglion cells are consistent with the different properties of A-fiber and C-fiber afferent receptors in the bladder. Immunohistochemical studies have shown that a large percentage of bladder afferent neurons contain the peptides CGRP, VIP, pituitary adenyl cyclase–activating peptide (PACAP), tachykinins, galanin, and opioid peptides.43,74,83,98 In the spinal cord, peptidergic nerve terminals have a distribution very similar to the distribution of pelvic nerve afferents labeled with horseradish peroxidase. Nerves containing these peptides are also common in the bladder, in the submucosal and epithelial layers, and around blood vessels.56 Peptidergic bladder afferent neurons in the rat also express tyrosine kinase A, a high-affinity receptor for nerve growth factor (NGF), and receptors for capsaicin (transient receptor potential protein V1 [TRPV1]) and tachykinins (neurokinin [NK]–2 and NK-3 receptors). Capsaicin, a neurotoxin that can release peptides from afferent terminals, produces inflammatory responses, including plasma extravasation and vasodilation, when applied locally to the bladder in experimental animals.83 These findings suggest that the neuropeptides may be important transmitters in the afferent pathways from the lower urinary tract. Tachykinins may also act back on afferent terminals in an auto-feedback manner to modulate the excitability of the terminals.98

Urothelial-Afferent Interactions Recent studies have revealed that the urothelium, which has been traditionally viewed as a passive barrier at the bladder luminal surface,79,81 also has specialized sensory and signaling properties that allow urothelial cells to respond to their chemical and physical environment and to engage in reciprocal chemical communication with neighboring nerves in the bladder wall.14–18,40,60,98,136 These properties include (1) expression of nicotinic, muscarinic, tachykinin, adrenergic, and capsaicin (TRPV1) receptors; (2) responsiveness to transmitters released from sensory nerves; (3) close physical association with afferent nerves; and (4) ability to release chemical mediators such as ATP and NO that can regulate the activity of adjacent nerves and thereby trigger local vascular changes and/or reflex bladder contractions. The role of ATP in urothelial-afferent communication has attracted considerable attention

306

Function of the Peripheral Nervous System

because bladder distention releases ATP from the urothelium and intravesical administration of ATP induces bladder hyperactivity, an effect blocked by administration of P2X purinergic receptor antagonists that suppress the excitatory action of ATP on bladder afferent neurons.98 Mice in which the P2X3 receptor was knocked out exhibited hypoactive bladder activity and inefficient voiding,40 suggesting that activation of P2X3 receptors on bladder afferent nerves by ATP released from the urothelium was essential for normal bladder function. It has also been reported that urothelial cells obtained from patients or cats with a chronic painful bladder condition (interstitial cystitis) release significantly larger amounts of ATP in response to mechanical stretching than do urothelial cells from normal patients.15,128 This raises the possibility that ATP-mediated signaling between the urothelium and afferent nerves is involved in the triggering of painful bladder sensations.

REFLEX MECHANISMS CONTROLLING THE LOWER URINARY TRACT The neural pathways controlling lower urinary tract function are organized as simple on-off switching circuits (Figs. 14–4 and Fig. 14–5; see also Fig. 14–1) that maintain a reciprocal relationship between the urinary bladder and urethral outlet.48,52 The principal reflex components of these switching circuits are listed in Table 14–2 and illustrated in Figure 14–5. Intravesical

FIGURE 14–4 Diagram illustrating the anatomy of the lower urinary tract and the switchlike function of the micturition reflex pathway. During urine storage, a low level of afferent activity activates efferent input to the urethral sphincter. A high level of afferent activity induced by bladder distention activates the switching circuit in the central nervous system (CNS), producing firing in the efferent pathways to the bladder, inhibition of the efferent outflow to the sphincter, and urine elimination.

pressure measurements during bladder filling in both humans and animals reveal low and relatively constant bladder pressures when bladder volume is below the threshold for inducing voiding (see Fig. 14–1). The accommodation of the bladder to increasing volumes of urine is primarily a passive phenomenon dependent upon the intrinsic properties of the vesical smooth muscle and quiescence of the parasympathetic efferent pathway. In addition, in some species urine storage is also facilitated by sympathetic reflexes that mediate an inhibition of bladder activity, closure of the bladder neck, and contraction of the proximal urethra (see Table 14–2 and Fig. 14–5).51,52 During bladder filling the activity of the sphincter EMG also increases (see Fig. 14–1), reflecting an increase in efferent firing in the pudendal nerve and an increase in outlet resistance that contributes to the maintenance of urinary continence. The storage phase of the urinary bladder can be switched to the voiding phase either involuntarily (reflexly) or voluntarily (see Fig. 14–1). The former is readily demonstrated in the human infant (Fig. 14–1A) or in the anesthetized animal when the volume of urine exceeds the micturition threshold. At this point, increased afferent firing from tension receptors in the bladder reverses the pattern of efferent outflow, producing firing in the sacral parasympathetic pathways and inhibition of sympathetic and somatic pathways. The expulsion phase consists of an initial relaxation of the urethral sphincter (Fig. 14–1) followed in a few seconds by a contraction of the bladder, an increase in bladder pressure, and flow of urine. Relaxation of the urethral outlet is mediated by activation of a parasympathetic reflex pathway to the urethra that triggers the release of NO, an inhibitory transmitter, as well as by removal of adrenergic and somatic cholinergic excitatory inputs to the urethra. Secondary reflexes elicited by flow of urine through the urethra facilitate bladder emptying.48 These reflexes require the integrative action of neuronal populations at various levels of the neuraxis (see Fig. 14–5). Certain reflexes, for example, those mediating the excitatory outflow to the sphincters and the sympathetic inhibitory outflow to the bladder, are organized at the spinal level (Fig. 14–5A), whereas the parasympathetic outflow to the detrusor has a more complicated central organization involving spinal and spinobulbospinal pathways (Fig. 14–5B).

ANATOMY OF CENTRAL NERVOUS PATHWAYS CONTROLLING THE LOWER URINARY TRACT The reflex circuitry controlling micturition consists of four basic components: spinal efferent neurons, spinal interneurons, primary afferent neurons, and neurons in

Autonomic Systems to the Urinary Bladder and Sexual Organs

A

307

B

FIGURE 14–5 Diagram showing neural circuits controlling continence and micturition. A, Urine storage reflexes. During the storage of urine, distention of the bladder produces low-level vesical afferent firing, which in turn stimulates (1) the sympathetic outflow to the bladder outlet (base and urethra) and (2) pudendal outflow to the external urethral sphincter. These responses occur by spinal reflex pathways and represent guarding reflexes, which promote continence. Sympathetic firing also inhibits detrusor muscle and modulates transmission in bladder ganglia. A region in the rostral pons (the pontine storage center) increases external urethral sphincter activity. B, Voiding reflexes. During elimination of urine, intense bladder afferent firing activates spinobulbospinal reflex pathways passing through the pontine micturition center, which stimulate the parasympathetic outflow to the bladder and internal sphincter smooth muscle and inhibit the sympathetic and pudendal outflow to the urethral outlet. Ascending afferent input from the spinal cord may pass through relay neurons in the periaqueductal gray (PAG) before reaching the pontine micturition center.

the brain that modulate spinal reflex pathways.44,98 Transneuronal virus tracing, measurements of gene expression, and patch-clamp recording in spinal cord slice preparations have recently provided many new insights into the morphologic and electrophysiologic properties of these reflex components. Neurotropic viruses, such as pseudorabies virus (PRV), have been particularly useful because they can be injected into a target organ (urinary bladder, urethra, urethral sphincter) and then move intra-axonally from the periphery to the central nervous system, where they replicate and then pass retrogradely across synapses to infect secondand third-order neurons in the neural pathways.101,126,135 Because PRV can be transported across many synapses,

it could sequentially infect all of the neurons that connect directly or indirectly to the lower urinary tract (see Fig. 14–3).

Pathways in the Spinal Cord Spinal Cord Anatomy The spinal cord gray matter is divided into three general regions: (1) the dorsal horn, which contains interneurons that process sensory input; (2) the ventral horn, which contains motoneurons; and (3) the intermediate region, which is located between the dorsal and ventral horns and contains interneurons and autonomic PGNs. These regions are further subdivided into layers, or laminae, that are

308

Function of the Peripheral Nervous System

Table 14–2. Receptors for Putative Transmitters in the Lower Urinary Tract* Tissue

Cholinergic

Adrenergic

Other

Bladder body

⫹ (M2) ⫹ (M3)

⫺ (␤3)

Bladder base

⫹ (␣1)

Urothelium

⫹ (M2) ⫹ (M3) ⫹ (M2) ⫹ (M3) ⫹ (N)

Ganglia

⫹ (N) ⫹ (M1)

⫺ (␣2) ⫹ (␣1) ⫹ (␤)

Urethra

⫹ (M)

⫹ (␣1) ⫹ (␣2) ⫺ (␤)

⫹ Purinergic (P2X) ⫺ VIP ⫹ Substance P ⫹ Purinergic (P2X) ⫺ VIP ⫹ TRPV1 ⫹ Purinergic (P2X) ⫹ 5-HT ⫹ Substance P ⫺ Enkephalinergic (␦) ⫹/⫺ Purinergic (P2X, P1) ⫹ Substance P ⫺ Nitric oxide ⫹/⫺ Purinergic (P2X, P1) ⫺ VIP ⫹ NPY

Sphincter striated muscle

⫹ (N)

⫹ (␣) ⫹ (␤)

NPY ⫽ neuropeptide Y; TRPV1 ⫽ transient receptor potential protein V1; VIP ⫽ vasoactive intestinal polypeptide. *Letters in parentheses indicate receptor type: M (muscarinic) and N (nicotinic). ⫹ and ⫺ indicate excitatory and inhibitory, respectively.

numbered starting with the superficial layer of the dorsal horn (lamina I) and extending to the ventral horn (lamina IX) and the commissure connecting the two sides of the spinal cord (lamina X) (Fig. 14–6D). Efferent Pathways Parasympathetic PGNs innervating the lower urinary tract are located in the intermediolateral gray matter (laminae V through VII) in the sacral segments of the spinal cord (see Fig. 14–6),6,47,94,96 whereas sympathetic PGNs are located in both medial (lamina X) and lateral (laminae V through VII) sites in the intermediate gray matter of the rostral lumbar spinal cord. Parasympathetic PGNs send dendrites to discrete regions of the spinal cord, including (1) the lateral and dorsolateral funiculus, (2) lamina I on the lateral edge of the dorsal horn, (3) the dorsal gray commissure (lamina X), and (4) gray matter and lateral funiculus ventral to the autonomic nucleus.44,48 As discussed later, this dendritic structure very likely indicates the origin of important synaptic inputs to these cells. Pudendal motoneurons innervating the EUS in the cat are located in the ventrolateral division of Onuf’s nucleus and send dendritic projections (1) into the lateral funiculus, (2) into lamina X, (3) into the intermediolateral gray matter, and (4) rostrocaudally within the nucleus.44 The dendritic distribution of sphincter motoneurons (i.e., lateral, dorsolateral, and dorsomedial) is similar to that of sacral PGNs, indicating that these two populations of neurons may receive synaptic inputs from the same interneuronal sites and fiber tracts in the spinal cord.

Afferent Projections in the Spinal Cord Afferent pathways from the lower urinary tract project to discrete regions of the dorsal horn that contain the interneurons as well as the soma and/or dendrites of efferent neurons innervating the lower urinary tract. Pelvic nerve afferent pathways from the urinary bladder of the cat and rat project into Lissauer’s tract at the apex of the dorsal horn and then pass rostrocaudally, giving off collaterals that extend laterally and medially through the superficial layer of the dorsal horn (lamina I) into the deeper layers (laminae V through VII and X) at the base of the dorsal horn (see Fig. 14–6A).97,123 The lateral pathway, which is the most prominent projection, terminates in the region of the sacral parasympathetic nucleus and also sends some axons to the dorsal commissure (Fig. 14–6A). Pudendal afferent pathways from the urethra and urethral sphincter exhibit a similar pattern of termination in the sacral spinal cord,48,98 whereas pudendal afferent pathways from cutaneous receptors (e.g., penis) have a prominent projection to deeper layers of the dorsal horn and the dorsal commissure. The overlap of bladder and urethral afferents in the lateral dorsal horn and dorsal commissure indicates these regions are likely to be important sites of viscerosomatic integration and to be involved in coordinating bladder and sphincter activity. Spinal Interneurons As shown in Figure 14–6B and C, interneurons retrogradely labeled by injection of PRV into the urinary bladder of the rat are located in the regions of the

Autonomic Systems to the Urinary Bladder and Sexual Organs

309

FIGURE 14–6 Comparison of the distribution of bladder afferent projections to the L6 spinal cord of the rat (A) with the distribution of c-fos–positive cells in the L6 spinal segment following chemical irritation of the lower urinary tract of the rat (B) and the distribution of interneurons in the L6 spinal cord labeled by transneuronal transport of pseudorabies virus injected into the urinary bladder (C). Afferents are labeled by wheat germ agglutinin–horseradish peroxidase injected into the urinary bladder. C-fos immunoreactivity is present in the nuclei of cells. CC ⫽ central canal; DH ⫽ dorsal horn; SPN ⫽ sacral parasympathetic nucleus. Calibration represents 500 ␮m. D, Drawing shows the laminar organization of the cat spinal cord.

spinal cord receiving afferent input from the bladder.101,126 Interneuronal locations also overlap in many respects with the dendritic distribution of the efferent neurons. A similar distribution of labeled interneurons has been noted following injections of virus into the urethra135 or the external urethral sphincter, indicating a prominent overlap of the interneuronal pathways controlling the various target organs of the lower urinary tract. The spinal neurons involved in processing afferent input from the lower urinary tract have been identified by the expression of the immediate early gene c-fos (see Fig. 14–6B). In the rat, noxious or non-noxious stimulation of the bladder and urethra increases the levels of Fos protein primarily in the dorsal commissure, the superficial dorsal horn, and the area of the sacral parasympathetic nucleus (Fig. 14–6B). Some of these interneurons send long projections to the brain, whereas others make local connections in the spinal cord and participate in segmental spinal reflexes. Patch-clamp recordings from parasympathetic PGNs in the neonatal rat spinal slice preparation have revealed that interneurons located immediately dorsal and medial to

the parasympathetic nucleus make direct monosynaptic connections with the PGNs. 6 Microstimulation of interneurons in both locations elicits glutamatergic, N-methyl-D-aspartic acid (NMDA) and non-NMDA excitatory postsynaptic currents in PGNs. Stimulation of a subpopulation of medial interneurons elicits ␥-aminobutyric acid (GABA)ergic and glycinergic inhibitory postsynaptic currents. Thus local interneurons are likely to play an important role in both excitatory and inhibitory reflex pathways controlling the preganglionic outflow to the lower urinary tract. Glutamatergic excitatory inputs have also been elicited by stimulation of the projections from lamina X and the lateral funiculus.94,95

Pathways in the Brain The neurons in the brain that control the lower urinary tract have been studied with a variety of anatomic tracing techniques in several species. In the rat, transneuronal virus tracing methods have identified many populations of central neurons that are involved in the control of

310

Function of the Peripheral Nervous System

the bladder, urethra, and urethral sphincter, including Barrington’s nucleus (the pontine micturition center [PMC]); the medullary raphe nucleus, which contains serotonergic neurons; the locus ceruleus, which contains noradrenergic neurons; periaqueductal gray (PAG); and the A5 noradrenergic cell group. Several regions in the hypothalamus and the cerebral cortex also exhibited virusinfected cells. Neurons in the cortex were located primarily in the medial frontal cortex. Other anatomic studies in which anterograde tracer substances were injected into brain areas and then identified in terminals in the spinal cord are consistent with the virus tracing data. Tracer injected into the paraventricular nucleus of the hypothalamus labeled terminals in the sacral parasympathetic nucleus as well as the sphincter motor nucleus. In contrast, neurons in the anterior hypothalamus project to the PMC. Neurons in the PMC in turn project primarily to the sacral parasympathetic nucleus and the lateral edge of the dorsal horn and the dorsal commissure, areas containing dendritic projections from the PGNs, sphincter motoneurons, and afferent inputs from the bladder. Finally, projections from neurons in the lateral pons terminate rather selectively in the sphincter motor nucleus. Thus the sites of termination of descending projections from the PMC are optimally located to regulate reflex mechanisms at the spinal level.

ORGANIZATION OF URINE STORAGE AND VOIDING REFLEXES Sympathetic Pathways The integrity of the sympathetic input to the lower urinary tract is not essential for the performance of micturition.48,131 However, physiologic experiments in animals indicate that, during bladder filling, the sympathetic system does provide a tonic inhibitory input to the bladder as well as an excitatory input to the urethra. This sympathetic input is physiologically significant because surgical interruption or pharmacologic blockade of the sympathetic innervation can reduce urethral outflow resistance, reduce bladder capacity, and increase the frequency and amplitude of bladder contractions recorded under constant-volume conditions. Sympathetic reflex activity is elicited by a sacrolumbar intersegmental spinal reflex pathway that is triggered by vesical afferent activity in the pelvic nerves (see Fig. 14–5A).51 The reflex pathway is inhibited when bladder pressure is raised to the threshold for producing micturition. This inhibitory response is abolished by transection of the spinal cord at the lower thoracic level, indicating that it originates at a supraspinal site, possibly the PMC. Therefore, the vesicosympathetic reflex represents a negative feedback mechanism whereby an increase in bladder pressure tends to increase inhibitory input to

vesical ganglia and smooth muscle, thus allowing the bladder to accommodate large volumes (Fig. 14–5A). Increased sympathetic excitatory input to the bladder base and urethra would complement these mechanisms by increasing outflow resistance.

Somatic Pathways to the Urethral Sphincter Motoneurons innervating the striated muscles of the urethral sphincter exhibit a tonic discharge that increases during bladder filling. This activity is mediated in part by low-level afferent input from the bladder (see Fig. 14–5A). During micturition the firing of sphincter motoneurons is inhibited. This inhibition is dependent in part on supraspinal mechanisms, because it is not as prominent in chronic spinal cord–transected animals. Electrical stimulation of the PMC induces sphincter relaxation, suggesting that bulbospinal pathways from the pons may be responsible for maintaining the normal reciprocal relationship between bladder and sphincter.19,33

Voiding Reflexes Spinobulbospinal Micturition Reflex Pathway Micturition is mediated by activation of the sacral parasympathetic efferent pathway to the bladder and the urethra (see Fig. 14–5B) as well as reciprocal inhibition of the somatic pathway to the urethral sphincter (see Table 14–1 and Fig. 14–5B). Studies in animals using brain-lesioning techniques revealed that neurons in the brainstem at the level of the inferior colliculus (i.e., the PMC) have an essential role in the control of the parasympathetic component of micturition.48,85,131 Removal of areas of the brain above the inferior colliculus by intercollicular decerebration usually facilitates micturition by elimination of inhibitory inputs from more rostral centers.138 However, transections at any point below the colliculi abolish micturition. Bilateral lesions in the rostral pons in the region of the locus ceruleus in cats or Barrington’s nucleus in rats also abolish micturition, whereas electrical or chemical stimulation at these sites triggers bladder contractions and micturition.85 These observations led to the concept of a spinobulbospinal micturition reflex pathway that passes through the PMC (Fig. 14–7; see also Fig. 14–5B). The pathway functions as an “on-off” switch (see Fig. 14–4) that is activated by a critical level of afferent activity arising from tension receptors in the bladder and is in turn modulated by inhibitory and excitatory influences from areas of the brain rostral to the pons (e.g., diencephalon and cerebral cortex) (see Fig. 14–7).48,131 In contrast to the reflex control of the bladder, the parasympathetic control of the urethra in the rat appears to be dependent on pathways organized in the spinal cord

Autonomic Systems to the Urinary Bladder and Sexual Organs

311

FIGURE 14–7 Diagram showing the organization of the parasympathetic excitatory reflex pathway to the detrusor muscle. Scheme is based on electrophysiologic studies in cats. In animals with an intact spinal cord, micturition is initiated by a supraspinal reflex pathway passing through a center in the brainstem. The pathway is triggered by myelinated afferents (A␦ fibers), which are connected to the tension receptors in the bladder wall. Injury to the spinal cord above the sacral segments interrupts the connections between the brain and spinal autonomic centers and initially blocks micturition. However, over a period of several weeks following cord injury, a spinal reflex mechanism emerges that is triggered by unmyelinated vesical afferents (C fibers); the A-fiber afferent inputs are ineffective. The C-fiber reflex pathway is usually weak or undetectable in animals with an intact nervous system. Stimulation of the C-fiber bladder afferents by instillation of ice water into the bladder (cold stimulation) activates voiding responses in patients with spinal cord injury. Capsaicin (20 to 30 mg subcutaneously) blocks the C-fiber reflex in chronic spinal cord–injured cats, but does not block micturition reflexes in intact cats. Intravesical capsaicin also suppresses detrusor hyperreflexia and cold-evoked reflexes in patients with neurogenic bladder dysfunction.

that are modulated by input from the brain. NO-mediated relaxation of the urethra that occurs in response to bladder distention is reduced but not eliminated by acute transection of the spinal cord.71 The reflex relaxation of the urethra is also very prominent in chronic spinal cord–transected rats. Electrophysiologic studies in cats and rats have confirmed that the parasympathetic efferent outflow to the urinary bladder is activated by a long-latency supraspinal reflex pathway.36,44,47,52,84 In cats, recordings from sacral parasympathetic PGNs innervating the urinary bladder show that reflex firing occurs with a long latency (65 to 100 ms) following stimulation of myelinated (A␦) vesical afferents in the pelvic nerve. Afferent stimulation also evokes negative field potentials in the rostral pons at latencies of 30 to 40 ms, whereas electrical stimulation in the pons excites sacral PGNs at latencies of 45 to 60 ms. The sum of the latencies for the spinobulbar and bulbospinal components of the reflex pathway approximates the latency for the entire

reflex. In cats it is believed that the ascending afferent pathways from the spinal cord project to a relay station in the PAG, which then connects to the PMC (see Fig. 14–5B).19 Pontine Micturition Center Physiologic and anatomic experiments have provided substantial support for the concept that neuronal circuitry in the PMC functions as a switch in the micturition reflex pathway. The switch seems to regulate bladder capacity and also coordinate the activity of the bladder and EUS. Electrical or chemical stimulation in the PMC of the rat, cat, and dog induces (1) a suppression of urethral sphincter EMG, (2) firing of sacral PGNs, (3) bladder contractions, and (4) release of urine.19,48,84,131 Conversely, microinjections of putative inhibitory transmitters into the PMC of the cat can increase the volume threshold for inducing micturition and in high doses completely block reflex voiding, indicating that synapses in this region are

312

Function of the Peripheral Nervous System

important for regulating the set point for reflex voiding and also are an essential link in the reflex pathway.84 Brain imaging studies using positron emission tomography (PET) or functional magnetic resonance imaging (fMRI) have identified increased neuronal activity in the PMC and PAG during voiding.9,10,21,22 Suprapontine Control of Micturition The organization of suprapontine pathways controlling micturition is less well defined, despite the fact that there is a large body of literature dealing with the responses of the lower urinary tract to lesions or electrical stimulation of the brain. In brief, it appears that the voluntary control of micturition in humans is dependent upon (1) connections between the frontal cortex and the septal and preoptic regions of the hypothalamus and (2) connections between the paracentral lobule and the brainstem and spinal cord.48,131 Lesions to these areas of cortex resulting from tumors, aneurysms, or cerebrovascular disease appear to remove inhibitory control over the anterior hypothalamic area, which normally provides an excitatory input to micturition centers in the brainstem.138 Electrical stimulation of anterior and lateral hypothalamic regions in animals induces bladder contractions and voiding, whereas stimulation of posterior and medial hypothalamic areas inhibits bladder activity. According to results obtained in cats, the inhibitory and excitatory effects of hypothalamic stimulation are believed to be mediated, respectively, by activation of sympathetic inhibitory pathways and activation of parasympathetic excitatory pathways to the bladder. Human PET scan studies have revealed that two cortical areas (the right dorsolateral prefrontal cortex and the anterior cingulate gyrus) were active (i.e., exhibited increased blood flow) during voiding.19,21,22 The hypothalamus, including the preoptic area as well as the pons and the PAG, also showed activity in concert with voluntary micturition. It is noteworthy that the active areas were predominately on the right side of the brain, which is consistent with reports that urge incontinence is correlated with lesions in the right hemisphere. Other PET studies that examined the changes in brain activity during filling of the bladder revealed that increased activity occurred in the PAG, the midline pons, the mid-cingulate gyrus, and bilaterally in the frontal lobes.9,10,90 It was concluded that the results were consistent with the notion that the PAG receives information about bladder fullness and then relays this information to other brain areas involved in the control of bladder storage. A PET study was also conducted in adult female volunteers to identify brain structures involved in voluntary control of pelvic floor muscles.20 The results revealed that the superomedial precentral gyrus, the most medial portion of the motor cortex, is activated during pelvic floor contraction. In addition, the right anterior cingulate gyrus was activated during sustained pelvic floor straining.

NEUROTRANSMITTERS IN MICTURITION REFLEX PATHWAYS Excitatory Neurotransmitters Excitatory transmission in the central pathways to the lower urinary tract may depend on several types of transmitters, including glutamic acid, neuropeptides (substance P, VIP, PACAP), norepinephrine, ACh, and NO.48,98 Experiments in rats have revealed that glutamic acid is an essential transmitter in the ascending, pontine, and descending limbs of the spinobulbospinal micturition reflex pathway and in the reflex pathways controlling the EUS. NMDA and non-NMDA glutamatergic synaptic mechanisms appear to interact synergistically to mediate transmission in these pathways.

Inhibitory Neurotransmitters Damage to central inhibitory mechanisms following disease or injury to the nervous system leads to failure of urine storage and bladder hyperactivity and incontinence.137–139,142,143 Animal studies indicate that transmission in the PMC and spinal cord is regulated by multiple transmitters, including opioid peptides (enkephalins), inhibitory amino acids (GABA, glycine), 5-hydroxytryptamine, ACh, and dopamine.48,98,139,142,143 Experiments in anesthetized animals indicate that GABA and enkephalins exert a tonic inhibitory control in the PMC and regulate bladder capacity. The inhibitory effects are mediated by GABAA and ␮ opioid receptors, respectively. Administration of GABAA or opioid receptor antagonists into the PMC reduces the micturition volume threshold, indicating that the set point for reflex voiding is regulated by inhibitory mechanisms in the brain (see Fig. 14-5).48,55,85 GABA and enkephalins also have inhibitory actions in the spinal cord. Baclofen, a GABAB agonist that mimics the inhibitory effect of GABA, has been used clinically via intrathecal administration in patients with hyperactive bladders to suppress bladder activity and to promote urine storage. Patients with idiopathic Parkinson’s disease often exhibit bladder hyperactivity, suggesting a role of dopamine in the control of bladder function. Animal models for Parkinson’s disease have been developed in monkeys and rats by administering neurotoxins (1-methyl-4-phenyl-1,2,3,6tetrahydropyridine [MPTP] or 6-hydroxydopamine) to destroy dopamine neurons.55,142,143 Following treatment the animals show motor symptoms typical of Parkinson’s disease and also have hyperactive bladders. Pharmacologic studies in MPTP-treated monkeys revealed that the bladder hyperactivity was due to the loss of dopaminergic inhibition mediated by D1 dopaminergic receptors. D2 dopaminergic receptors can mediate a facilitation of micturition. Activation of these receptors also contributes to the bladder hyperactivity after cerebral infarction in the rat.139

Autonomic Systems to the Urinary Bladder and Sexual Organs

NEUROGENIC DYSFUNCTION OF THE LOWER URINARY TRACT Neurogenic disturbances of micturition can be classified into two general categories: failure to store and failure to eliminate urine.137 Problems with storage occur with differing degrees of severity, ranging from reduced bladder capacity and frequency of urination to urgency and incontinence. A common finding is that disorders affecting the brain, particularly suprapontine areas, produce hyperactive or uninhibited bladders. Cerebrovascular accidents, Parkinson’s disease, tumors, and demyelinating diseases are common causes of this problem.13,24,41,61,112 Failure to eliminate urine occurs in various conditions that interrupt the detrusor-to-detrusor excitatory reflex pathway or that interfere with the coordination between detrusor and sphincters.33,137 Areflexic bladders can occur with (1) lower motor neuron lesions, including damage to the pelvic nerve or the sacral spinal cord; (2) lesions of the afferent pathways (e.g., diabetes, tabes dorsalis, pernicious anemia, herniated intervertebral disc); or (3) the acute stage of spinal cord injury (an upper motor neuron lesion).

Spinal Injury Complete spinal cord injury rostral to the lumbosacral level eliminates voluntary and supraspinal control of voiding, leading initially to an areflexic bladder and complete urinary retention followed by a slow development of automatic micturition and bladder hyperactivity mediated by spinal reflex pathways.24,131,137 Following recovery of reflex bladder activity, these patients usually still exhibit urinary retention as a result of a loss of coordination between the bladder and sphincter (detrusor-sphincter dyssynergia). Because sphincter dyssynergia is rarely seen in patients with suprapontine lesions who also exhibit hyperactive bladders, the condition has been attributed to the elimination of bulbospinal inhibitory pathways, which originate in the PMC (see Fig. 14–7). A similar mechanism may account for dyssynergia of the smooth muscle sphincter, which is mediated by sympathetic efferent pathways. Dyssynergia is treated surgically (sphincterotomy) or with drugs (skeletal muscle relaxants, botulinum toxin, adrenergic blocking agents) that depress somatic or sympathetic reflex pathways. The mechanisms contributing to the recovery of bladder function have been studied in rats and cats using electrophysiologic techniques.36,47,50,84,98,140 These studies revealed that the micturition reflex pathways in spinal cord–intact and chronic spinal cord–transected animals are markedly different. In both species, the central delay for the micturition reflex in chronic spinal

313

cord–transected animals is considerably shorter (⬍5 ms in rats; 15 to 40 ms in cats) than in intact animals (60 to 75 ms). In addition, in chronic spinal cord–transected cats the afferent limb of the micturition reflex consists of unmyelinated (C-fiber) afferents, whereas in intact cats it consists of myelinated (A␦) afferents (see Fig. 14–7). This was demonstrated not only with electrophysiologic recording but also by administering capsaicin, a neurotoxin that is known to disrupt the function of C-fiber afferents. In normal cats, capsaicin injected systemically in large doses did not block reflex contractions of the bladder or the A␦-fiber evoked bladder reflex. However, in chronic spinal cord–transected cats (3 to 6 weeks after spinal cord transection), capsaicin completely blocked the rhythmic bladder contractions induced by bladder distention and blocked the C-fiber–evoked reflex firing recorded on bladder postganglionic nerves.36 These data indicate that two distinct central pathways (supraspinal and spinal) utilizing different peripheral afferent limbs (A and C fiber) can mediate detrusor-to-detrusor reflexes in the cat (see Fig. 14–7). The properties of the peripheral C-fiber afferent receptors also appear to be changed in the spinal cord–injured cat. C-fiber bladder afferents in the cat usually do not respond to bladder distention (i.e., silent C fibers).65 However, in chronic spinal cord–transected cats bladder distention initiates automatic micturition by activating C-fiber afferent neurons.36,50 Thus spinal injury must change the properties of C-fiber afferent receptors in the bladder. Experiments in spinal cord–injured rats revealed that C-fiber dorsal ganglion neurons innervating the bladder increase in size,78,140 exhibit increased excitability related in part to a downregulation of A-type K⫹ channels, and exhibit a shift in the expression of Na⫹ channels from a high-threshold TTX-resistant subtype to a TTX-sensitive subtype of Na⫹ channels. Neurotrophic factors appear to contribute to this change because, after spinal cord injury, expression of neurotrophic factors, including NGF, increases in the bladder in concert with bladder hypertrophy.98 The latter most likely occurs as a result of overdistention-elicited detrusor-sphincter dyssynergia and urinary retention. A contribution of NGF to the change in bladder afferent neuron excitability and detrusor hyperreflexia in spinal cord–injured rats is supported by several observations, including (1) administration of exogenous NGF to the bladder or spinal cord can induce bladder hyperactivity and increased excitability of bladder afferent neurons accompanied by downregulation of A-type K⫹ channels, and (2) immunoneutralization of NGF in the spinal cord reduces the bladder hyperreflexia in spinal cord–injured rats and reduces the NGF levels in the lumbosacral dorsal root ganglia.114 The effects of NGF immunoneutralization can be duplicated by systemic or intravesical administration of the C-fiber

314

Function of the Peripheral Nervous System

neurotoxins capsaicin or resiniferatoxin.36,50 Intravesical administration of these toxins in patients with neurogenic bladder dysfunction also reduces bladder hyperactivity and incontinence episodes and increases bladder capacity.55,98 Other reflexes that are unmasked following spinal cord injury also appear to be mediated by C-fiber afferents. For example, it is known that instillation of cold water into the bladder of patients with upper motoneuron lesions induces reflex voiding (the Bors ice water test).24,33,48 This reflex does not occur in normal patients. Recently, it has been shown in the cat that C-fiber bladder afferents are responsible for cold-induced bladder reflexes (see Fig. 14–7). The ice water test is also positive in patients with multiple sclerosis, cerebrovascular disease, Parkinson’s disease, and benign prostatic hypertrophy, as well as in normal infants. These observations suggest that cold-evoked bladder reflexes are mediated by a primitive spinal pathway that is present in the immature nervous system and then is suppressed during postnatal development as supraspinal mechanisms assume the dominant role in controlling micturition. However, when supraspinal controls are eliminated by spinal cord injury or neurologic diseases, such as multiple sclerosis, it appears that the spinal reflexes reemerge. Other reflexes also emerge after spinal cord injury. For example, arterial pressor responses induced by bladder distention (autonomic dysreflexia) occur in spinal cord–injured patients and animals. In normal and spinal cord–injured rats, the increase in blood pressure induced by isometric bladder contractions or distention is suppressed by capsaicin, the C-fiber neurotoxin.39 This suggests that C-fiber afferents mediate bladder vascular reflexes in this species. Similar capsaicin-sensitive afferents may be responsible for autonomic dysreflexia in quadriplegic patients.33

Sex Organs The physiologic changes initiated by erotic stimuli can be divided into four distinct phases (excitement, plateau, orgasm, and resolution) that have been designated collectively the sexual response cycle.87 Although anatomic differences obviously preclude identical responses in male and female during each phase of the cycle, it is clear that similar vascular responses (skin flush, penile and clitoral erection), secretory responses (stimulation of the prostate, bulbourethral gland, and glands of Littre in the male and Bartholin’s and paraurethral glands in the female), and responses of smooth and striated muscles occur in both sexes.4,12,31,45,110,111,115 This section reviews the neural control of the principal autonomic and somatic responses in the male sexual response cycle (erection, secretion, emission, and ejaculation).

INNERVATION Parasympathetic Pathways As described earlier with regard to micturition, three sets of nerves provide an innervation to the urogenital system (see Fig. 14–2). Parasympathetic preganglionic axons from the sacral spinal cord provide the efferent outflow to erectile tissue in the penis and to the seminal vesicles, prostate, and urethral glands (Table 14–3). The sacral pathways have cholinergic ganglionic relay stations in the pelvic plexus and possibly in the effector organs. The postganglionic parasympathetic neurons synthesize and release several transmitters, including NO, ACh, VIP, and ATP. NO is thought to be the major transmitter mediating neurally induced erections,1–4,7,27,29,30,45,59 whereas ACh appears to be important in stimulating

Table 14–3. Male Sexual Reflexes Response

Afferent Source

Efferent Nerves

Central Pathway

Effector Organ

Penile erection Reflexogenic

Pudendal nerve

Sacral parasympathetic

Sacral spinal reflex

Dilation of arterial supply to corpus cavernosum and spongiosum

Auditory, olfactory, visual, imaginative Pudendal nerve

Lumbar sympathetic

Supraspinal origin

Lumbosacral sympathetic and parasympathetic Lumbar sympathetic

Sacral spinal reflex

Psychogenic Glandular secretion Seminal emission

Pudendal nerve

Ejaculation

Pudendal nerve

Somatic efferents in pudendal nerves

Intersegmental spinal reflex (sacrolumbar) Sacral spinal reflex

Seminal vesicles and prostate Contraction of vas deferens, ampulla, prostate and bladder neck closure Rhythmic contractions of bulbo- and ischiocavernous muscles

Autonomic Systems to the Urinary Bladder and Sexual Organs

secretion in glands.25 The functions of VIP and ATP are uncertain.

Sympathetic Pathways The sympathetic innervation of the genital organs consists of PGNs in the thoracolumbar segments of the spinal cord and postganglionic neurons in the paravertebral and prevertebral ganglia (inferior mesenteric and pelvic ganglia), which provide an input to the penile erectile tissue as well as to the smooth muscle of the ductus deferens, seminal vesicles, urethra, and prostate (see Table 14–3).1,3–5,45,77,82,102 Sympathetic postganglionic axons are carried in the pudendal nerve and in nerves arising from the pelvic plexus. Most sympathetic postganglionic neurons are noradrenergic and release norepinephrine, ATP, and neuropeptides such as NPY. These nerves produce constriction of blood vessels and cause a contraction of the vas deferens and urethra– bladder neck.45,57 Some sympathetic nerves arising in the pelvic plexus release NO and presumably induce vasodilation.45,109,111,115

Somatic Pathways The pudendal nerve, arising from the S2 to S4 segments of the spinal cord, provides an efferent excitatory input to the striated muscles (bulbocavernosus and ischiocavernosus) involved in ejaculation.86 The pudendal nerve also contains the principal afferent pathway from the penis (see Table 14–3), and from the clitoris and vagina in the female. Afferent innervation to the uterine cervix and uterus travels in the pelvic and sympathetic nerves.

ERECTION Peripheral Mechanisms Penile erection, which is one of the first responses to occur during sexual arousal, is a vascular phenomenon resulting from a neurally mediated increase in blood flow to the penile erectile tissue (corpora cavernosa and corpus spongiosum).3,27,111 The erectile tissue consists of large venous sinuses that contain very little blood when the penis is flaccid, but distend considerably when blood flow is increased. Dilation in the arterial supply to the cavernous tissue coupled with a relaxation of the sinusoidal smooth muscle in the trabecular tissue (increased compliance) is responsible for erection. During the initiation of erection, corpus blood flow increases 7 to 30 times.111 The neural pathways producing erection arise from both the sacral (parasympathetic) and the thoracolumbar (sympathetic) segments of the spinal cord (see Table 14–3). The postganglionic neurons in these pathways express neuronal

315

NOS (nNOS), and all smooth muscle regions of the penis are richly innervated by nerves containing nNOS.2,27,29,30 The endothelial cells in the penis also express endothelial NOS (eNOS) and can release NO in response to mechanical stimuli (shear stress) associated with changes in blood flow.3,27,29,69 NO directly activates soluble guanylate cyclase in the penile smooth muscle to increase the formation of cyclic GMP (cGMP), which in turn activates cGMPdependent protein kinase I (cGK I).3,66,67,111 Inactivation of cGK I in mice severely reduces reproductive function and markedly reduces the ability of corpus cavernosal tissue to relax in response to neurally, endothelially, or exogenously administered NO.3,67 cGK I is thought to act by multiple mechanisms, including suppression of membrane and sarcoplasmic reticulum Ca2⫹ channels and suppression of inositol trisphosphate (IP3)–induced intracellular Ca2⫹ release. The effects of NO are terminated by the enzymatic breakdown of cGMP by phosphodiesterase. Pharmacologic studies in animals have shown that erections elicited by stimulation of autonomic nerves are reduced by NOS inhibitors and enhanced by phosphodiesterase (PDE) inhibitors.26,27,30 One type of PDE (PDE-5) is highly expressed in penile tissues. Several PDE-5 inhibitors are currently used to treat erectile dysfunction.63 eNOS has also been implicated in neurally mediated erections.3,69 It has been suggested that synthesis of NO in nerves by nNOS initiates erections. However, this response, which is relatively brief, triggers increased blood flow and expansion of the penile vasculature and sinusoidal spaces. The resulting shear force on the endothelium activates a phosphatidylinositol 3-kinase (PI3-kinase) pathway that in turn stimulates the serine/threonine protein kinase Akt, causing direct phosphorylation of eNOS in endothelial cells. After phosphorylation the Ca2⫹ requirement for eNOS activity is reduced, causing increased production of NO, which induces a prolonged penile erection. Intercellular communication via gap junctions is thought to promote the spread of signals throughout the smooth muscle of the penis and amplify the relaxation.3 Exogenous ACh acts on postjunctional muscarinic receptors to elicit a contraction or relaxation of corpus cavernosum preparations. The latter effect is mediated by M3 receptors and release of NO from the endothelium. However, the failure of exogenous ACh to produce erection and the failure of atropine to block erections indicates that cholinergic mechanisms are not essential for neurally mediated increases in penile blood flow.45 However, ACh may act on prejunctional inhibitory receptors on adrenergic nerve terminals to suppress the release of norepinephrine and thereby facilitate NO-mediated erections.111 VIP activates G protein–coupled receptors to stimulate adenylate cyclase and increase cyclic AMP (cAMP), which in turn stimulates cAMP-dependent kinases to suppress contractile mechanisms in vascular smooth muscle and smooth muscle of the trabecular tissue of the penis.1,2,4

316

Function of the Peripheral Nervous System

Although it is clear that exogenous VIP can relax human cavernosal tissue strips in vitro, it has never been shown that VIP released from nerves is responsible for relaxation of penile smooth muscle in vitro or in vivo. The adrenergic innervation of the penis, which provides an excitatory or constrictor input to penile blood vessels, is thought to be involved primarily in detumescence.1,3,4,37,38,45,57,92,93 Electrical stimulation of sympathetic axons in either the hypogastric or pudendal nerves in various animals produces a substantial reduction in penile blood flow.45,110 The effect is blocked by ␣-adrenergic blocking agents. In some animals, a partial erection can be elicited by the administration of ␣-adrenergic blocking agents, suggesting that tonic sympathetic vasoconstrictor mechanisms may have an inhibitory influence on erection.45 Several mechanisms have been implicated in the noradrenergic vasoconstrictor effect, including activation of phospholipase C followed by formation of IP3 and diacylglycerol, which leads to release of intracellular Ca2⫹ as well as sensitization of contractile mechanisms to Ca2⫹.1–3,37,38,93 Ca2⫹ sensitization has been linked to activation of an intracellular signaling pathway involving Rho kinase, which can be stimulated by G protein–coupled receptors such as ␣-adrenergic and endothelin receptors, both of which can elicit contraction of penile smooth muscle. Activation of Rho kinase leads to a change in the phosphorylation state of myosin light chain kinase, resulting in phosphorylation of myosin and subsequent smooth muscle contraction. Administration of a Rho kinase inhibitor triggered a significant increase in intracavernous pressure and an NO-independent erection.38 Similarly, the expression of a dominant negative construct to downregulate Rho kinase enhanced erectile function in rats.37 Other studies have revealed that NO may act to relax penile smooth muscle by suppressing Rho kinase activity.3 These studies have raised the possibility that antagonism of Rho kinase activity may yield new treatments for erectile dysfunction.

Central Mechanisms Penile erection is primarily an involuntary or reflex phenomenon that can be elicited by a variety of stimuli and by at least two distinct central mechanisms (see Table 14–3). For example, psychogenic erections are initiated by supraspinal centers in response to auditory, visual, olfactory, tactile, and imaginative stimuli.4,87,91,111 The efferent limb of the reflex pathway can be in either the thoracolumbar or the sacral autonomic outflow.34,109 Reflexogenic erections, which are initiated by exteroceptive stimulation of the genital regions, are mediated by a sacral spinal reflex mechanism having an afferent limb in the pudendal nerves and an efferent limb in the sacral parasympathetic nerves. Electrophysiologic studies in animals have revealed that the reflex occurs with a short central delay and is not altered by transection of the spinal cord above the lumbar level.124

In patients with lower motor neuron lesions involving the sacral spinal cord, reflexogenic erections are abolished but psychogenic erections may still occur via the sympathetic innervation to the penis.34,45,110 In patients with spinal cord lesions above the level of T12, psychogenic erections are abolished but reflexogenic reactions persist. Under normal conditions it is likely that psychic and reflexogenic stimuli act synergistically in producing erections. It is also known that psychological factors, such as guilt and hostility, or endocrine disturbances that influence libido or supraspinal centers, can interfere with erectile reflexes.12,13,41,62,82,87,112 In addition, any factor that restricts blood flow to erectile tissue will alter erection. Thus atherosclerosis, thrombosis, or a depression of autonomic transmission by either disease or drugs can compromise the vascular responses necessary for tumescence. In adult diabetics, in whom both vascular and neural pathology are common, the incidence of erectile dysfunction is very high.26,62,92

GLANDULAR SECRETION During the second phase of the sexual response cycle (plateau), activity in parasympathetic pathways stimulates mucus secretion from bulbourethral and Littre’s glands and secretion from the seminal vesicles and the prostate gland.25,87,110 Mucus secretion contributes to lubrication of the penis, whereas secretions from the seminal vesicles and prostate provide the bulk of the fluid and chemical factors that contribute to the viability and motility of the spermatozoa. Glandular secretion is thought to be mediated by the parasympathetic system; however, the transmitters have not been identified. Acetylcholine may be involved because cholinomimetic agents stimulate secretion from some glands and AChE-containing nerves have been identified in the prostate and seminal vesicles. However, VIP-containing nerves are also present in these organs. The seminal vesicles and prostate gland are also sensitive to the levels of sex hormones, as evidenced by the stimulation of glandular size and secretion by testosterone and the reverse effects after the administration of estrogens or castration.

EMISSION-EJACULATION The third phase of the sexual act (orgasm), which is accompanied by emission and ejaculation of semen, involves the coordination of autonomic and somatic reflex mechanisms at different levels of the lumbosacral spinal cord. During the first step in this process (emission), reflex activity in the thoracolumbar sympathetic outflow elicits rhythmic contractions of the smooth muscle of the seminal vesicles, prostate, ductus deferens, and ampulla, resulting in the

Autonomic Systems to the Urinary Bladder and Sexual Organs

ejection of sperm and glandular secretions into the urethra and at the same time a closure of the vesical neck to prevent the backflow of semen into the bladder.59,82,102 Pharmacologic studies have shown that these responses are mediated by the adrenergic transmitters norepinephrine and ATP, interacting with ␣-adrenergic receptors and purinergic receptors, respectively.12,76 Thus surgical interruption of sympathetic nerves or the administration of drugs that either block ␣-adrenergic receptors (e.g., tamsulosin), deplete norepinephrine stores, or block norepinephrine release (e.g., guanethidine) blocks emission.5,76 Seminal emission may also be modulated by activity in cholinergic nerves. For example, physostigmine, an anticholinesterase agent, or exogenous ACh depresses the responses of the ductus deferens and other sex organs to electrical stimulation of the sympathetic nerves. The depression, which is accompanied by a reduction in norepinephrine release, is blocked by atropine. These findings suggest that one action of the cholinergic innervation to the sex organs is a muscarinic suppression of excitatory adrenergic transmission. After emission of semen into the proximal urethra, rhythmic contractions of the bulbocavernosus, ischiocavernous, and paraurethral striated muscles result in ejaculation. The afferent and efferent limbs of the ejaculation reflex are contained in the pudendal nerve (see Table 14–3). The sensations accompanying ejaculation constitute the orgasm.102 Orgasm is not necessarily affected by sympathectomy, provided that the pudendal nerves remain intact. Thus neither efferent fibers in the sympathetic nerves nor contractions of smooth muscles of the seminal vesicles and ductus deferens are essential for the occurrence of orgasm.

CENTRAL REFLEX PATHWAYS In the spinal cord, ascending fibers carrying sensory information from the sex organs seem to lie primarily within the anterolateral tracts. Thus bilateral cordotomy usually diminishes or completely abolishes orgasm. Cordotomy also severely compromises ejaculatory and erectile mechanisms, although the latter may recover over a period of time. With more extensive damage of the spinal cord in paraplegics, when the site of injury is located rostral to T12, ejaculation occurs in a relatively small percentage of patients in comparison to reflexogenic erections, which are readily elicited.34,35,64 This observation is consistent with the greater complexity of spinal reflex pathways underlying ejaculation and indicates a considerable dependence of these pathways on supraspinal coordinating mechanisms. Brain imaging studies using fMRI or PET during sexual arousal induced by erotic visual stimuli have identified several brain regions in both males and females that are activated during sexual stimulation.8,100,105,107,125 PET studies in males detected activation in three general regions:

317

limbic/paralimbic areas (anterior cingulate gyrus, orbitofrontal cortex), the striatum (head of the caudate nucleus), and the posterior hypothalamus. Strong signals specifically associated with penile turgidity were observed in the right subinsular region (claustrum, caudate, cingulate gyrus) using fMRI. A PET study also showed increased blood flow in the right prefrontal cortex during orgasm in males, whereas all other cortical areas showed decreases in blood flow.130 It has been suggested that the activation of paralimbic areas is correlated with the emotional and motivational states associated with sexual arousal, whereas the activation of the anterior cingulate and hypothalamic regions is related to the affective, autonomic, and endocrine responses.8 More recent fMRI studies in males have also detected activation in parietal areas known to be involved in attentional processes.100 The human data correlate well with results from animal experiments, which indicate that connections between the limbic system and the hypothalamus (medial preoptic area [MPOA] and paraventricular nucleus) are essential for the stimulation of the descending autonomic projections to the spinal cord that trigger psychogenic penile erections.45,91 Recent studies in rats have identified a population of lumbar spinal neurons located around the central canal in lamina X and medial lamina VII that participate in the ejaculatory reflex.133 These neurons express NK-1 receptors, receive afferent input from the penis, and project to the thalamus. It was initially hypothesized that these neurons might function primarily in transmitting sensory information from the penis to pleasure centers in the brain. However, it was discovered that the neurons were only activated during ejaculation and not during other components of male sexual behavior. When the neurons were destroyed by intrathecal administration of a toxin (saporin-SSP) that selectively destroys neurons expressing NK-1 receptors, ejaculatory responses were completely abolished without altering other components of copulatory behavior (penile erection, mounting, intromission). It was concluded that the neurons may be part of a spinal ejaculation generator. Pharmacologic studies in animals have implicated many neurotransmitter systems in the central control of sexual function.3,7,26,45,91 The dopaminergic system in the MPOA and the oxytocin and NO systems in the paraventricular nucleus exert major excitatory effects on erection, whereas the serotonergic pathways arising in the raphe nuclei seem to be inhibitory. Elsewhere in the brain noradrenergic, GABAergic, and serotonergic pathways are generally inhibitory, whereas dopaminergic, NO, and oxytocin pathways facilitate sexual function. NO appears to tonically modulate the hypothalamic excitatory circuitry because intracerebroventricular injections of NOS inhibitors prevented the penile erectile responses induced by dopamine agonists and oxytocin.113 It is uncertain whether all of these findings in animals are entirely applicable to man.

318

Function of the Peripheral Nervous System

However, recently apomorphine, a nonselective dopamine receptor agonist that stimulates sexual behavior and penile erection in animals, has been marketed to treat penile erectile dysfunction in patients.26

17.

18.

REFERENCES

19.

1. Andersson, K. E.: Pharmacology of lower urinary tract smooth muscle and penile erection tissues. Pharmacol. Rev. 45:253, 1993. 2. Andersson, K. E.: Pharmacology of penile erection. Pharmacol. Rev. 53:417, 2001. 3. Andersson, K. E.: Erectile physiological and pathophysiological pathways involved in erectile dysfunction. J. Urol. 170:S6, 2003. 4. Andersson, K. E., and Wagner, G.: Physiology of penile erection. Physiol. Rev. 75:191, 1995. 5. Andersson, K.-E., and Wyllie, M. G.: Ejaculatory dysfunction: why all ␣-blockers are not equal. BJU Int. 92:876, 2003. 6. Araki, I., and de Groat, W. C.: Synaptic modulation associated with developmental reorganization of visceral reflex pathways. J. Neurosci. 17:8402, 1997. 7. Argiolas, A., and Melis, M.: Neuromodulation of penile erection: an overview of the role of neurotransmitters and neuropeptides. Prog. Neurobiol. 47:235, 1995. 8. Arrow, B. A., Desmond, J. E., Banner, L. L., et al.: Brain activation and sexual arousal in healthy, heterosexual males. Brain 125:1014, 2002. 9. Athwal, B. S., Berkley, K. J., Brennan, A., et al.: Brain activity associated with the urge to void and bladder fill volume in normal men: preliminary data from a PET study. BJU Int. 84:148, 1999. 10. Athwal, B. S., Berkley, K. J., Hussain, I., et al.: Brain responses to changes in bladder volume and urge to void in healthy men. Brain 124:369, 2001. 11. Bahns, E., Ernsberger, U., Janig, W., and Nelke, A.: Functional characteristics of lumbar visceral afferent fibers from the urinary bladder and urethra in the cat. Eur. J. Physiol. 407:510, 1998. 12. Beck, R. O.: Physiology of male sexual function and effect of neurologic disease. In Fowler, C. J. (ed.): Neurology of Bladder, Bowel, and Sexual Dysfunction. Boston, Butterworth-Heinemann, p. 47, 1999. 13. Betts, C. D.: Bladder and sexual function in multiple sclerosis. In Fowler, C. J. (ed.): Neurology of Bladder, Bowel, and Sexual Dysfunction. Boston, Butterworth-Heinemann, p. 289, 1999. 14. Birder, L. A., Apodaca, G., de Groat, W. C., and Kanai, A. J.: Adrenergic and capsaicin evoked nitric oxide release from urothelium and afferent nerves in urinary bladder. Am. J. Physiol. 275:F226, 1998. 15. Birder, L. A., Barrick, S., Roppolo, J. R., et al.: Feline interstitial cystitis results in mechanical hypersensitivity and altered ATP release from bladder urothelium. Am. J. Physiol. 285:F423, 2003. 16. Birder, L. A., Kanai, A. J., de Groat, W. C., et al.: Vanilloid receptor expression suggests a sensory role for urinary blad-

20.

21.

22.

23.

24.

25.

26. 27. 28.

29.

30.

31.

32.

33.

34.

35.

der epithelial cells. Proc. Natl. Acad. Sci. U.S.A. 98:13396, 2001. Birder, L. A., Nakamura, Y., Kiss, S., et al.: Altered urinary bladder function in mice lacking the vanilloid receptor TRPV1. Nature Neurosci. 5:856, 2002. Birder, L. A., Nealon, M., Kiss, S., et al.: ␤-Adrenoceptor agonists stimulate endothelial nitric oxide synthase in rat urinary bladder urothelial cells. J. Neurosci. 22:8063, 2002. Blok, B. F. M.: Brain control of the lower urinary tract. Scand. J. Urol. Nephrol. 36(Suppl. 210):11, 2002. Blok B. F. M., Sturms, L. M., and Holstege, G.: A PET study on cortical and subcortical control of the pelvic floor musculature in women. J. Comp. Neurol. 389:535, 1997. Blok B. F. M., Sturms, L. M., and Holstege, G.: Brain activity during micturition in women. Brain 121:2033, 1998. Blok, B. F. M., Willemsen, A. T. M., and Holstege, G.: A PET study on the brain control of micturition in humans. Brain 120:111, 1997. Booth, A. M., and de Groat, W. C.: A study of facilitation in vesical parasympathetic ganglia of the cat using intracellular recording techniques. Brain Res. 169:388, 1979. Bors, E., and Comarr, A. E.: Neurological Urology: Physiology of Micturition, Its Neurological Disorders and Sequelae. Baltimore, University Park Press, 1971. Bruschini, H., Schmidt, R. A., and Tanagho, E. A.: Neurologic control of prostatic secretion in the dog. Invest. Urol. 15:288, 1978. Burnett, A. L.: Novel pharmacological approaches in the treatment of erectile dysfunction. World J. Urol. 19:57, 2001. Burnett, A. L.: Nitric oxide regulation of penile erection: biology and therapeutic implications. J. Androl. 23:S20, 2002. Burnett, A. L., Calvin, D. C., Chamness, S. L., et al.: Urinary bladder-urethral sphincter dysfunction in mice with targeted disruption of neuronal nitric oxide synthase models idiopathic voiding disorders in humans. Nature Med. 3:571, 1997. Burnett, A. L., Chang, A. G., Crone, J. K., et al.: Noncholinergic penile erection in mice lacking the gene for endothelial nitric oxide synthase. J. Androl. 23:92, 2002. Burnett, A. L., Lowenstein, C. J., Bredt, D. S., et al.: Nitric oxide: a physiologic mediator of penile erection. Science 257:401, 1992. Burnett, A. L., and Truss, M. C.: Mediators of the female sexual response: pharmacotherapeutic implications. World J. Urol. 20:101, 2002. Burnstock, G.: Purinergic signaling in the lower urinary tract. In Abbracchio, M. P., and Williams, M. (eds.): Handbook of Experimental Pharmacology. Berlin, Springer Verlag, p. 423, 2001. Chancellor, M. B., and Yoshimura, N.: Physiology and pharmacology of the bladder and urethra. In Walsh, P. C., Retik, A. B., Vaughn, E. D., and Wein, A. J. (eds.): Campbell’s Urology. Philadelphia, W. B. Saunders, p. 831, 2002. Chapelle, P. A., Durand, J., and Lacert, P.: Penile erection following complete spinal cord injury in man. Br. J. Urol. 52:216, 1980. Chapelle, P. A., Roby-Brami, A. Yakovleff, A., and Bussel, B.: Neurological correlations of ejaculation and testicular size

Autonomic Systems to the Urinary Bladder and Sexual Organs

36.

37.

38.

39.

40.

41. 42. 43. 44.

45.

46.

47.

48.

49.

50.

51.

52.

in men with a complete spinal cord section. J. Neurol. Neurosurg. Psychiatry 51:197, 1988. Cheng, C-L., Liu, J-C., Chang, S-Y., et al.: Effect of capsaicin on the micturition reflex in normal and chronic spinal cats. Am. J. Physiol. 277:R786, 1999. Chitaley, K., Bivalacqua, T. J., Champion, H. C., et al.: Adeno-associated viral gene transfer of dominant negative RhoA enhances erectile function in rats. Biochem. Biophys. Res. Commun. 298:427, 2002. Chitaley, K., Webb, R. C., and Mills, T.: RhoA/Rho-kinase: a novel player in the regulation of penile erection. Int. J. Impot. Res. 13:67, 2001. Chuang, Y. C., Fraser, M. O., Yu, Y. B., et al.: Analysis of the afferent limb of the vesicovascular reflex using the neurotoxins, resiniferatoxin and capsaicin. Am. J. Physiol. Reg. Integ. Comp. Physiol. 281:R1302, 2001. Cockayne, D. A., Hamilton, S. G., Zhu, Q. M., et al.: Urinary bladder hyporeflexia and reduced pain behaviour in P2X3 deficient mice. Nature 407:1011, 2000. DasGupta, R., and Fowler, C. J.: Bladder, bowel, and sexual dysfunction in multiple sclerosis. Drugs 63:153, 2003. de Groat, W. C.: Spinal cord projections and neuropeptides in visceral afferent neurons. Prog. Brain Res. 67:165, 1986. de Groat, W. C.: Neuropeptides in pelvic afferent pathways. Experientia 43:801, 1987. de Groat, W. C., Araki, I., Vizzard, M. A., et al.: Developmental and injury induced plasticity in the micturition reflex pathway. Behav. Brain Res. 92:127, 1998. de Groat, W. C., and Booth, A. M.: Neural control of penile erection. In Maggi, C. A. (ed.): The Autonomic Nervous System, Vol. 3. Nervous Control of the Urogenital System. London, Harwood Academic Publishers, p. 467, 1993. de Groat, W. C., and Booth, A. M.: Synaptic transmission in pelvic ganglia. In Maggi, C. A. (ed.): The Autonomic Nervous System, Vol. 3. Nervous Control of the Urogenital System. London, Harwood Academic Publishers, p. 291, 1993. de Groat, W. C., Booth, A. M., Milne, R. J., and Roppolo, J. R.: Parasympathetic preganglionic neurons in the sacral spinal cord. J. Auton. Nerv. Syst. 5:23, 1982. de Groat, W. C., Booth, A. M., and Yoshimura, N.: Neurophysiology of micturition and its modification in animal models of human disease. In Maggi, C. A. (ed.): The Autonomic Nervous System, Vol. 3. Nervous Control of the Urogenital System. London, Harwood Academic Publishers, p. 227, 1993. de Groat, W. C., and Kawatani, M.: Enkephalinergic inhibition in parasympathetic ganglia of the urinary bladder of the cat. J. Physiol. (Lond.) 413:13, 1989. de Groat, W. C., Kawatani, M., Hisamitsu, T., et al.: Mechanisms underlying the recovery of urinary bladder function following spinal cord injury. J. Auton. Nerv. Syst. 30:S71, 1990. de Groat, W. C., and Lalley, P. M.: Reflex firing in the lumbar sympathetic outflow to activation of vesical afferent fibers. J. Physiol. (Lond.) 226:289, 1972. de Groat, W. C., Nadelhaft, I., Milne, R. J., et al.: Organization of the sacral parasympathetic reflex pathways to the urinary bladder and large intestine. J. Auton. Nerv. Syst. 3:135, 1981.

319

53. de Groat, W. C., and Saum, W. R.: Sympathetic inhibition of the urinary bladder and of pelvic ganglionic transmission in the cat. J. Physiol. (Lond.) 220:297, 1972. 54. de Groat, W. C., and Saum, W. R.: Synaptic transmission in parasympathetic ganglia in the urinary bladder of the cat. J. Physiol. (Lond.) 256:137, 1976. 55. de Groat, W. C., and Yoshimura, N.: Pharmacology of the lower urinary tract. Annu. Rev. Pharmacol. Toxicol. 41:691, 2001. 56. Delancey, J., Gosling, J., Creed, K., et al.: Gross anatomy and cell biology of the lower urinary tract. In Abrams, P., Cardozo, L., Khoury, S., and Wein, A. (eds.): Incontinence. Paris, Health Publications, Ltd., p. 17, 2002. 57. Diederichs, W., Stief, C. G., Lue, T. F., and Tanagho, E. A.: Norepinephrine involvement in penile detumescence. J. Urol. 143:1264, 1990. 58. Drake, M. J., Hedlund, P., Andersson, K.-E., et al.: Morphology, phenotype, and ultrastructure of fibroblastic cells from normal and neuropathic human detrusor: absence of myofibroblast characteristics. J. Urol. 169:1573, 2003. 59. Elbadawi, A., and Goodman, D. C.: Autonomic innervation of accessory male genital glands. In Spring-Mills, E., and Hafez, E. S. E. (eds.): Male Accessory Sex Glands. Amsterdam, Elsevier, p. 101, 1980. 60. Ferguson, D. R., Kennedy, I., and Burton, T. J.: ATP is released from rabbit urinary bladder cells by hydrostatic pressure changes—a possible sensory mechanism? J. Physiol. (Lond.) 505:503, 1997. 61. Fowler, C. J., Jewekes, D., McDonald, W. I., et al.: Intravesical capsaicin for neurogenic bladder dysfunction. Lancet 339:1239, 1992. 62. Goldstein, I., Graziottin, J. R., Heiman, J. R., et al.: Sexual dysfunction in females. In Jardin, A., Wagner, G., Khoury, S., et al. (eds.): Erectile Dysfunction. Paris, Health Publications, Ltd., pp. 507–556, 2000. 63. Goldstein, I., Lue, T. F., Padma-Nathan, H., et al.: Oral sildenafil in the treatment of erectile dysfunction. Sildenafil Study Group. N. Engl. J. Med. 338:1397, 1998. 64. Grossiord, A., Chapelle, P. A., Lacert, P. H., et al.: Le segment medullare lesionnel chez le paraplegique. Rev. Neurol. (Paris) 134:729, 1978. 65. Habler, H. J., Janig, W., and Koltzenberg, M.: Activation of unmyelinated afferent fibers by mechanical stimuli and inflammation of the urinary bladder. J. Physiol. (Lond.) 425:545, 1990. 66. Hashitani, H., Fukuta, H., Dickens, E. J., and Suzuki, H.: Cellular mechanisms of nitric oxide-induced relaxation of corporeal muscle in the guinea pig. J. Physiol. (Lond.) 538:573, 2002. 67. Hedlund, P., Aszodi, A., Pfeifer, A., et al.: Erectile dysfunction in cyclic GMP-dependent kinase I-deficient mice. Proc. Nat. Acad. Sci. U.S.A. 97:2349, 2000. 68. Ho, K. M., Ny, L., McMurray, G., et al.: Co-localization of carbon monoxide and nitric oxide synthesizing enzymes in the human urethral sphincter. J. Urol. 161:1968, 1999. 69. Hurt, J. K., Musicki, B., Palese, M. A., et al.: Akt-dependent phosphorylation of endothelial nitric-oxide synthase mediates penile erection. Proc. Natl. Acad. Sci. U.S.A. 99:1461, 2002.

320

Function of the Peripheral Nervous System

70. Jankowski, R. J., Vorp, D. A., Prantil, R. L., et al.: Development of a system for studying urethral biomechanical function. Am. J. Physiol. 286:F225, 2004. 71. Kakizaki, H., Fraser, M. O., and de Groat, W. C.: Reflex pathways controlling urethral striated and smooth muscle function in the male rat. Am. J. Physiol. 272:R1647, 1997. 72. Kamo, I., Torimoto, K., Chancellor, M. B., et al.: Urethral closure mechanisms under sneeze-induced stress condition in rats: a new animal model for evaluation of stress urinary incontinence. Am. J. Physiol. 285:R356, 2003. 73. Kawatani, M., Shioda, S., Nakai, Y., et al.: Ultrastructural analysis of enkephalinergic terminals in parasympathetic ganglia innervating the urinary bladder of the cat. J. Comp. Neurol. 288:81, 1989. 74. Keast, J. R., and de Groat, W. C.: Segmental distribution and peptide content of primary afferent neurons innervating the urogenital organs and colon of male rats. J. Comp. Neurol. 319:615, 1992. 75. Keast, J., Kawatani, M., and de Groat, W. C.: Sympathetic modulation of cholinergic transmission in cat vesical ganglia is mediated by ␣1 and ␣2 adrenoceptors. Am. J. Physiol. 258:R44, 1990. 76. Kedia, K., and Markland, C.: The effect of pharmacological agents on ejaculation. J. Urol. 114:569, 1975. 77. Kihara, K., and de Groat, W. C.: Sympathetic efferent pathways projecting bilaterally to the vas deferens in the rat. Anat. Rec. 248:291, 1997. 78. Kruse, M. N., Bray, L. A., and de Groat, W. C.: Influence of spinal cord injury on the morphology of bladder afferent and efferent neurons. J. Auton. Nerv. Syst. 54:215, 1995. 79. Lavelle, J. P., Meyers, S., Ruiz, G., et al.: Urothelial pathophysiological changes in feline interstitial cystitis: a human model. Am. J. Physiol. 278: F540, 2000. 80. Lee, H. Y., Bardini, M., and Burnstock, G.: Distribution of P2X3 in the urinary bladder and ureter in the rat. J. Urol. 163:2002, 2000. 81. Lewis, S. A.: Everything you wanted to know about the bladder epithelium but were afraid to ask. Am. J. Physiol. 278:F867, 2000. 82. Lundberg, P. O., Brockett, N. L., Denys, P., et al.: Neurological disorders: erectile and ejaculatory dysfunction. In Jardin, A., Wagner, G., Khoury, S., et al. (eds.): Erectile Dysfunction. Paris, Health Publications, Ltd., p. 591, 2000. 83. Maggi, C. A.: The dual sensory and efferent functions of the capsaicin-sensitive primary sensory neurons in the urinary bladder and urethra. In Maggi, C. A. (ed.): The Autonomic Nervous System, Vol. 3. Nervous Control of the Urogenital System. London, Harwood Academic Publishers, p. 383, 1993. 84. Mallory, B., Steers, W. D., and de Groat, W. C.: Electrophysiological study of micturition reflexes in the rat. Am. J. Physiol. 257:R410, 1989. 85. Mallory, B. S., Roppolo, J. R., and de Groat, W. C.: Pharmacological modulation of the pontine micturition center. Brain Res. 546:310, 1991. 86. Marson, L., and McKenna K.: CNS cell groups involved in the control of the ischiocavernous and bulbospongiosus muscles: a transneural tracing study using pseudorabies virus. J. Comp. Neurol. 374:161, 1996.

87. Masters, W. H., and Johnson, V. E.: Human Sexual Response. Boston, Little Brown & Co., 1966. 88. Matsui, M., Motomura, D., Fujikawa, T., et al.: Mice lacking M2 and M3 muscarinic acetylcholine receptors are devoid of cholinergic smooth muscle contractions but still viable. J. Neurosci. 22:10627, 2002. 89. Matsui, M., Motomura, D., Karasawa, H., et al.: Multiple functional defects in peripheral autonomic organs in mice lacking muscarinic acetylcholine receptor gene for the M3 subtype. Proc. Nat. Acad. Sci. U.S.A. 97:9579, 2000. 90. Matsuura, S., Kakizaki, H., Mitsui, T., et al.: Human brain region response to distension or cold stimulation of the bladder: a positron emission tomography study. J. Urol. 168:2035, 2002. 91. McKenna, K.: Central control of penile erection. Int. J. Impot. Res. 10(Suppl.):S25, 1998. 92. Melman, A., Henry, D. P., Felten, D. F., and O’Connor, B. L.: Effect of diabetes upon penile sympathetic nerves in impotent patients. South. Med. J. 73:307, 1980. 93. Mills, T. M., Chitaley, K., and Lewis, R. W.: Vasoconstrictors in erectile physiology. Int. J. Impot. Res. 5(Suppl. 13):S29, 2001. 94. Miura, A., Kawatani, M., and de Groat, W. C.: Excitatory synaptic currents in lumbosacral parasympathetic preganglionic neurons elicited from the lateral funiculus. J. Neurophysiol. 86:1587, 2001. 95. Miura, A., Kawatani, M., and de Groat, W. C.: Excitatory synaptic currents in lumbosacral parasympathetic preganglionic neurons evoked by stimulation of the dorsal commissure. J. Neurophysiol. 89:382, 2003. 96. Morgan, C., Nadelhaft, I., and de Groat, W. C.: Location of bladder preganglionic neurons within the sacral parasympathetic nucleus of the cat. Neurosci. Lett. 14:189, 1979. 97. Morgan, C., Nadelhaft, I., and de Groat, W. C.: The distribution of visceral primary afferents from the pelvic nerve within Lissauer’s tract and the spinal gray matter and its relationship to the sacral parasympathetic nucleus. J. Comp. Neurol. 201:415, 1981. 98. Morrison, J., Steers, W. D., Brading, A., et al.: Neurophysiology and neuropharmacology. In: Abrams, P., Cardozo, L., Khoury, S., and Wein, A. (eds.): Incontinence. Paris, Health Publications, Ltd., p. 83, 2002. 99. Morrison, J., Wen, J., and Kibble, A.: Activation of pelvic afferent nerves from the rat bladder during filling. Scand. J. Urol. Nephrol. Suppl. 201:73, 1999. 100. Mouras, H., Stoleru, S., Bittoun, J., et al.: Brain processing of visual sexual stimuli in healthy men: a functional magnetic resonance imaging study. Neuroimage 20:855, 2003. 101. Nadelhaft, I., and Vera, P. L.: Central nervous system neurons infected by pseudorabies virus injected into the rat urinary bladder following unilateral transection of the pelvic nerve. J. Comp. Neurol. 359:443, 1995. 102. Newman, H. F., Reiss, H., and Northup, J. C.: Physical basis of emission, ejaculation, and orgasm in the male. Urology 19:341, 1982. 103. O’Reilly, B. A., Kosaka, A. H., Knight, G. F., et al.: P2X receptors and their role in female idiopathic detrusor instability. J. Urol. 167:157, 2002. 104. Palea, S., Artibani, W., Ostardo, E., et al.: Evidence for purinergic neurotransmission in the human urinary bladder affected by interstitial cystitis. J. Urol. 150:2007, 1993.

Autonomic Systems to the Urinary Bladder and Sexual Organs 105. Park, K., Kang, H. K., Seo, J. J., et al.: Blood-oxygenationlevel-dependent functional magnetic resonance imaging for evaluating cerebral regions of female sexual arousal response. Urology 57:1189, 2001. 106. Ralevic, V., and Burnstock, G.: Receptors for purines and pyrimidines. Physiol. Rev. 50:413, 1998. 107. Redoute, J., Stoleru, S., Gregoire, M. C., et al.: Brain processing of visual sexual stimuli in human males. Human Brain Mapping 11:162, 2000. 108. Rong, W., Spyer, K. M., and Burnstock, G.: Activation and sensitization of low and high threshold afferent fibres mediated by P2X receptors in the mouse urinary bladder. J. Physiol. (Lond.) 541:591, 2002. 109. Root, W. S., and Bard, P.: The mediation of feline erection through sympathetic pathways with some remarks on sexual behavior after deafferentation of the genitalia. Am. J. Physiol. 151:80, 1947. 110. Sachs, B. D., and Meisel, R. L.: The physiology of male sexual behavior. In Knobil, E., and Neill, J. (eds.): The Physiology of Reproduction. New York, Raven, p. 1393, 1988. 111. Saenz de Tejada, I., Cadavid, N. G., Heaton, J., et al.: Anatomy, physiology, and pathophysiology of erectile function. In Jardin, A., Wagner, G., Khoury, S., et al. (eds.): Erectile Dysfunction. Paris, Health Publications, Ltd., p. 65, 2000. 112. Sakakibara, R., and Fowler, C. J.: Cerebral control of bladder, bowel, and sexual function and effects of brain disease. In Fowler, C. J. (ed.): Neurology of Bladder, Bowel, and Sexual Dysfunction. Boston, Butterworth-Heinemann, p. 229, 1999. 113. Sato, Y., Zhao, W., and Christ, G. J.: Central modulation of the NO/cGMP pathway affects the MPOA-induced intracavernous pressure response. Am. J. Physiol. Reg. Integ. Comp. Physiol. 281:R269, 2001. 114. Seki, S., Sasaki, K., Fraser, M. O., et al.: Immunoneutralization of nerve growth factor in the lumbosacral spinal cord reduces bladder hyperreflexia in spinal cord injured rats. J. Urol. 168:2269, 2002. 115. Semans, J. H., and Langworthy, O. R.: Observations on the neurophysiology of sexual function in the male cat. J. Urol. 40:836, 1938. 116. Sengupta, J. N., and Gebhart, G. F.: Mechanosensitive properties of pelvic afferent nerve fibers innervating the urinary bladder of the rat. J. Neurophysiol. 72:2420, 1994. 117. Shea, V. K., Cai, R., Crepps, B., et al.: Sensory fibers of the pelvic nerve innervating the rat’s bladder. J. Neurophysiol. 84:1924, 2000. 118. Somogyi, G. T., Tanowitz, M., and de Groat, W. C.: Prejunctional facilitatory ␣1-adrenoceptors in the rat urinary bladder. Br. J. Pharmacol. 114:1710, 1995. 119. Somogyi, G. T., Tanowitz, M., Zernova, G., and de Groat, W. C.: M1 muscarinic receptor facilitation of ACh and noradrenaline release in the rat urinary bladder is mediated by protein kinase C. J. Physiol. (Lond.) 496:245, 1996. 120. Somogyi, G. T., Zernova, G. V., Tanowitz, M., and de Groat, W. C.: Role of L and N type Ca⫹2 channels in muscarinic receptor mediated facilitation of ACh and noradrenaline release in the rat urinary bladder. J. Physiol. (Lond.) 499:645, 1997. 121. Somogyi, G. T., Zernova, G. V., Yoshiyama, M., et al.: Frequency dependence of muscarinic facilitation of trans-

122.

123.

124.

125.

126.

127.

128.

129. 130.

131. 132.

133.

134.

135.

136.

137.

138.

321

mitter release in urinary bladder strips from neurally intact or chronic spinal cord transected rats. Br. J. Pharmacol. 125:241, 1998. Steers, W. D., Ciambotti, J., Erdman, S., and de Groat, W. C.: Morphological plasticity in efferent pathways to the urinary bladder of the rat following urethral obstruction. J. Neurosci. 10:1943, 1990. Steers, W. D., Ciambotti, J., Etzel, B., et al.: Alterations in afferent pathways from the urinary bladder of the rat in response to partial urethral obstruction. J. Comp. Neurol. 310:1, 1991. Steers, W. D., Mallory, B., and de Groat, W. C.: Electrophysiological study of neural activity in penile nerve of the rat. Am. J. Physiol. 254:R989, 1988. Stoleru, S., Gregoire, M. C., Gerard, D., et al.: Neuroanatomical correlates of visually evoked sexual arousal in human males. Arch. Sexual Behav. 28:1, 1999. Sugaya, K., Roppolo, J. R., Yoshimura, N., et al.: The central neural pathways involved in micturition in the neonatal rats as revealed by the injection of pseudorabies virus into the bladder. Neurosci. Lett. 223:197, 1997. Sui, G. P., Wu, C., and Fry, C. H.: Electrical characteristics of suburothelial cells isolated from the human bladder. J. Urol. 171:938, 2004. Sun, Y., Keay, S., de Deyne, P. G., and Chai, T. C.: Augmented stretch activated adenosine triphosphate release from bladder uroepithelial cells in patients with interstitial cystitis. J. Urol. 166:1951, 2001. Tanagho, E. A., and Miller, E. R.: Initiation of voiding. Br. J. Urol. 42:175, 1970. Tiihonen, J., Kuikka, J., Kupila, J., et al.: Increase in cerebral blood flow of right prefrontal cortex in man during orgasm. Neurosci. Lett. 170:241, 1994. Torrens, M., and Morrison, J. F. B.: The Physiology of the Lower Urinary Tract. Berlin, Springer-Verlag, 1986. Tran, L. V., Somogyi, G. T., and de Groat, W. C.: Inhibitory effects of neuropeptide Y on cholinergic and adrenergic transmission in the rat urinary bladder and urethra. Am. J. Physiol. 266:R1411, 1994. Truitt, W. A., and Coolen, L. M.: Identification of a potential ejaculation generator in the spinal cord. Science 297:1566, 2002. Truschel, S. T., Wang, E., Ruiz, G., et al.: Stretch-related exocytosis/endocytosis in bladder umbrella cells. Mol. Biol. Cell 13:830, 2002. Vizzard, M. A., Erickson, V. L., Card, J. P., et al.: Transneuronal labeling of neurons in the adult rat brain and spinal cord after injection of pseudorabies virus into the urethra. J. Comp. Neurol. 355:629, 1995. Vlaskovoska, M., Kasakov, L., Rong, W., et al.: P2X3 knockout mice reveal a major sensory role for urothelially released ATP. J. Neurosci. 21:5670, 2001. Wein, A. J.: Pathophysiology and categorization of voiding dysfunction. In Walsh, P. C., Retik, A. B., Vaughn, E. D., and Wein, A. J. (eds.): Campbell’s Urology. Philadelphia: W. B. Saunders, p. 887, 2002. Yokoyama, O., Yoshimura, M., Namiki, M., and de Groat, W. C.: Role of the forebrain in bladder hyperactivity following cerebral infarction in the rat. Exp. Neurol. 163:469, 2000.

322

Function of the Peripheral Nervous System

139. Yokoyama, O., Yoshiyama, M., Namiki, M., and de Groat, W. C.: Changes in dopaminergic and glutamatergic excitatory mechanisms of the micturition reflex after middle cerebral artery occlusion in conscious rats. Exp. Neurol. 173:129, 2002. 140. Yoshimura, N., and de Groat, W. C.: Plasticity of Na⫹ channels in afferent neurons innervating rat urinary bladder following spinal cord injury. J. Physiol. (Lond.) 503:269, 1997. 141. Yoshimura, N., and de Groat, W. C.: Increased excitability of afferent neurons innervating rat urinary bladder following chronic bladder inflammation. J. Neurosci. 19:4644, 1999. 142. Yoshimura, N., Kuno, S., Chancellor, M. B., et al.: Dopaminergic mechanisms underlying bladder hyperactivity in rats with a unilateral 6-hydroxydopamine (6-OHDA) lesion of the nigrostriatal pathway. Br. J. Pharmacol. 139:1425, 2003. 143. Yoshimura, N., Mizuta, E., Yoshida, O., and Kuno, S.: Therapeutic effects of dopamine D1/D2 receptor

144.

145.

146.

147.

agonists on detrusor hyperreflexia in 1-methyl-4-phenyl1,2,3,6-tetrahydropyridine-lesioned parkinsonian monkeys. J. Pharmacol. Exp. Ther. 286:228, 1998. Yoshimura, N., Seki, S., Erickson, K. A., et al.: Histological and electrical properties of rat dorsal root ganglion neurons innervating the lower urinary tract. J. Neurosci. 23:4355, 2003. Yoshimura, N., Seki, S., Novakovic, S. D., et al.: The role of the tetrodotoxin-resistant sodium channel Nav1.8 (PN3/SNS) in a rat model of visceral pain. J. Neurosci. 21:8690, 2001. Yoshimura, N., White, G., Weight, F. F., and de Groat, W. C.: Different types of Na⫹ and A-type K⫹ currents in dorsal root ganglion neurons innervating the rat urinary bladder. J. Physiol. (Lond.) 494:1, 1996. Zoubek, J. A., Somogyi, G. T., and de Groat, W. C.: A comparison of inhibitory effects of neuropeptide Y on rat urinary bladder, urethra and vas deferens. Am. J. Physiol. 265:R537, 1993.

15 Sympathetic Nerves and Control of Blood Vessels to Human Limbs MICHAEL J. JOYNER, PAUL M. VANHOUTTE, JOHN R. HALLIWILL, AND JOHN T. SHEPHERD

Baroreflexes Direct Measurement of Sympathetic Traffic in Humans via Microneurography Characteristics of Peripheral Sympathetic Traffic in Humans Baroreceptor Control of MSNA MSNA Responses to Orthostatic Challenge MSNA Responses to Chemoreceptor Activation Central Chemoreflexes and Autonomic Function

MSNA and Autonomic Responses to Exercise Sensory Feedback from Active Muscles Baroreflex Resetting during Exercise Interactions between Sympathetic Vasoconstriction and Metabolic Vasodilatation ␣-Receptor Subtypes and Sympathetic Control of Blood Flow in Active Muscles Exercise: Summary

Arterial blood pressure is the key regulated variable in the cardiovascular system, and the autonomic nervous system has a central role in this regulation. In response to a variety of behavioral situations, blood pressure can rise (e.g., exercise) or fall (sleep), but for each situation arterial pressure is usually maintained in a fairly narrow range. Central to the regulation of arterial pressure is the balance that is maintained between local factors such as metabolites from exercising muscle that can cause vasodilatation and the actions of the sympathetic vasoconstrictor nerves.77 In some species there are sympathetic dilator nerves to skeletal muscle, but their existence in humans is now doubtful.35 Additionally, during increases in core temperature, sympathetic vasodilator nerves to the skin evoke marked cutaneous vasodilatation.53 In previous editions of Peripheral Neuropathy, detailed structural, pharmacologic, and biochemical information concerning the autonomic nerves has been presented.77 In this chapter, we focus on how the sympathetic nerves innervating the blood vessels of the human limbs respond to common stimuli.

Autonomic and MSNA Responses to Mental Stress Neurally Mediated Dilatation in Human Skeletal Muscle Aging, Hypertension, and Congestive Heart Failure Autonomic Responses to Body Heating Cutaneous Vascular Responses to Changes in Temperature Summary

BAROREFLEXES For the autonomic nervous system to regulate arterial pressure, the brain must receive information about the level of the arterial pressure. This is accomplished primarily through sensory nerves located in the carotid sinus and aortic arch. Mechanosensitive afferent nerves in the vessel walls at these locations respond to physical deformation in the blood vessels associated with changes in arterial pressure.10,20,45 The carotid receptors are innervated by the glossopharyngeal nerve and the aortic receptors by the vagus nerve (cranial nerves IX and X, respectively). Some of the sensory nerves respond to tonic levels of stretch, while others are more active when the receptors deform with each heart beat.10,20 The baroreceptor afferents are initially integrated at the nucleus of the solitary tract of the brainstem, and the information relayed by the afferents is processed and “sent” to other areas involved in cardiovascular control.20,45 In responses to increases or decreases in arterial pressure, predictable changes in heart rate and efferent 323

324

Function of the Peripheral Nervous System

sympathetic outflow occur. When blood pressure falls, there is an increase in heart rate as a result of vagal withdrawal. When blood pressure increases, there is increased vagal activity and heart rate slows. Likewise, when arterial pressure falls, there is an increase in efferent sympathetic traffic that evokes vasoconstriction to “defend” arterial pressure. By contrast, when pressure increases, sympathetic outflow is suppressed in an effort to reduce the pressure to a normal level. In general, these responses can be described by a sigmoidal curve and the “resting” blood pressure is on the steep linear region of this curve. Over a wide range of pressure, there is a linear relationship (gain) between pressure and either heart rate or sympathetic outflow. However, when blood pressure is extremely low, it is below the “threshold” for normal baroreflex function. By contrast, when pressure is extremely high, the afferent nerves are maximally activated and the system is “saturated.”10,20,45 Figure 15–1 is a schematic of a typical baroreflex response curve.

DIRECT MEASUREMENT OF SYMPATHETIC TRAFFIC IN HUMANS VIA MICRONEUROGRAPHY Microneurography consists of placing small tungsten microelectrodes in peripheral nerves of conscious humans. This technique was initially developed in the 1960s to study the function of motor nerves and skeletal muscle afferents in humans.25,75 During that time, other signals were noted that could be activated by breath holding or the Valsalva maneuver or eliminated by ganglionic blockade, and it was determined that it was also possible to record from multiunit bundles of sympathetic nerves that innervate skeletal muscle or skin.16,17 Although a number of important observations were made using this technique in the 1970s, as the technique became more widespread, a host of findings on almost every aspect of cardiovascular control in humans have been made since the early 1980s. Figure 15–2 depicts arterial blood pressure, breathing, and muscle sympathetic nerve activity (MSNA) from a human volunteer. It is interesting to note that, as blood pressure falls, MSNA increases, and as blood pressure increases, MSNA falls.

Sympathetic nerve activity or heart rate

Threshold

Linear region 150 mm Hg

Rest

100

Saturation 2s

Arterial pressure

FIGURE 15–1 Schematic representation of a baroreflex curve in humans. As blood pressure increases (x axis), there is a progressive fall in both heart rate and efferent sympathetic traffic (y axis). As pressure falls, there are reflex increases in heart rate in efferent sympathetic traffic. These changes are thought to be governed primarily by arterial baroreflexes. The increase in heart rate is due primarily to vagal withdrawal to values of approximately 100 beats/min and thereafter is the result of activation of sympathetic nerves to the heart. The “linear” portion of the stimulus response curve has been described as the gain of the arterial baroreflex, while the flat portion on the upper left is known as the “threshold” and the flat portion in the lower right corner, where further increases in pressure do not evoke further reductions in heart rate of sympathetic traffic, is known as the “saturation point.”

50

FIGURE 15–2 Classic recording of muscle sympathetic nerve activity (MSNA) and mean arterial pressure in a conscious human. In most humans there are continuous subtle changes in arterial pressure associated with either intrinsic cardiac rhythms or breathing. Under these circumstances, as blood pressure increases, baroreflexes are activated and muscle sympathetic nerve activity is reduced. Likewise, when pressure falls, MSNA increases. There is an approximately 1.3-s time lag between changes in arterial pressure and changes in MSNA to account for the time it takes for the baroreceptors to transduce the signal, for the various neural conduction delays, and for central integration of the signal along with the efferent response measured in the lower extremity. Note the large burst associated with aberrant heartbeat in the middle of the record. (From Wallin, B. G., and Fagius, J.: The sympathetic nervous system in man—aspects derived from microelectrode recordings. Trends Neurosci. 9:63, 1986, with permission.)

Sympathetic Nerves and Control of Blood Vessels to Human Limbs

CHARACTERISTICS OF PERIPHERAL SYMPATHETIC TRAFFIC IN HUMANS The most common site for recording peripheral sympathetic traffic in humans is the peroneal nerve. At this site it is possible to record MSNA and skin sympathetic nerve activity (SSNA).17,24 MSNA and SSNA have different characteristics. MSNA is primarily vasoconstrictive under baroreflex control and is “pulse synchronous,” with only one burst seen for each heartbeat. Additionally, MSNA is responsive to chemoreceptor stimuli and rises during hypercapnia or hypoxia.17,75,82,83 Finally, arousal stimuli such as a loud, unexpected noise do not evoke a rise in MSNA.17 In young, healthy, supine humans, the level of resting MSNA typically ranges from 10 to 30 bursts/min. Interestingly, there is little correlation between resting MSNA and baseline values of arterial pressure in normotensive subjects.66,67,83 Skin sympathetic nerve traffic is more complex because it is directed at blood vessels, sweat glands, and piloerector organs. In addition to vasoconstrictor nerves, there is neurally mediated vasodilatation in the skin. In a thermoneutral environment, SSNA is not pulse synchronous, is not under obvious baroreceptor or chemoreceptor control, and is responsive to arousal stimuli.16,75,82 Additionally, the characteristic sharp peaks seen in records of MSNA are less obvious in records of SSNA.24 The differences in MSNA and SSNA seen at rest and during other stimuli also highlight the idea that the sympathetic nervous system does not function in an “all

FIGURE 15–3 Original record of a “modified” Oxford test to assess integrated baroreflex function in conscious humans. Arterial pressure, MSNA, and heart rate (electrocardiogram) are being measured. In this technique, a 100-␮g intravenous bolus dose of nitroprusside is given to lower arterial pressure. One minute later, a 150-␮g bolus of phenylephrine is given intravenously to raise arterial pressure. In response to the fall in pressure, the MSNA bursts increase (middle tracing), and as pressure begins to rise as a result of the phenylephrine bolus, MSNA is suppressed. Likewise, heart rate increases as pressure falls and slows as pressure increases.

or none” unified fashion, and that sympathetic outflow to different organs and regions can differ.

BARORECEPTOR CONTROL OF MSNA Several techniques are available to assess baroreceptor control of MSNA in humans. One of the most useful is the “modified Oxford technique.” In this technique, beat-tobeat measurements of arterial pressure are made and MSNA is recorded during an intravenous bolus infusion of the vasodilator drug sodium nitroprusside followed 1 minute later by a bolus infusion of the vasoconstrictor phenylephrine.19,26,56 This paradigm permits measurement of the heart rate and MSNA responses to a wide range of pressures, and allows for the calculation of baroreflex function curves for both variables. Figure 15–3 is a record of a modified Oxford trial. One limitation with the technique is that the systemic infusion of drugs causes changes in pressure of both the carotid and aortic baroreceptors, and perhaps the cardiopulmonary receptors as well, so determining the relative contribution of each receptor population to the total response is not possible. It also appears that MSNA but not SSNA is under tonic inhibition by arterial baroreflexes. In conscious humans, administration of local anesthesia to the glossopharyngeal and vagus nerves has been used to eliminate afferent traffic from the carotid and aortic receptors.21 Figure 15–4 demonstrates that this evokes large increases in MSNA and arterial pressure with little change in SSNA.

“Modified Oxford” method for baroreflex assessment

Nitroprusside 100 ug bolus iv

Phenylephrine 150 ug bolus iv

60 s

Arterial pressure

Muscle sympathetic nerve activity

ECG

325

326

Function of the Peripheral Nervous System

1 Control

2 After atropine

3 Unilateral block

ECG

MSA SSA 200 mm Hg BP

100 0 5s

4 Bilateral block

5 Bilateral block

6 30 min after bilateral block

ECG

MSA

SSA

200 mm Hg

BP

100 30 s

110 s

0 5s

FIGURE 15–4 Individual record of the effects of bilateral carotid sinus nerve block on MSNA, SSNA, arterial pressure, and heart rate in a conscious human. 1, Control conditions. 2, The effects of atropine. Most notably, atropine caused heart rate to increase and blood pressure to rise slightly but had little impact on MSNA or SSNA. 3, Effect of unilateral glossopharyngeal nerve block on these variables. MSNA begins to rise, SSNA is unchanged, and arterial pressure continues to increase. 4, Bilateral nerve block is associated with a huge increase in MSNA, no change in SSNA, and a large increase in blood pressure. 5, This response is sustained. 6, The response “wears off” as the local anesthetic block recedes. This figure demonstrates that MSNA, but not SSNA, is under tonic baroreceptor control in humans. It also demonstrates the key role that baroreceptors play in suppressing sympathetic traffic and modulating mean arterial pressure. (From Fagius, J., Wallin, B. G., Sundlof, G., et al.: Sympathetic outflow in man after anaesthesia of the glossopharyngeal and vagus nerves. Brain 108:423, 1985, with permission.)

MSNA RESPONSES TO ORTHOSTATIC CHALLENGE One of the most common challenges to the autonomic nervous system in humans is assumption of the upright posture. As a result of gravity, there is a shift of approximately 0.5 to 1.0 L of blood from the heart, thorax, and abdomen to the dependent veins in the legs. Additionally, the head is now above heart level and the effective perfusion pressure to the carotid sinus and brain is reduced. In response to these changes, there is typically a 50% to 100% increase in MSNA that is sustained9 (Fig. 15–5). Additionally, if the orthostatic

stress is prolonged, there is a slow rise in MSNA above the immediate increase. This continues up to the point of syncope, when there is the sudden onset of sympathetic silence that occurs as the subject faints.64,84 An individual record of such a response is shown in Figure 15–6. The mechanisms that evoke the sympathetic silence with syncope are poorly understood. One common idea is that the heart becomes relatively empty, and that with vigorous contractions sympathoinhibitory afferents (normally active when the heart is “full”) are paradoxically stimulated and suppress sympathetic outflow. However, fainting can be seen in patients after heart transplantation who have

Sympathetic Nerves and Control of Blood Vessels to Human Limbs

Heart rate/min

327

denervated ventricles.23 It is also possible that hypotension in the carotid vessels or aortic vessels evokes paradoxical stimulation of arterial baroreceptor afferents and suppresses sympathetic outflow. However, experimental evidence for both mechanisms is far from clear-cut.

Bursts/min

120

100

80

80

60

60

40

40

20

20

0

MSNA RESPONSES TO CHEMORECEPTOR ACTIVATION

0 A

B

C

A

B

C

FIGURE 15–5 Effects of standing on heart rate and MSNA in conscious humans. Left, Heart rate increased from approximately 65 beats/min to approximately 80 to 90 beats/min. Right, The burst of MSNA essentially doubled from 30 to approximately 60/min. (From Burke, D., Sundlof, G., and Wallin, G.: Postural effects on muscle nerve sympathetic activity in man. J. Physiol. [Lond.] 272:399, 1977, with permission.)

Symp activity 200 mm Hg Blood pressure

100

Heart rate

0 90 Beats/min 60

Finger pleth 10 s

FIGURE 15–6 MSNA tracing from an individual subject during syncope. The tracing shows sympathetic activity, blood pressure, heart rate, and finger plethysmographic blood flow. The star indicates the time of the faint. The key observation is that, before the faint, there was a progressive rise of MSNA, which suddenly became silent at the time of the faint. This occurred concurrently with the fall in blood pressure and the marked bradycardia associated with vasovagal syncope. These observations demonstrate that a profound sympathetic silence is normally associated with standard vasovagal syncope. (From Wallin, B. G., and Sundlof, G.: Sympathetic outflow to muscles during vasovagal syncope. J. Auton. Nerv. Syst. 6:287, 1982, with permission.)

The peripheral chemoreceptors located in the carotid and aortic bodies are activated by a fall in O2, and to a lesser degree by a rise in CO2 or acidity.22 These receptors send signals to the medulla via cranial nerves IX and X, and, like the baroreceptor afferents, synapse initially in the nucleus of the tractus solitarius. The best known effect of peripheral chemoreceptor activation is hyperpnea, but these pathways also affect efferent sympathetic and parasympathetic outflow. Studies in animals have documented that exposure of the peripheral chemoreceptors to a fall in partial pressure of O2 results in a rise in sympathetic nerve activity that can be measured in peripheral sympathetic nerves and a bradycardia that is dependent on increased parasympathetic traffic to the heart. Likewise, studies in humans have observed increases in sympathetic vasoconstrictor outflow to muscle vascular beds during hypoxia.54,58,65 Figure 15–7 shows an example of the sympathetic responses to chemoreceptor activation. It should be noted that these primary chemoreflex responses are often modified, or overridden, by the effects of hyperpnea on autonomic outflow and cardiovascular function.57 For instance, in animals the bradycardia produced by peripheral chemoreceptor stimulation is often abolished, or even converted to tachycardia, by ongoing changes in ventilation.22 This may account for the rise in heart rate frequently reported during studies on hypoxia in humans.54,58,65 However, one series of studies suggests that increased ventilation or changes in breathing patterns in humans modify the pattern of sympathetic nerve activity within breaths but do not change the overall level of sympathetic outflow.60,61 This suggests that the sympathoexcitation produced by hypoxia in humans is a direct effect of peripheral chemoreflex activation.

Central Chemoreflexes and Autonomic Function The central chemoreceptors, located on the ventral aspect of the medulla, are activated by an increase in CO2 or acidity.4 The best known effects of central chemoreceptor activation are increases in ventilation. Indeed, studies have suggested that central chemoreceptor activation, depending on conditions, mediates 50% to 90% of the ventilatory response to hypercapnia.4 Under hyperoxic conditions, the response to hypercapnia is almost exclusively mediated by the central

328

Function of the Peripheral Nervous System

Control

8% Oxygen

Peroneal neurogram ECG Respiration Total MSNA 170 units

Burst frequency 17 min-1

Total MSNA 650 units

Burst frequency 50 min-1

Heart rate 52 beats min-1

Mean pressure 95 mm Hg

Heart rate 73 beats min-1

Mean pressure 94 mm Hg

FIGURE 15–7 Effects of systemic hypoxia, MSNA, heart rate, and respiration in a conscious human. Breathing 8% oxygen (right side) increases MSNA via the peroneal nerve. Heart rate and respiration also increased.

chemoreceptors.30 These pathways also affect the level of activity in the sympathetic outflow tracts, but this aspect of chemoreflex control is poorly understood. Studies in animals have documented that exposure of the central chemoreceptors to a rise in partial pressure of CO2 results in a rise in sympathetic nerve activity that can be measured in peripheral sympathetic nerves.23 Analogous studies in humans have shown that hyperoxic hypercapnia produces a rise in muscle sympathetic nerve activity and heart rate46,65; however, the vagal withdrawal that mediates the rise in heart rate may be secondary to changes in arterial pressure that were produced by hyperpnea.

MSNA AND AUTONOMIC RESPONSES TO EXERCISE Exercise is another frequent stressor that confronts the autonomic nervous system, and the increases in heart rate and blood pressure that occur with exercise are obvious examples of altered autonomic responses. In general, the increase in heart rate from resting to approximately 100 beats/min is largely due to withdrawal of vagal tone, with subsequent increases to maximum largely the result of increased sympathetic traffic.49,55,78,79 Three main mechanisms are thought to control the autonomic responses to exercise. These include central command, sensory feedback from contracting muscles, and baroreflex resetting. The behavior of MSNA and SSNA as contributors to these mechanisms is more complex. Central command is a term that describes a feedforward signal from the brain motor centers to the cardiovascular centers.81 This signal is thought to account for the nearly instantaneous rise in heart rate and arterial pressure that occurs with the onset of exercise. In this context, when conscious humans perform a static handgrip, there is usually no immediate rise in MSNA but there can be a rise in SSNA.41,80 These general observations suggest that central command is not a major regulator of MSNA in humans. Additionally, the initial burst in SSNA at the onset of contractions is consistent with the increased SSNA responses

seen with arousal stimuli. These observations again demonstrate that MSNA and SSNA behave differently and emphasize the idea that sympathetic outflow can be differentiated and selective in humans.

Sensory Feedback from Active Muscles Although MSNA does not rise at the onset of contractions if a static handgrip of sufficient force is maintained for a few minutes, there is a progressive rise in MSNA.41 This increase in MSNA can be maintained after the contractions if an arm cuff is inflated to supersystolic levels just before stopping the exercise. The general interpretation of this observation is that afferent nerves (i.e., ergoreceptors) in the active muscles that are sensitive to muscle metabolites can evoke a reflex increase in MSNA and arterial pressure. It does not appear that SSNA is under ergoreceptor control. Figure 15–8 details the MSNA responses to static handgripping and postexercise ischemia in humans. In general, the ergoreceptors are thinly myelinated (group III) or unmyelinated (group IV) afferents. In resting muscle, group III afferents are primarily mechanosensitive and group IV afferents chemosensitive. However, with contraction, the group III afferents become sensitized and respond to chemical stimuli. Studies in animals suggest that, although a variety of substances can stimulate the ergoreceptors, the factor that is most prominent in vivo is the H⫹ ion. This idea has been generally confirmed in humans, but it also appears that other unknown substances can stimulate the ergoreceptors when H⫹ ion is absent.34,78

Baroreflex Resetting during Exercise Blood pressure rises during exercise, and it appears that this increase in pressure is also facilitated by a resetting of the arterial baroreflexes so that the “set point” (Fig. 15–9) is elevated and a higher pressure is regulated. Baroreflex resetting is thought to be a second way that central command contributes to the autonomic responses to exercise.49,55,81 The neural substrates responsible for this

329

Sympathetic Nerves and Control of Blood Vessels to Human Limbs

10 sec. Mean voltage neurogram of MSA Control

Handgrip 1st min.

Handgrip 2nd min.

MIR 1st min.

MIR 2nd min.

Recovery

HR (beats/min)

76

88

83

67

63

67

MSA (burst/min X mean burst ampl.)

476

410

1182

1473

1386

603

MAP (mmHg)

100

114

127

122

122

107

FIGURE 15–8 Effects of isometric handgrip exercise and postexercise circulatory rest on MSNA responses to handgripping in humans. With the onset of contractions (central command), there is little increase in MSNA, but there are marked increases in heart rate and arterial pressure. As the contractions become fatiguing, MSNA slowly rises. When contractions stop but the metabolites associated with contractions are “trapped” in previously active muscles by inflation of an arm cuff (postexercise circulatory rest; MIR), the rise in sympathetic traffic seen at the end of exercise is maintained. These data demonstrate that central command is not associated with a large increase in MSNA at the onset of contractions and that, during exercise, MSNA is primarily under the control of “ergoreceptors” in the active muscles. These receptors appear to be sensitive to acidosis and the metabolic by-products of contraction. One idea is that they signal a “mismatch” between muscle metabolism and muscle blood flow, and the purpose of the increase in arterial pressure is to improve the blood flow to the “ischemic” contracting muscles. (From Mark, A. L., Victor, R. G., Nerhed, C., and Wallin, B. G.: Microneurographic studies of the mechanisms of sympathetic nerve responses to static exercise in humans. Circ. Res. 57:461, 1985, with permission.)

Sympathetic nerve activity or heart rate

Exercise

Rest

Arterial pressure

FIGURE 15–9 Schematic representation of arterial baroreflex “resetting” during exercise in humans. Exercise is associated with an increase in heart rate and arterial pressure. It is also associated with a “resetting” of baroreflex control of these variables. In general, the shape of the stimulus-response curve relating arterial pressure to either sympathetic nerve activity or heart rate is unchanged with exercise but the “set point” of the curve is shifted to the right. The mechanisms responsible for this are unknown, but one idea is that the “central command” associated with the motor effort needed to produce the contractions also stimulates the brainstem cardiovascular centers and causes the baroreflex resetting.

response in humans are poorly understood, but baroreflex resetting is thought to maintain or support the autonomic adjustments made by central command and ergoreceptors. To study central command and baroreflex resetting in humans, it is necessary to increase or decrease the effort associated with a given level of exercise and measure arterial baroreflex behavior during the various conditions. When humans are made weak with small doses of curarelike drugs, the effort associated with a given level of exercise is disproportionately increased. Under these conditions, when subjects are weak, the arterial pressure and heart rate responses to contraction are increased and there is clear resetting of arterial baroreflexes.49 The brief discussion above on the MSNA and autonomic response to exercise has generally used static handgrip exercise as a model to illustrate the role of central command, ergoreceptors, and baroreflex resetting. It should also be noted that MSNA is normally measured in resting muscle, but using a variety of approaches, it is clear that sympathetic activity increases in exercising muscle. The mechanisms that regulate MSNA during static handgripping also participate in the autonomic responses to whole-body dynamic exercise such as running or cycling. However, there is an additional challenge to consider during whole-body exercises55,62: how does the autonomic nervous system respond to the marked skeletal muscle vasodilatation seen in the active muscles?

330

Function of the Peripheral Nervous System

Interactions between Sympathetic Vasoconstriction and Metabolic Vasodilatation Metabolic vasodilatation in active muscles is one of the hallmarks of exercise, and blood flow to a small mass of active skeletal muscle can reach very high values (⬃300 mL/100 g tissue/min).1 Additionally, if only a modest fraction (⬃10 kg) of total skeletal muscle mass is “maximally” vasodilated during exercise, the ability of the heart to generate the cardiac output required to support mean arterial pressure would be surpassed.55 However, peak skeletal muscle blood flow is lower during large versus small muscle mass exercise, suggesting that blood flow to active skeletal muscles is “restrained” by the sympathetic nervous system so

that mean arterial pressure can be regulated. Figure 15–10 illustrates the importance of sympathetic restraint of blood flow to active muscles, showing that arterial pressure falls during supine (or 15-degree head-down) cycle exercise in a patient who had undergone multilevel surgical sympathectomy in the 1960s for the treatment of severe hypertension.42 In contrast to the argument above, which favors sympathetic restraint of metabolic vasodilatation, evidence from a variety of skeletal muscle preparations in animals demonstrates that the constrictor responses to either sympathetic nerve stimulation or administration of ␣-adrenergic agonists are blunted or eliminated during exercise-induced vasodilatation.6,28,29,51,70–74 This observation has been termed functional sympatholysis.51

Radial artery (mm. Hg) Supine leg exercise

200

100

0 Baseline

Radial artery (mm. Hg) 200

15 Head down tilt Supine leg exercise

100

0 Baseline

FIGURE 15–10 Effects of sympathetic denervation on blood pressure responses to exercise in conscious humans. This figure comes from a patient who underwent extensive surgical sympathectomy for the treatment of malignant hypertension around 1960. The key finding is that, with supine exercise, blood pressure falls as a result of unopposed metabolic vasodilatation in the active muscles (top). This fall in pressure occurs even when the subject is 15% head down to maximize venous return and cardiac output (bottom). This figure argues for the idea that the sympathetic nerves must restrain the metabolic vasodilatation that can occur during exercise so that arterial pressure can be regulated and blood flow directed appropriately to both vital organs and the contracting skeletal muscle. (From Marshall, R. J., Schirger, A., and Shepherd, J. T.: Blood pressure during supine exercise in idiopathic orthostatic hypotension. Circulation 24:76, 1961, with permission.)

Sympathetic Nerves and Control of Blood Vessels to Human Limbs

␣-Receptor Subtypes and Sympathetic Control of Blood Flow in Active Muscles In contrast to earlier ideas that ␣1 receptors were primarily postjunctional and vasoconstrictive while ␣2 receptors were primarily prejunctional and inhibited norepinephrine release, it is now clear that resistance vessels have both postsynaptic ␣1 and ␣2 receptors that are vasoconstricting.18,32,33,76 These postsynaptic ␣2 receptors are major contributors to limb vascular tone at rest in many species, including humans. For example, forearm vasodilatation in response to selective ␣1 blockade is less than that seen with nonselective blockade.18 Additionally, the vasoconstriction in the human forearm caused by various maneuvers is only partially blunted by brachial artery administration of selective ␣1 antagonists, but eliminated by nonselective ␣-antagonists.33 In this context, several observations have important implications for sympathetic control of blood flow in active muscles. First, experiments in animals demonstrate that contraction, hypoxia, and acidosis inhibit the ability of postsynaptic ␣2 receptors (and to a much lesser extent ␣1 receptors) to cause vasoconstriction in skeletal muscle resistance vessels.2,44 Second, there are relatively more postsynaptic ␣2 receptors in “fast-twitch” muscles, and in general there are relatively more ␣2-receptors in smaller versus larger skeletal muscle arterioles. These differences in receptor subtype distribution therefore favor “sympatholysis” because the postsynaptic ␣2 receptors are likely to be located near muscle fibers that become acidotic with moderate or heavy contractile activity.70 This distribution should also permit some ␣1-mediated vasoconstriction in the larger arterioles, which play a greater role in the control of muscle blood flow when the downstream vessels are dilated. The third line of evidence that clearly suggests a strong role for postsynaptic ␣2 receptors in “sympatholysis” comes from a series of studies in animals.70–74 These studies used sympathetic nerve stimulation along with selective ␣1 and ␣2 agonists and antagonists to show that skeletal muscle contraction clearly limits (or even eliminates) postsynaptic ␣2but not ␣1-mediated vasoconstriction in fast-twitch muscles from rodents. Additionally, studies with nitric oxide synthase inhibitors and nitric oxide–deficient animals strongly implicate locally produced NO as the major factor inhibiting locally produced ␣2-mediated vasoconstriction in the active muscles.73 Finally, in conscious instrumented dogs there is evidence of ongoing sympathetic restraint of blood flow to active muscles, but there is also evidence that postsynaptic vasoconstriction is blunted (especially ␣2 receptor–mediated constrictions) in a manner that is consistent with the studies discussed above.5–8,48 Taken together, it appears that sympathetic control of blood flow to active muscles is reduced during exercise, and that NO produced in the contracting muscles acting on vasoconstricting postsynaptic ␣2 receptors is the major site of “sympatholysis.”

331

Exercise: Summary MSNA and SSNA behave differently during exercise in humans. The main regulator of the increased MSNA seen in humans during static exercise appears to be ergoreceptors in the active muscles. During large muscle mass exercise, arterial baroreflexes probably contribute. In general, central command probably does not cause MSNA to rise but can influence SSNA. The idea that central command causes a resetting of arterial baroreflexes is now gaining wide acceptance. The way that increased sympathetic activity interacts with metabolic vasodilatation via postsynaptic ␣1 and ␣2 receptors is currently a major area of investigation.

AUTONOMIC AND MSNA RESPONSES TO MENTAL STRESS In humans, mental or emotional stress can evoke marked increases in arterial pressure and heart rate. These changes are also typically associated with measurable forearm vasodilatation.3,52 This has led to the concept that there might be sympathetic vasodilator nerves to skeletal muscle in humans. This concept is supported by animal data from a variety of species that provide neurophysiologic, pharmacologic, and histologic evidence for sympathetic dilator nerves to skeletal muscle.35 In general, these nerves are either sympathetic cholinergic nerves and/or nitroxidergic nerves. One idea is that acetylcholine released from sympathetic cholinergic nerves acts on the vascular endothelium to evoke NO release.12,43 Sympathetic vasodilator nerves to skeletal muscle in animals appear to be active during the so-called defense reaction, which occurs when certain brainstem areas are stimulated and evoke skeletal muscle vasodilatation in conjunction with hypertension and tachycardia.35,52 This response is thought to prepare the animal for a classic “fight or flight” response.

Neurally Mediated Dilatation in Human Skeletal Muscle A variety of lines of evidence suggest there might be neurally mediated skeletal muscle vasodilatation in humans. Some of the most obvious evidence comes from “mental stress” experiments conducted in the 1950s. In these studies, severe mental stress was associated with hypertension, tachycardia, and marked skeletal muscle vasodilatation.3 In a select number of subjects who had undergone surgical sympathectomy of the forearm, the vasodilatation was absent in the sympathectomized forearm. These and related studies were seen as roughly analogous to the defense reaction studies in animals and offered strong support for the existence of sympathetic dilator fibers to human muscle. However, in contrast to the older human studies, recent studies suggest that sympathetic vasodilator nerves to

332

Function of the Peripheral Nervous System

human skeletal muscle do not exist. First, in both primates and humans there is no histologic evidence for the existence of sympathetic cholinergic vasodilator nerves in muscle.35 Second, the vasodilator nerve responses to mental stress persist after various forms of nerve block, which should abolish sympathetic traffic to the forearm.27 Third, microneurographic records of sympathetic traffic to the forearm during mental stress demonstrate that there is sympathetic withdrawal during mental stress and not sympathetic activation.27 In other words, there is no neurophysiologic evidence for sympathetic dilator traffic. Although it is clear that the vasodilatation evoked by mental stress is at least partly dependent on NO, the current mechanism for the release of the NO appears to be sympathetic withdrawal, epinephrine-mediated activation of ␤2 receptors, and local mechanical factors evoking NO release.35,50

AGING, HYPERTENSION, AND CONGESTIVE HEART FAILURE When healthy older subjects who are free of disease (diabetes, hypertension, heart disease, etc.) are studied, it is clear that there is generalized increase in sympathetic outflow to many vascular beds and organ systems.59,67 Resting norepinephrine values are frequently higher, and norepinephrine spillover to many organ systems is increased. In this context, there is clear-cut evidence that sympathetic outflow to skeletal muscle, gut, brain, and perhaps the heart is increased with normal aging59 (Fig. 15–11). By contrast, sympathetic outflow to the kidney does not appear to increase with advancing age. Although baseline sympathetic outflow to many organ systems is increased with advancing age, the rise in sympathetic outflow above baseline during

A 20 yr old woman

24 yr old man

AP=122/78

BF

BI

11

19

17

29

25

38

38

66

AP=116/70

65 yr old woman

68 yr old man

AP=138/84

AP=116/68

5s

B 500

*

40

PNA (pg ml–1)

MSNA (bursts min–1)

50

*

30 * 20 * 10 0

400 300 200 100 0

YW

YM

OW

OM

YW

YM

OW

OM

FIGURE 15–11 A, Individual records of MSNA in older and younger men and women. B, MSNA and plasma noradrenaline levels in various groups. Young women (YW) have lower sympathetic traffic than young men (YM). Both younger groups have lower values than old women (OW). The values seen in older men (OM) are roughly twice those seen in their younger counterparts. These findings demonstrate that there is a progressive rise in sympathetic tone with aging. The mechanisms responsible for this rise are currently unknown. (From Seals, D. R., and Esler, M. D.: Human ageing and the sympathoadrenal system. J. Physiol. [Lond.] 528:407, 2000, with permission.)

Sympathetic Nerves and Control of Blood Vessels to Human Limbs

acute stress (standing, exercise, etc.) appears to be essentially normal.13–15,47,59,68 Currently the mechanisms responsible for the age-related changes in sympathetic outflow are poorly understood, but one idea is that increased subcortical central nervous sympathetic drive is responsible.39,59 Another possible contributor to the increased sympathetic outflow seen with aging is a progressive stiffening of the arterial blood vessels. If the vessels at the carotid and aortic receptors stiffen with age, a given change in arterial pressure will evoke less deformation and hence less baroreceptor inhibition of sympathetic outflow.11,31,69 In view of the aging population throughout the world, and the key role that the sympathetic nervous system plays in many disease states, further understanding of the autonomic responses to normal aging and aging in the context of disease are likely to become increasingly important. The pathophysiologic mechanisms that contribute to hypertension in humans are complex, and dysfunction in a host of key regulatory systems has been proposed. In this context, one important question is whether sympathetic outflow to skeletal muscle is chronically increased in individuals with essential hypertension. Although a detailed description of all the evidence for and against changes in MSNA as contributors to hypertension in humans is beyond the scope of this chapter, currently available evidence suggests that increases in MSNA are not major contributors to hypertension in most situations.67,83 The current idea is that baseline MSNA in individuals with essential hypertension is either normal or only slightly elevated. In contrast to hypertension, MSNA is clearly elevated in congestive heart failure. In many patients resting MSNA is roughly twofold greater than normal. Additionally, congestive heart failure is also associated with other signs of sympathetic activation, including increased resting norepinephrine levels.40 For many years the sympathetic activation and increase in MSNA with congestive heart failure were assumed to be due to the inability of the arterial cardiopulmonary receptors to inhibit sympathetic outflow as a result of poor “pump function” in people with congestive heart failure. More recent evidence from animal models suggests that congestive heart failure is associated with a complex series of events in the central nervous system that is associated with reduced centrally mediated inhibition of sympathetic outflow.85

AUTONOMIC RESPONSES TO BODY HEATING Increases in core temperature are sensed at the level of the hypothalamus and evoke widespread autonomic responses. Most notably, when the core temperature is increased by 0.5° to 1.0° C, there can be marked cutaneous vasodilatation. As core temperature increases between 1° and 2° C, skin blood flow can reach values

333

of 7 to 8 L/min. This means that skin blood flow can reach levels of 300 to 500 mL/100 g/min. This increase in flow is neurally mediated via sympathetic dilator fibers.53 Additionally, emerging evidence suggests that (as is the case with high skeletal muscle blood flows), when skin blood flow is very high, arterial baroreflexes act to either limit skin vasodilator nerve activity or restrain it via concurrent activation of vasoconstrictor activity.38

Cutaneous Vascular Responses to Changes in Temperature At rest, skin blood flow is probably on the order of 1 to 3 mL/100 g/min. When subjects are cooled, thermoregulatory reflexes are activated and there can be intense vasoconstriction of skin to protect core temperature.37,53 This vasoconstriction is mediated via noradrenergic sympathetic nerves. When core temperature increases initially, there is a withdrawal of vasoconstrictor sympathetic activity to skin and skin blood flow rises. As core temperature rises between 0.5° and 1.0° C, there is a marked increase in skin blood flow that occurs about the same time as the onset of sweating. As core temperature continues to rise, dilatation progresses. Temporary elimination of sympathetic outflow to the skin (nerve block) or surgical sympathectomy eliminates both the sweating response and the marked cutaneous vasodilator responses to body heating.53 In general, the sweating response is governed by sympathetic cholinergic nerves and can be blocked by administration of atropine. In contrast, the neurally mediated vasodilatation that can occur in skin can be blunted modestly by atropine administration but cannot be eliminated completely.53,63 This observation has led to the conclusion that acetylcholine released from the sympathetic cholinergic (sudomotor) nerves is not responsible for the marked cutaneous vasodilatation that occurs with body heating (Fig. 15–12). It also suggests that either there is a separate population of vasodilator nerve fibers that are independent from the sympathetic cholinergic fibers, or perhaps that some of the sympathetic cholinergic nerves release a vasodilating cotransmitter. This latter possibility has recently been supported in studies by Kellogg and colleagues.37 They demonstrated that, although muscarinic blockade had little response on the cutaneous vasodilator responses to body heating in humans, prejunctional inhibition of sympathetic cholinergic fibers eliminated the dilator response. They interpreted these findings as strong evidence that some substance cotransmitted with acetylcholine from the sympathetic cholinergic nerves is responsible for the marked cutaneous vasodilatation during body heating. Finally, for many years bradykinin was postulated to play a major role in the cutaneous vasodilator responses to body heating, but recent evidence using selective blockers of bradykinin clearly demonstrate that this substance is not obligatory for the response.36

334

Function of the Peripheral Nervous System

SUMMARY

A

Relative humidity (mV)

20

Control 15

10 Atropine

5 Botulinum 0

B

100 Control

Atropine

60

40 Cold stress

% Maximal CVC

80

20

Botulinum

ACKNOWLEDGMENTS Whole body heat stress

Local warming to 42 C

0 0

In the last 20 to 30 years, microneurographic and other sophisticated measurements have been used to probe the complexities of the autonomic nervous system in conscious humans. These studies have shown that sympathetic outflow and changes in sympathetic outflow to various organ systems and vascular beds can be highly selective and targeted. An increase in sympathetic outflow to one vascular bed or organ system does not necessarily mean a generalized increase in sympathetic outflow throughout the body. Additionally, these new techniques have permitted mechanistic studies on the reflex control of the cardiovascular system in humans and better understanding of how baroreflexes and chemoreflexes govern the cardiovascular system in health and disease. Finally, the autonomic responses to common stressors such as standing, exercise, and body heating have been explored and new insights of how the autonomic nervous system governs these responses have been provided. The success of these approaches demonstrates the continued power of integrative physiology in conscious humans in an era of molecular biology and genomics.

10

20

30

40

50

60

70

80

90 100

Minutes

FIGURE 15–12 A, The effect of whole-body heating on relative humidity of a small area of skin. The relative humidity is proportional to the sweat rate. Under normal conditions whole-body heating is associated with a large increase in sweating. This can be blocked by atropine, which postsynaptically limits the ability of the sudomotor cholinergic nerves to evoke the sweating response. Additionally, it can be inhibited presynaptically with botulinum toxin, which limits the release of acetylcholine from these nerves. B, The impact of these treatments on the skin blood flow responses to body heating. Normally during body heating there is intense activation of sympathetic vasodilator nerves. Whether these nerves are the same as or different from the sudomotor cholinergic nerves responsible for the sweating has been under lengthy and controversial investigation. B demonstrates the control response and also demonstrates that atropine blunts the cutaneous dilator response modestly but does not eliminate it. By contrast, botulinum toxin eliminated this response. This experiment argues that some substance co-released with acetylcholine by the sudomotor nerves is responsible for the marked cutaneous vasodilatation seen during body heating in humans. Despite more than 70 years of work on this topic, the substance responsible for this dilatation is currently unknown. (Adapted from Kellogg, D. L. Jr., Pergola, P. E., Piest, K. K., et al.: Cutaneous active vasodilation in humans is mediated by cholinergic nerve cotransmission. Circ. Res. 77:1222, 1995, with permission.)

The authors would like to thank Janet Beckman for her excellent secretarial assistance. They would also like to thank the many volunteer subjects who have participated in their studies over the years. This work was supported in part by grants HL 46493, NS 32352, HL 65305, and RR00585.

REFERENCES 1. Andersen, P., and Saltin, B.: Maximal perfusion of skeletal muscle in man. J. Physiol. (Lond.) 366:233, 1985. 2. Anderson, K. M., and Faber, J. E.: Differential sensitivity of arteriolar alpha 1- and alpha 2-adrenoceptor constriction to metabolic inhibition during rat skeletal muscle contraction. Circ. Res. 69:174, 1991. 3. Blair, D. A., Glover, W. E., Greenfield, A. D. M., and Roddie, I. C.: Excitation of cholinergic vasodilator nerves to human skeletal muscles during emotional stress. J. Physiol. (Lond.) 148:633, 1959. 4. Bruce, E. N., and Cherniack, N. S.: Central chemoreceptors. J. Appl. Physiol. 62:389, 1987. 5. Buckwalter, J. B., and Clifford, P. S.: Alpha-adrenergic vasoconstriction in active skeletal muscles during dynamic exercise. Am. J. Physiol. 277:H33, 1999. 6. Buckwalter, J. B., and Clifford, P. S.: The paradox of sympathetic vasoconstriction in exercising skeletal muscle. Exerc. Sport Sci. Rev. 29:159, 2001. 7. Buckwalter, J. B., Mueller, P. J., and Clifford, P. S.: Alpha 1-adrenergic-receptor responsiveness in skeletal muscle during dynamic exercise. J. Appl. Physiol. 85:2277, 1998.

Sympathetic Nerves and Control of Blood Vessels to Human Limbs 8. Buckwalter, J. B., Naik, J. S., Valic, Z., and Clifford, P. S.: Exercise attenuates alpha-adrenergic-receptor responsiveness in skeletal muscle vasculature. J. Appl. Physiol. 90:172, 2001. 9. Burke, D., Sundlof, G., and Wallin, G.: Postural effects on muscle nerve sympathetic activity in man. J. Physiol. (Lond.) 272:399, 1977. 10. Chapleau, M. W., and Abboud, F. M.: Contrasting effects of static and pulsatile pressure on carotid baroreceptor activity in dogs. Circ. Res. 61:648, 1987. 11. Chapleau, M. W., Cunningham, J. T., Sullivan, M. J., et al.: Structural versus functional modulation of the arterial baroreflex. Hypertension 26:341, 1995. 12. Davisson, R. L., Johnson, A. K., and Lewis, S. J.: Nitrosyl factors mediate active neurogenic hindquarter vasodilation in the conscious rat. Hypertension 23:962, 1994. 13. Davy, K. P., Jones, P. P., and Seals, D. R.: Influence of age on the sympathetic neural adjustments to alterations in systemic oxygen levels in humans. Am. J. Physiol. 273:R690, 1997. 14. Davy, K. P., Seals, D. R., and Tanaka, H.: Augmented cardiopulmonary and integrative sympathetic baroreflexes but attenuated peripheral vasoconstriction with age. Hypertension 32:298, 1998. 15. Davy, K. P., Tanaka, H., Andros, E. A., et al.: Influence of age on arterial baroreflex inhibition of sympathetic nerve activity in healthy adult humans. Am. J. Physiol. 275:H1768, 1998. 16. Delius, W., Hagbarth, K. E., Hongell, A., and Wallin, B. G.: General characteristics of sympathetic activity in human muscle nerves. Acta Physiol. Scand. 84:65, 1972. 17. Delius, W., Hagbarth, K. E., Hongell, A., and Wallin, B. G.: Manoeuvres affecting sympathetic outflow in human muscle nerves. Acta Physiol. Scand. 84:82, 1972. 18. Dinenno, F. A., Eisenach, J. H., Dietz, N. M., and Joyner, M. J.: Post-junctional alpha-adrenoceptors and basal limb vascular tone in healthy men. J. Physiol. (Lond.) 540:1103, 2002. 19. Ebert, T. J., and Cowley, A. W. Jr.: Baroreflex modulation of sympathetic outflow during physiological increases of vasopressin in humans. Am. J. Physiol. Heart Circ. Physiol. 262:H1372, 1992. 20. Eckberg, D. L., and Sleight, P.: Human Baroreflexes in Health and Disease. New York, Oxford University Press, 1992. 21. Fagius, J., Wallin, B. G., Sundlof, G., et al.: Sympathetic outflow in man after anaesthesia of the glossopharyngeal and vagus nerves. Brain 108:423, 1985. 22. Fitzgerald, R. S., and Lahiri, S.: Reflex responses to chemoreceptor stimulation. In Cherniack, N. S., and Widdicombe, J. G. (eds.): Handbook of Physiology: The Respiratory System. Vol. II: Control of Breathing. Washington, DC, American Physiological Society, p. 313, 1986. 23. Fitzpatrick, A. P., Banner, N., Cheng, A., et al.: Vasovagal reactions may occur after orthotopic heart transplantation. J. Am. Coll. Cardiol. 21:1132, 1993. 24. Hagbarth, K. E., Hallin, R. G., Hongell, A., et al.: General characteristics of sympathetic activity in human skin nerves. Acta Physiol. Scand. 84:164, 1972. 25. Hagbarth, K. E., and Vallbo, A. B.: Pulse and respiratory grouping of sympathetic impulses in human muscle nerves. Acta Physiol. Scand. 74:96, 1968.

335

26. Halliwill, J. R.: Segregated signal averaging of sympathetic baroreflex responses in humans. J. Appl. Physiol. 88:767, 2000. 27. Halliwill, J. R., Lawler, L. A., Eickhoff, T. J., et al.: Forearm sympathetic withdrawal and vasodilatation during mental stress in humans. J. Physiol. (Lond.) 504:211, 1997. 28. Hansen, J., Sander, M., and Thomas, G. D.: Metabolic modulation of sympathetic vasoconstriction in exercising skeletal muscle. Acta Physiol. Scand. 168:489, 2000. 29. Hansen, J., Sayad, D., Thomas, G. D., et al.: Exerciseinduced attenuation of alpha-adrenoceptor mediated vasoconstriction in humans: evidence from phase-contrast MRI. Cardiovasc. Res. 41:220, 1999. 30. Heeringa, J., Berkenbosch, A., de Goede, J., and Olievier, C. N.: Relative contribution of central and peripheral chemoreceptors to the ventilatory response to CO2 during hyperoxia. Respir. Physiol. 37:365, 1979. 31. Hunt, B. E., Farquhar, W. B., and Taylor, J. A.: Does reduced vascular stiffening fully explain preserved cardiovagal baroreflex function in older, physically active men? Circulation 103:2424, 1984. 32. Jie, K., van Brummelen, P., Vermey, P., et al.: Identification of vascular postsynaptic alpha 1- and alpha 2-adrenoceptors in man. Circ. Res. 54:447, 1984. 33. Jie, K., van Brummelen, P., Vermey, P., et al.: Postsynaptic alpha 1- and alpha 2-adrenoceptors in human blood vessels: interactions with exogenous and endogenous catecholamines. Eur. J. Clin. Invest. 17:174, 1987. 34. Joyner, M. J.: Muscle chemoreflexes and exercise in humans. Clin. Auton. Res. 2:201, 1992. 35. Joyner, M. J., and Halliwill, J. R.: Sympathetic vasodilatation in human limbs. J. Physiol. (Lond.) 526:471, 2000. 36. Kellogg, D. L. Jr., Liu, Y., McAllister, K., et al.: Bradykinin does not mediate cutaneous active vasodilation during heat stress in humans. J. Appl. Physiol. 93:1215, 2002. 37. Kellogg, D. L. Jr., Pergola, P. E., Piest, K. K., et al.: Cutaneous active vasodilation in humans is mediated by cholinergic nerve cotransmission. Circ. Res. 77:1222, 1995. 38. Kenney, W. L., Tankersley, C. G., Newswanger, D. L., and Puhl, S. M.: Alpha 1-adrenergic blockade does not alter control of skin blood flow during exercise. Am. J. Physiol. 260:H855, 1991. 39. Lambert, G. W., Kaye, D. M., Thompson, J. M., et al.: Internal jugular venous spillover of noradrenaline and metabolites and their association with sympathetic nervous activity. Acta Physiol. Scand. 163:155, 1998. 40. Leimbach, W. N. Jr., Wallin, B. G., Victor, R.G., et al.: Direct evidence from intraneural recordings for increased central sympathetic outflow in patients with heart failure. Circulation 73:913, 1986. 41. Mark, A. L., Victor, R. G., Nerhed, C., and Wallin, B. G.: Microneurographic studies of the mechanisms of sympathetic nerve responses to static exercise in humans. Circ. Res. 57:461, 1985. 42. Marshall, R. J., Schirger, A., and Shepherd, J. T.: Blood pressure during supine exercise in idiopathic orthostatic hypotension. Circulation 24:76, 1961. 43. Matsukawa, K., Shindo, T., Shirai, M., and Ninomiya, I.: Nitric oxide mediates cat hindlimb cholinergic vasodilation induced by stimulation of posterior hypothalamus. Jpn. J. Physiol. 43:473, 1993.

336

Function of the Peripheral Nervous System

44. McGillivray-Anderson, K. M., and Faber, J. E.: Effect of acidosis on contraction of microvascular smooth muscle by alpha 1- and alpha 2-adrenoceptors: implications for neural and metabolic regulation. Circ. Res. 66:1643, 1990. 45. Mifflin, S. W.: What does the brain know about blood pressure? News Physiol. Sci. 16:266, 2001. 46. Morgan, B. J., Crabtree, D. C., Palta, M., and Skatrud, J. B.: Combined hypoxia and hypercapnia evokes long-lasting sympathetic activation in humans. J. Appl. Physiol. 79:205, 1995. 47. Ng, A. V., Callister, R., Johnson, D. G., and Seals, D. R.: Sympathetic neural reactivity to stress does not increase with age in healthy humans. Am. J. Physiol. 267:H344, 1994. 48. O’Leary, D. S., Robinson, E. D., and Butler, J. L.: Is active skeletal muscle functionally vasoconstricted during dynamic exercise in conscious dogs? Am. J. Physiol. 272:R386, 1997. 49. Raven, P. B., Fadel, P. J., and Smith, S. A.: The influence of central command on baroreflex resetting during exercise. Exerc. Sport Sci. Rev. 30:39, 2002. 50. Reed, A. S., Tschakovsky, M. E., Minson, C. T., et al.: Skeletal muscle vasodilatation during sympathoexcitation is not neurally mediated in humans. J. Physiol. (Lond.) 525:253, 2000. 51. Remensnyder, J. P., Mitchell, J. H., and Sarnoff, S. J.: Functional sympatholysis during muscular activity. Circ. Res. 11:370, 1962. 52. Roddie, I. C.: Human responses to emotional stress. Ir. J. Med. Sci. 146:395, 1977. 53. Roddie, I. C.: Circulation to skin and adipose tissue. In Shepherd, J. T., and Abboud, F. M. (eds.): Handbook of Physiology, Sect. 2: The Cardiovascular System. Vol. III: Peripheral and Organ Blood Flow. Bethesda, MD, American Physiological Society, p. 285, 1983. 54. Rowell, L. B., Johnson, D. G., Chase, P. B., et al.: Hypoxemia raises muscle sympathetic activity but not norepinephrine in resting humans. J. Appl. Physiol. 66:1736, 1989. 55. Rowell, L. B., and O’Leary, D. S.: Reflex control of the circulation during exercise: chemoreflexes and mechanoreflexes. J. Appl. Physiol. 69:407, 1990. 56. Rudas, L., Crossman, A. A., Morillo, C. A., et al.: Human sympathetic and vagal baroreflex responses to sequential nitroprusside and phenylephrine. Am. J. Physiol. Heart Circ. Physiol. 276:H1691, 1999. 57. Rutherford, J. D., and Vatner, S. F.: Integrated carotid chemoreceptor and pulmonary inflation reflex control of peripheral vasoreactivity in conscious dogs. Circ. Res. 43:200, 1978. 58. Saito, M., Mano, T., Iwase, S., et al.: Responses in muscle sympathetic activity to acute hypoxia in humans. J. Appl. Physiol. 65:1548, 1988. 59. Seals, D. R., and Esler, M. D.: Human ageing and the sympathoadrenal system. J. Physiol. (Lond.) 528:407, 2000. 60. Seals, D. R., Suwarno, N. O., and Dempsey, J. A.: Influence of lung volume on sympathetic nerve discharges in normal humans. Circ. Res. 67:130, 1990. 61. Seals, D. R., Suwarno, N. O., Joyner, M. J., et al.: Respiratory modulation of muscle sympathetic nerve activity in intact and lung denervated humans. Circ. Res. 72:440, 1993. 62. Seals, D. R., and Victor, R. G.: Regulation of muscle sympathetic nerve activity during exercise in humans. Exerc. Sport Sci. Rev. 19:313, 1991.

63. Shastry, S., Minson, C. T., Wilson, S. A., et al.: Effects of atropine and L-NAME on cutaneous blood flow during body heating in humans. J. Appl. Physiol. 88:467, 2000. 64. Smith, M. L., Carlson, M. D., and Thames, M. D.: Naloxone does not prevent vasovagal syncope during simulated orthostasis in humans. J. Auton. Nerv. Syst. 45:1, 1993. 65. Somers, V. K., Mark, A. L., Zavala, D. C., and Abboud, F. M.: Influence of ventilation and hypocapnia on sympathetic nerve responses to hypoxia in normal humans. J. Appl. Physiol. 67:2095, 1989. 66. Sundlof, G., and Wallin, B. G.: The variability of muscle nerve sympathetic activity in resting recumbent man. J. Physiol. (Lond.) 272:383, 1977. 67. Sundlof, G., and Wallin, B. G.: Human muscle nerve sympathetic activity at rest: relationship to blood pressure and age. J. Physiol. (Lond.) 274:621, 1978. 68. Tanaka, H., Davy, K. P., and Seals, D. R.: Cardiopulmonary baroreflex inhibition of sympathetic nerve activity is preserved with age in healthy humans. J. Physiol. (Lond.) 515:249, 1999. 69. Tanaka, H., Dinenno, F. A., Monahan, K. D., et al.: Aging, habitual exercise, and dynamic arterial compliance. Circulation 102:1270, 2000. 70. Thomas, G. D., Hansen, J., and Victor, R. G.: Inhibition of alpha 2-adrenergic vasoconstriction during contraction of glycolytic, not oxidative, rat hindlimb muscle. Am. J. Physiol. 266:H920, 1994. 71. Thomas, G. D., Hansen, J., and Victor, R. G.: ATP-sensitive potassium channels mediate contraction-induced attenuation of sympathetic vasoconstriction in rat skeletal muscle. J. Clin. Invest. 99:2602, 1997. 72. Thomas, G. D., Sander, M., Lau, K. S., et al.: Impaired metabolic modulation of alpha-adrenergic vasoconstriction in dystrophin-deficient skeletal muscle. Proc. Natl. Acad. Sci. U. S. A. 95:15090, 1998. 73. Thomas, G. D., and Victor, R. G.: Nitric oxide mediates contraction-induced attenuation of sympathetic vasoconstriction in rat skeletal muscle. J. Physiol. (Lond.) 506:817, 1998. 74. Thomas, G. D., Zhang, W., and Victor, R. G.: Impaired modulation of sympathetic vasoconstriction in contracting skeletal muscle of rats with chronic myocardial infarctions: role of oxidative stress. Circ. Res. 88:816, 2001. 75. Vallbo, A. B., Hagbarth, K. E., Torebjork, H. E., and Wallin, B. G.: Somatosensory, proprioceptive, and sympathetic activity in human peripheral nerves. Physiol. Rev. 59:919, 1979. 76. van Brummelen, P., Jie, K., and van Zwieten, P. A.: Alphaadrenergic receptors in human blood vessels. Br. J. Clin. Pharmacol. 21:33S, 1986. 77. Vanhoutte, P. M., and Shepherd, J. T.: Autonomic nerves to the systemic blood vessels. In Dyck, P. J., Thomas, P. K., Lambert, E. H., and Bunge, R. (eds.): Peripheral Neuropathy, 2nd ed. Philadelphia, W. B. Saunders, p. 301, 1984. 78. Victor, R. G., and Seals, D. R.: Reflex stimulation of sympathetic outflow during rhythmic exercise in humans. Am. J. Physiol. 257:H2017, 1989. 79. Victor, R. G., Seals, D. R., and Mark, A. L.: Differential control of heart rate and sympathetic nerve activity during

Sympathetic Nerves and Control of Blood Vessels to Human Limbs dynamic exercise: insight from direct intraneural recordings in humans. J. Clin. Invest. 79:508, 1987. 80. Vissing, S. F., and Hjortso, E. M.: Central motor command activates sympathetic outflow to the cutaneous circulation in humans. J. Physiol. (Lond.) 492:931, 1996. 81. Waldrop, T. G., Eldridge, F. L., Iwamoto, A., and Mitchell, J. H.: Central neural control of respiration and circulation during exercise. In Rowell, L. B., and Shepherd, J. T. (eds.): Handbook of Physiology, Sect. 12: Exercise: Regulation and Integration of Multiple Systems. New York, Oxford University Press, p. 333, 1996.

337

82. Wallin, B. G., and Fagius, J.: The sympathetic nervous system in man—aspects derived from microelectrode recordings. Trends Neurosci. 9:63, 1986. 83. Wallin, B. G., and Fagius, J.: Peripheral sympathetic neural activity in conscious humans. Annu. Rev. Physiol. 50:565, 1988. 84. Wallin, B. G., and Sundlof, G.: Sympathetic outflow to muscles during vasovagal syncope. J. Auton. Nerv. Syst. 6:287, 1982. 85. Zucker, I. H., Wang, W., Brandle, M., et al.: Neural regulation of sympathetic nerve activity in heart failure. Prog. Cardiovasc. Dis. 37:397, 1995.

16 Molecular Signaling in Schwann Cell Development RHONA MIRSKY AND KRISTJÁN R. JESSEN

Outline of Schwann Cell Development Differentiation Markers Associated with Embryonic Schwann Cell Development Brain-Specific Fatty Acid–Binding Protein Protein Zero Desert Hedgehog Oct6 Protein Growth-Associated Protein-43, CD9, Peripheral Myelin Protein 22, and Proteolipid Protein Neural Crest Origins of Schwann Cells Schwann Cell Precursors

Developmental Potential of Early Peripheral Glia Importance of Neuregulin-1 in Schwann Cell Development Neuregulin-1 Signaling Pathways in Schwann Cells Regulation of the Precursor–Schwann Cell Transition Signals That Control Schwann Cell Survival Cell Culture Studies on Schwann Cell Proliferation and Differentiation Transcription Factors Involved in Myelination

This chapter reviews glial development in embryonic and early neonatal nerve trunks. The origin of the Schwann cell lineage from the neural crest and the cell transformations that result in the emergence of immature Schwann cells are discussed. The extracellular signaling molecules and matrix receptors involved in controlling fate and lineage progression are considered, and the role of transcription factors that are known to be involved in regulating embryonic Schwann cell development and the onset of myelination is described. The myelinating and nonmyelinating Schwann cells of adult peripheral nerves are the two main peripheral glial cell types. Nevertheless, several other distinct categories of peripheral glia exist. These include satellite cells that surround neuronal cell bodies in sympathetic, parasympathetic, and sensory ganglia; the astrocyte-like enteric glial cells of the enteric nervous system; the teloglia (terminal Schwann cells) of somatic motor nerve terminals; and the specialized glia that associate with sensory terminals such as pacinian corpuscles and olfactory ensheathing cells.55,92,96,103,135,164,186,240,241,259,263,268,305,371 Evidence to date suggests that the molecular and morphologic differences

Signaling Pathways and Growth Factors That Regulate Myelination Neural Activity and Schwann Cell Development Extracellular Matrix, Matrix Receptors, and the Cytoskeleton in Schwann Cell Development Laminin Integrin Signaling in Schwann Cells and Interactions with the Actin Cytoskeleton Dystroglycan Signaling in Schwann Cells

between these various cells depend on the specific location and cellular environment in which they are found, and that the glial cells of the peripheral nervous system (PNS) retain unusual plasticity throughout life.

OUTLINE OF SCHWANN CELL DEVELOPMENT The myelinating and nonmyelinating Schwann cells of adult nerves originate from the neural crest in a process reviewed by several authors.136,184,208,209,277,336 Three main developmental transitions lie between migrating crest cells and mature Schwann cells (Fig. 16–1). The first involves the generation of Schwann cell precursors from neural crest cells. Next, the precursors transform into immature Schwann cells, which finally generate either myelinating or nonmyelinating cells. The first and last of these transition points involve a choice of fate, because neural crest cells give rise to several other cell types besides glia, and immature Schwann cells will become either myelinating 341

342

Neurobiology of the Peripheral Nervous System

FIGURE 16–1 The Schwann cell lineage in rat and mouse. There are three main transitions in the Schwann cell lineage: (1) the formation of Schwann cell precursors from undifferentiated migrating neural crest cells, (2) the precursor–Schwann cell transition, and (3) the formation of mature myelinating and nonmyelinating Schwann cells. The phenotypic changes that characterize each developmental step are summarized in Figure 16–2. The reversibility of postnatal Schwann cell development is indicated by stippled arrows. (From Jessen, K. R., and Mirsky, R.: Schwann cell development. In Lazzarini, R. A. [ed.]: Myelin Biology and Disorders, Vol. 1. San Diego, CA, Academic Press, p. 329, 2003, with permission.)

or nonmyelinating cells. These two transitions are also reversible, because myelinating and nonmyelinating cells revert to a phenotype comparable to that of immature Schwann cells, and Schwann cell precursors can be reprogrammed to generate other crest derivatives, at least in vitro. The transition from precursors to Schwann cells involves lineage progression rather than fate choice, because the only known fate of Schwann cell precursors in normal development is to become Schwann cells or die by programmed cell death. The extracellular signals that control fate choice in the lineage are not well understood. It is unclear what signals or mechanisms enable or direct crest cells to enter the glial lineage, and remarkably, the signals from larger diameter axons that are believed to instruct the Schwann cells associated with them to myelinate are unknown. Despite the lack of definitive information about what controls signaling at the choice points, two factors that regulate embryonic Schwann cell development have been identified by a combination of in vitro and in vivo experiments. These are neuregulin-1, which carries out a number of major functions in developing nerves, and endothelin, which is involved in the precursor–Schwann cell transition.26,99,136 (Neuregulin-1, also known as heregulin, ARIA, glial growth factor [GGF], and sensory neuron and motor–derived factor [SMDF], is a growth factor that, besides being involved in Schwann cell development, is also implicated in the progression of breast and other cancers.) Two transcription factors are known to be indispensable for Schwann cell development. The

first, Sox10, is essential for the generation of the earliest cells in the Schwann cell lineage, while the second, Krox20 (Egr2), is essential for myelination.27,336 (Sox10 is a member of the Sox family of the HMG [high mobility group] domain family of transcription factors. Another member of this family, SRY [sex-determining region of the Y chromosome], is essential for proper development of the male gonads, and other family members play crucial roles in development. Krox20 is one of four members of the EGR [early growth response] zinc finger transcription factor family. Another member of the family, Krox24 [Egr1], is essential for pituitary and ovarian development.) Additionally, the transcription factors Oct6 (SCIP, Tst1, POU3f1) and, to a lesser extent, Brn2 are important for the timing of myelination, while in vitro evidence suggests that the transcription factor nuclear factor-␬B (NF-␬B) is involved in the control of Oct6 expression.132,233,336 (Oct6 and Brn2 are both members of the POU domain family of transcription factors, which bend DNA and bind to DNA via an octamer sequence. They also contain a homeodomain and are involved in a variety of developmental and homeostatic processes. NF-␬B is a transcription factor involved in the regulation of many immune and inflammatory factors.) In vivo, Schwann cell precursors and immature Schwann cells proliferate rapidly, with a peak of DNA incorporation at the immature Schwann cell stage.30,313 Therefore, cell division is compatible with the differentiation processes that transform the phenotype of crest cells first to that of Schwann cell precursors and then to that of immature

Molecular Signaling in Schwann Cell Development

Schwann cells. Exit from the cell cycle occurs only at the final transition as immature Schwann cells differentiate into myelinating and nonmyelinating cells. A remarkable feature of the Schwann cell lineage is the rapid reversibility of this last step of Schwann cell development. Removal of Schwann cells from contact with axons, which initiates the reversal, can be achieved either by injuring nerves in vivo or by dissociating cells from adult nerves and placing them in culture without neurons. Both in vivo and in vitro, the process entails the developmental regression and dedifferentiation of individual Schwann cells and myelin breakdown. The characteristic molecular markers and structural features of myelinating and nonmyelinating cells are lost and the cells reenter the cell cycle and regress to a phenotype similar to that of immature Schwann cells. Such denervated Schwann cells redifferentiate readily on reassociation with axons during nerve regeneration.82,279 Are the other main transitions in the lineage also reversible? Reversal of the phenotype of immature Schwann cells leading to the re-formation of Schwann cell precursors has not yet been observed. Schwann cell precursors, in contrast, are likely to be more plastic. These cells have a phenotype that differentiates them unambiguously from both neural crest and early developing neurons, which indicates that they have been specified as glial cells. In their normal signaling environment in developing nerve trunks, they are also likely only to have a Schwann cell fate. As discussed further below, this does not imply that these cells are impotent to generate other cell types when challenged with alternative signals in an alien environment. It is possible that the entire cell population is not converted to the next stage at each of the three main transitions in Schwann cell development. For example, it has been suggested that, in addition to Schwann cell precursors, embryonic day (E) 14 rat nerves harbor a separate population of cells (10% to 15% of total cell number) that resembles neural crest cells.218,219 Evidence suggests that some precursor-like cells might be present in perinatal nerves21 and that even adult nerves may contain cells with an early phenotype corresponding to precursors or immature Schwann cells.267 Many other tissues, including the central nervous system (CNS), the enteric nervous system, muscle, and the hematopoietic system, retain cell populations with a phenotype and developmental potential reminiscent of early developing cells in their respective lineages.

DIFFERENTIATION MARKERS ASSOCIATED WITH EMBRYONIC SCHWANN CELL DEVELOPMENT The use of a set of well-defined molecular markers that characterize distinct stages in the development of myelinating and nonmyelinating cells has greatly facilitated analysis

343

of postnatal Schwann cell development and myelination. Comparable tools to study initial glial differentiation in the PNS are still relatively poorly developed despite the fact that the early development of neurons can be monitored using well-characterized molecular markers. These include neuron-specific tubulin (␤-III tubulin; TUJ1), neurofilament, peripherin, and early markers of distinct peripheral neuronal lineages such as neurogenins, Phox2, and Mash1.4,202,329,333 (Neurofilament and peripherin are intermediate filament proteins characteristic of neurons. Neurogenins 1 through 3 and Mash1 are members of the helix-loop-helix [HLH] family of transcription factors, while Phox2s are homeodomain transcription factors. All of these transcription factors play crucial roles in the specification and differentiation of distinct subsets of PNS and CNS neurons.) An orderly analysis of glial lineage progression in embryonic nerves has therefore not been straightforward, and markers that appear at a relatively late stage, in particular glial fibrillary acidic protein (GFAP) and S100, have often been used to monitor early events such as the emergence of the glial lineage from neural crest cells. (GFAP is an intermediate filament protein expressed by some astrocytes, enteric glia, and nonmyelinating Schwann cells. S100 is expressed by Schwann cells, astrocytes, neurons, and oligodendrocyte precursors, and is used in some circumstances as a glial marker.) Differentiation markers for embryonic Schwann cell development in rats and mice are, however, now appearing, and the task should become easier. Using stage specificity as the criterion, the markers shown in Figure 16–2 fall into four categories: 1. Markers that are expressed by embryonic PNS glia but do not differentiate between developmental stages, exemplified by the cell adhesion molecule L1 2. Markers that are expressed by crest cells and Schwann cell precursors but are expressed at very low levels on immature Schwann cells, namely N-cadherin (a cell adhesion molecule) and the transcription factor AP2␣ (a member of the AP2 family implicated in neural crest development) 3. Markers that differentiate Schwann cell precursors and immature Schwann cells from crest cells; three of these, brain-specific fatty acid–binding protein (B-FABP), protein zero (P0), and desert hedgehog (Dhh), are not found on early developing neurons and therefore distinguish between the neuronal and glial lineage at a very early stage 4. Markers that are expressed by immature Schwann cells that can be used to distinguish between them and Schwann cell precursors or crest cells, exemplified by S100 and GFAP Markers in the last two groups are particularly useful for monitoring the two main lineage transitions. Since the precursor–Schwann cell transition is relatively well characterized,136,209,210 the remarks below are limited to

344

Neurobiology of the Peripheral Nervous System

Neutal crest

Schwann-cell Precursors

Immature Schwann-cells

FIGURE 16–2 Markers of embryonic Schwann cell development. Neural crest cells, Schwann cell precursors, and Schwann cells are characterized by molecular markers as indicated in the graded gray boxes above the central diagrammatic scheme. The text below the boxes, graded from white to gray shades, summarizes some important differences in cellular properties between the three developmental stages. There are significant differences in survival properties between crest cells, precursors, and Schwann cells.364 Additional differences relate to the mitogenic response to FGF, because in rat cells FGF-2 is a mitogen for Schwann cells but not for precursors. Also precursors are highly motile in vitro in comparison to Schwann cells.69,134 (Modified from Jessen, K. R., and Mirsky, R.: The Schwann cell lineage. In Kettenmann, H., and Ransom, B. R. [eds.]: Neuroglia, 2nd ed. Oxford, UK, Oxford University Press, 2004.)

the third group, comprising markers that are likely to be useful for studying the important question of glial specification from the neural crest.

Brain-Specific Fatty Acid–Binding Protein This protein belongs to the fatty acid–binding protein family. It is closely related to the peripheral myelin protein P2 and the cellular retinoic acid–binding protein. In the

developing CNS, B-FABP is found in many regions, where it is expressed by radial glial cells in particular.85,159 Satellite cells in dorsal root sensory ganglia (DRGs) in embryonic and adult mice and glial cells in embryonic nerve trunks also express B-FABP, although adult Schwann cells do not.27,159 It is not expressed by migrating crest cells or developing neurons. It is first seen at thoracic levels of E10 mouse embryos, appearing first in DRGs and slightly later in limb nerves, and is seen in ganglia and

345

Molecular Signaling in Schwann Cell Development

nerves at all spinal levels at E11.27,364 This is consistent with the appearance of early glia, including Schwann cell precursors that have been identified in mouse nerves at E12 and E13.70 It has been suggested that B-FABP might characterize glia that are common precursors of Schwann and satellite cells.27,159 In the rat, although expression in the CNS appears comparable to that in the mouse at an equivalent developmental age, expression in both satellite cells and developing nerves at E13/14 is extremely low (K. R. Jessen and R. Mirsky, unpublished results, 2003). Therefore, in this species, it is unsuitable as a marker of early glial development.

Protein Zero Although P0 was initially identified as the major protein of Schwann cell myelin, it is now evident that P0 messenger RNA (mRNA) is expressed throughout the embryonic Schwann cell lineage in immature Schwann cells and Schwann cell precursors.52,168 P0 expression levels are, however, substantially lower than the axonally induced P0 expression in myelinating cells. In the chick, P0 protein is seen in a subpopulation of migrating crest cells in stage 19 embryos and subsequently in the earliest glia-associated axons in emerging nerves.19 In rats, P0 mRNA is first detected in a population of migrating crest cells located in a region ventral to condensing DRGs at E11, perhaps indicating differentiation into the glial lineage (see below).169 P0 expression is maintained in cells associating with the earliest motor axons emanating from the spinal cord and is subsequently seen in the large majority of cells in E14/15 rat nerves (i.e., Schwann cell precursors) and in most or all immature Schwann cells. P0 mRNA is not detectable in early neurons identified by the marker ␤-III tubulin (detected by TUJ1 antibodies), showing that P0 is not expressed even in very early cells of the neuronal lineage or at later stages (E. Calle and K. R. Jessen, unpublished observations, 1998).19,115,169 These findings suggest that P0 gene expression can be used as a very early marker of glial specification in neural crest development.19,168 Two comments should be made in relation to the acceptance of P0 as a marker specifically identifying cells that have just entered the glial lineage. First, in midgestation embryos, cells with detectable levels of P0 mRNA are, in the rat, seen in the notochord and in the otic vesicle and later on in cells in nonsensory regions of the ear, none of which contains, or gives rise to, PNS glia.169 The P0 gene is therefore transiently activated in restricted cell populations in other tissues in addition to its glial-restricted P0 expression within the crest sublineage. Second, it is clear that early crest-derived cells that express immunohistochemically detectable P0 are not irreversibly committed to a glial fate, because when these cells are exposed to differentiation signals in vitro, they can be induced to generate neurons and other crest

derivatives115,219,243 (see Developmental Potential of Early Peripheral Glia below). This raises an important general issue, that of the relationship between specification and developmental potential. These issues are discussed in the section on Developmental Potential of Early Peripheral Glia below. Briefly, the view that P0-positive Schwann cell precursors are specified glia, but retain the ability to be respecified when they are challenged by alternative signals in vitro away from their normal environment, accords well with current concepts and available data of developmental plasticity and transdifferentiation. At an equivalent developmental age in the mouse, it is more difficult to detect P0 mRNA in condensing DRGs and early nerves using digoxygenin-linked in situ hybridization than in the rat (R. Mirsky and K. R. Jessen, unpublished observations, 2003). Nevertheless, in E12 mouse nerves, P0 mRNA is readily detectable in Schwann cell precursors.253

Desert Hedgehog Dhh is a mammalian member of the hedgehog family of signaling molecules. It is secreted by Schwann cells and is crucial for the proper formation of the perineurium and epineurium around peripheral nerves.252 Dhh mRNA is present as early as E12 in mouse Schwann cell precursors but is not detectable in migrating crest cells or in developing PNS neurons20 (R. Mirsky and K. R. Jessen, unpublished observations, 2003).

Oct6 Protein Oct6, a member of the POU domain transcription factor family, controls the correct timing of Schwann cell myelination.15,212,336 In the perinatal period Oct6 protein is found in the nuclei of essentially all Schwann cells, with highest levels in promyelinating cells.7,280 It is also present at low levels earlier in development because it can be detected by reverse transcription–polymerase chain reaction and seen immunohistochemically in the nuclei of most or all late mouse Schwann cell precursors from E13 and in rat nerves from E16 at the time of the transition between precursors and Schwann cells.22 It is not seen in migrating neural crest cells or in developing peripheral neurons, although it is expressed by subsets of neurons in the CNS.336 Oct6 can therefore be used in vivo as a marker for late mouse Schwann cell precursors and in the rat for late precursors/early Schwann cells.171 It should be noted, however, that Oct6 expression depends in vivo on axonal signals280 and in vitro on factors that activate cyclic AMP (cAMP)-dependent pathways (e.g., forskolin or cAMP analogues). In vitro, therefore, neither Schwann cell precursors nor Schwann cells will express Oct6 unless cAMP pathways are activated.

346

Neurobiology of the Peripheral Nervous System

Growth-Associated Protein-43, CD9, Peripheral Myelin Protein 22, and Proteolipid Protein The first three of these markers are not expressed in migrating crest, although they are found on both rat Schwann cell precursors and DRG neurons at E14. Therefore, they cannot distinguish between early neurons and early glia, although they will distinguish these cells from the crest cells from which they originate. Growth-associated protein-43 (GAP-43; neuromodulin) is a phosphoprotein that has received much attention because it is expressed at high levels in growing neurites. Schwann cells at the neuromuscular junction and in nerve trunks also express detectable but lower levels of the protein.63,282,289,308 After nerve transection, GAP-43 expression is upregulated in Schwann cells in the distal stump of the nerve and at the denervated neuromuscular junction.254,365 In Schwann cell development, GAP-43 expression can be detected as early as the precursor stage at E14 by antibodies and immunohistochemistry.134 However, some of the currently available antibodies fail to detect GAP-43 at this early stage,216 while others detect GAP-43 at low levels or in a subpopulation of precursors (R. Mirsky and K. R. Jessen, unpublished observations, 2003). A possible explanation for this discrepancy could be that early glia may express a different form of the GAP-43 protein than that found in Schwann cells and neurons. CD9 is a tetraspan cell surface protein that is involved in Schwann cell migration in vitro. It has been described in Schwann cells as early as E18 using immunohistochemistry and tissue sections and is upregulated on myelination.5,10,142 CD9 is detectable immunohistochemically on E14 Schwann cell precursors and DRG neurons 2 to 3 hours after plating in vitro, although crest cells migrating from neural tubes are initially CD9 negative under identical culture conditions (K. R. Jessen, unpublished observations, 2003). In vitro, significant amounts of CD9 are shed from Schwann cell precursors onto the culture substrate. Peripheral myelin protein 22 (PMP22) is one of the important protein components of the PNS myelin sheath and, in common with P0, mutations or duplications in the gene frequently lead to peripheral neuropathy.229 It is present in some non-neural tissues in embryonic development,9 and is also expressed by some CNS and PNS neurons.115,250,251 PMP22 mRNA is present in the glia of E12 spinal nerves in the rat, and PMP22 mRNA and protein are expressed by E14 Schwann cell precursors. In contrast, the mRNA cannot be detected in migrating crest cells in E10 rat embryos.22,115 Proteolipid protein (PLP) is the major CNS myelin protein, but low quantities are also present in myelin sheaths in adult peripheral nerves. Null mutations in PLP can result in peripheral neuropathy.98 During

development, it is present in some non-neural tissues such as notochord and otic vesicle.304,332 In migrating mouse neural crest cells, mRNA can be detected as early as E9.5 in condensing sympathetic, trigeminal, and spinal ganglia and in peripheral nerves.304 In addition to glia, it may be expressed by neuroblasts or neurons at these early stages of development (B. Zalc, personal communication, 2003). It is therefore likely to detect neural cells at an early stage of their differentiation from the neural crest.

NEURAL CREST ORIGINS OF SCHWANN CELLS The neural crest is a transient group of cells that delaminates from the dorsal part of the neural tube during embryonic development. In the trunk region, cells of the neural crest give rise to glial cells; neurons of sensory, sympathetic, and parasympathetic ganglia; chromaffin cells; and melanocytes. The cardiac crest, situated in the most anterior part of the trunk, also gives rise to connective tissue and smooth muscle cells, while the cephalic crest, situated in the head region, generates in addition cartilage and bone cells.166 At the onset of migration, some neural crest cells may have already entered the glial lineage, while other cells appear to start glial development later as they migrate ventrally along the neural tube.121 The nature of the extrinsic signals that initiate glial development from crest cells is not established. As mentioned above, in embryonic nerves two growth factors, neuregulin-1 and endothelin,26,99 participate in the regulation of early Schwann cell development. At present it seems unlikely that these factors are required to trigger glial development from multipotent crest cells in vivo. The roles of neuregulin-1 and endothelin are discussed below in the section on Importance of Neuregulin-1 in Schwann Cell Development and the section on Regulation of the Precursor–Schwann Cell Transition, respectively. Cell culture studies indicate that bone morphogenetic proteins (BMPs), which are important for the generation of sympathetic neurons,283,290 may act as negative regulators of gliogenesis during crest development.290 Recent studies indicate that Notch activation may be involved in peripheral gliogenesis.95,97,155,218,346 (Notch is a transmembrane receptor activated by binding of membraneassociated ligands of the Delta family on neighboring cells. It is involved in cell fate determination in many different tissues.) This would be in line with observations on CNS glial development, because enforced Notch activation promotes the emergence of CNS glial cells in vivo.95,97,102,326 The evidence for Notch signaling in PNS gliogenesis is, however, less complete. It has been shown that, during division of avian crest cells, asymmetrical segregation of

Molecular Signaling in Schwann Cell Development

NUMB, a likely Notch antagonist, occurs. It has been suggested that, as a result, cells with higher levels of NUMB would be biased toward neuronal development within sensory ganglia.346 Indirect support for this prediction is provided by the finding that, in mice lacking NUMB, DRG neurons fail to form.374 Interestingly, in these animals sympathetic neurons are generated normally. Notch activation and neuregulin-1 share the ability to suppress neurogenesis from neural crest cells in vitro155,218,291,346 (see Importance of Neuregulin-1 in Schwann Cell Development below). In another parallel with neuregulin, it has been more difficult to demonstrate conclusively that Notch activation inductively triggers glial differentiation from neural crest cells. Clonal analysis experiments of migrating neural crest cells, treated with a soluble form of the Notch ligand Delta to activate Notch, reveal that Notch promotes glial differentiation, measured by an increase in the number of clones that contain GFAPpositive cells.155,218 In these experiments, lower levels of glial differentiation take place without Delta application, thus demonstrating that gliogenic signals are available to at least some of the cells without enforced Notch activation. Although this may be due to intrinsic Notch signaling, it is equally conceivable that the role of Notch activation in these experiments is to promote the action of other gliogenic signals. An established function of Notch is to block entry into nonglial lineages (see above). This could provide an indirect mechanism for an increase in glial numbers because it might well prolong the time during which crest cells are receptive to any instructive gliogenic signals present in the culture system. In conclusion, in the developing neural crest there is evidence that early developing neurons activate Notch in adjacent cells, which could be a mechanism for preventing excessive neurogenesis while allowing gliogenesis. It is still uncertain whether Notch activation also acts inductively to trigger gliogenesis or whether Notch promotes glial differentiation only indirectly. Experiments on cells from E14 rat nerves have raised the additional possibility that Notch signaling promotes the generation of Schwann cells from Schwann cell precursors (see Developmental Potential of Early Peripheral Glia below). The development of glia from crest cells depends on the transcription factor Sox10.27,157,243,253,302,303,353 In several other systems, Sox family members are also involved in regulating development.352 As the neural tube closes, Sox10 mRNA is expressed in restricted dorsal areas of the neural tube that give rise to neural crest cells.157 Sox10 mRNA is seen subsequently in migrating crest cells,27,257,303 and in vitro studies show that these cells all express Sox10 protein.243 In developing glia, both in ganglia and along nerve trunks, Sox10 expression persists, but in early developing neurons it is downregulated.157 During embryonic development the levels of Sox10 mRNA in peripheral nerves decline and levels in

347

Schwann cells in neonatal nerves are much lower than those in neural crest cells (U. Lange and K. R. Jessen, unpublished observations, 2002). The effects of Sox10 inactivation have been studied using both Sox10-null mice and dominant megacolon (Dom) mice, which have spontaneous nonsense or frameshift mutations in the Sox10 gene. In these mice, early peripheral glial cells are missing.27 Neither satellite cells within DRGs nor Schwann cell precursors in emerging peripheral nerves (identified in normal mice by expression of B-FABP) are present. In the mouse, B-FABP distinguishes early glial cells, which express B-FABP protein, from neural crest cells and early neurons, which are B-FABP negative (see Brain-Specific Fatty Acid–Binding Protein above). Despite the lack of B-FABP–positive glial precursors in DRGs in Sox10-deficient mice, sensory neurons, identified by expression of ␤-III tubulin, are initially generated in normal numbers, although they die later in development.302 This suggests that Sox10 has a particular role to play in glial development. One important function of Sox10 is likely to be the maintenance of ErbB3 expression. ErbB3, a key neuregulin-1 receptor, is expressed in crest cells as they emerge from the neural tube and in early glial cells. In Sox10-mutant mice, ErbB3 receptor mRNA appears on schedule in migrating crest cells, but as the crest cells migrate away from the dorsal neural tube and begin to differentiate into neurons and glia in the condensing DRGs, expression is not sustained, as it would be in normal mice.27 Although the survival of migrating crest cells in vivo does not appear to require neuregulin-1,28 the survival of Schwann cell precursors in nerve trunks depends on neuregulin-1 signaling and ErbB3 receptors (see Importance of Neuregulin-1 in Schwann Cell Development below). Neuregulin-1 is also a mitogen for these cells.28,69,70 Therefore, in Sox10 mutants, the predicted outcome of ErbB3 downregulation would be a selective reduction in the number of Schwann cell precursors in peripheral nerve trunks. This is in line with the finding that the number of Schwann cell precursors is severely reduced in nerve trunks of both Sox10 and ErbB3 mutants. Conversely, the principal reason for the absence of glia in Sox10-mutant DRGs cannot be due to the lack of neuregulin-1 signaling, because the development of DRG glia appears normal in mice in which neuregulin-1 signaling has been inactivated (see Importance of Neuregulin-1 in Schwann Cell Development below). Increased cell death and decreased proliferation of non-neuronal cells within condensing ganglia are seen in Sox10 mutants, although cell death in migrating crest cells prior to the condensation of the ganglia is normal.27,243,302 This would be explained if Sox10 were needed for survival signaling in DRG glia by factors other than neuregulin-1, such as insulin-like growth factors (IGFs) or fibroblast growth factors (FGFs).364 Another possible explanation for the lack of DRG glia

348

Neurobiology of the Peripheral Nervous System

might be that glial specification (i.e., the process by which crest cells change to become early glia) requires Sox10.27 Early DRGs of Sox10 mice, which lack B-FABP–positive glial cells (see above), contain, besides neurons, a cell population that appears to retain a neural crest cell phenotype, and a few such cells are present in nerve trunks.27,302 This is consistent with the idea that glial specification is blocked in the absence of Sox10, although crest cells thrive and can generate other cell types, including neurons. Cell culture experiments give indirect support to this idea. In Sox10 mutants, a proportion of cultured p75 low-affinity neurotrophin receptor (p75NTR)–positive non-neuronal cells from E13 mouse DRG, presumably undifferentiated neural crest cells,27 leaves the neural lineage and adopts a smooth muscle–like phenotype, one of the cell types in the crest repertoire, rather than the glial phenotype that cells from wild-type DRG adopt.243 The observation that enforced expression of Sox10 induces expression of the endogenous P0 gene, a marker of crest cell differentiation (see Protein Zero above), in the neuroblastoma cell line N2A provides a further link between Sox10 and glial differentiation.253 Thus at least two plausible mechanisms explain the effects of Sox10 mutations on early glial development, one relating to neuregulin-1 signaling and the other to glial specification. A wider role for Sox10 in the development of neural crest derivatives is indicated by deficiencies in pigmentation and in the enteric nervous system.352

SCHWANN CELL PRECURSORS The Schwann cell precursor defines the first stage of glial differentiation in peripheral nerve trunks.69,134,136 The large majority of cells in the limb nerves of E14/15 rats (E12/13 mice) are Schwann cell precursors. Analogous cells in dorsal and ventral roots and early ganglia may not have identical properties and are also likely to differ in the timing of appearance of glial differentiation markers.364 Initially, precursors appear mostly at the edge of nerves but are also found within more mature embryonic nerves. They are always in close apposition to axons. Their extensive sheetlike processes contact and form junctions with processes from neighboring cells, surrounding large groups of axons, which at this age are mostly of similar size (Fig. 16–3). In contrast to older nerves, in which there is abundant fibrous connective tissue containing blood vessels surrounding axon–Schwann cell units within a nerve fascicle, E14 rat or E12 mouse nerves possess no significant connective tissue or vessels. Examination of glia in rat hind limb nerves at various embryonic ages shows that, in E14 and E15 nerves, the large majority of cells have the phenotype of Schwann cell precursors, while by E18 most of these have

converted to cells with the distinct properties of immature Schwann cells. The precursor–Schwann cell transition therefore centers on the 16th embryonic day in these nerves. In mice the corresponding conversion occurs 2 days earlier.70 Early indications that there were significant differences between Schwann cells and Schwann cell precursors came from a study of cell survival and expression of the calciumbinding protein S100 in embryonic rat nerves. When cells from E14 and E18 nerves were compared, E14 cells showed very little cytoplasmic S100 immunoreactivity relative to older cells. Furthermore, while Schwann cells from E18 and older nerves survived at moderate or higher densities without added factors, the E14 cells died rapidly in vitro without added survival factors, irrespective of cell density. Autocrine survival circuits in Schwann cells were later revealed to be the cause of this difference in survival regulation. Schwann cell precursors do not possess autocrine survival circuits, relying instead on paracrine neuregulin-1 signaling from axons.134,200 Schwann cell precursors can now be differentiated from both the migrating crest cells they are derived from and the immature Schwann cells that they generate on the basis of a number of features (see Fig. 16–2). They are therefore specified glial cells. Presumably, during normal development they also have a single fate to become Schwann cells. This does not imply that Schwann cell precursors are determined or irreversibly committed to a glial fate (see Developmental Potential of Early Peripheral Glia below). The essential function of Schwann cell precursors is to generate Schwann cells, but an important additional role for these cells is suggested by studies on mutant mice lacking isoform III of neuregulin-1, the ErbB3 or ErbB2 neuregulin receptors, or the transcription factor Sox10. Schwann cell precursors are either missing or their number is severely reduced in all of these mutants. While DRG neurons and motor neurons are initially generated normally, by E14 and E18, respectively, at limb levels in the trunk the majority of these neurons have died. This cell death is most likely due to the absence of glialderived factors that are needed for the survival of sensory and motor neurons.27,266 Thus a key function of Schwann cell precursors and early Schwann cells may be to regulate the survival of discrete pools of embryonic CNS and PNS neurons. It is an interesting possibility that these cells control neuronal survival through the mechanism of back-signaling by the intracellular domain of neuregulin-1.11 If this were the case, neurons and glia in embryonic nerves would ensure mutual survival through a bidirectional effect of the same molecular interaction, the binding of neuregulin-1 to ErbB2/ErbB3 receptors. In the neuregulin mutants, loss of axonal contact with peripheral targets may also be an important contributor to sensory and motor neuron death.362

Molecular Signaling in Schwann Cell Development

349

FIGURE 16–3 Cells of the Schwann cell lineage as they appear during peripheral nerve development. Top, Four Schwann cell precursors (P1 through P4) in the hind limb nerves of a rat embryo at E15. The nuclei of three of the precursors (P1, P3, and P4) are visible. Schwann cell precursors form close contacts (arrows) with each other, and are either embedded between axons inside the nerve, as shown here, or found in close apposition to axons at the surface of the nerve. Because connective tissue spaces and extracellular matrix are essentially absent, these nerves are much more compact than older ones. The axons also have a smaller and more uniform diameter than those seen in mature nerves. (Courtesy of Y. Hashimoto.) Bar: 1 ␮m. Middle, Schwann cells in the sciatic nerve of a newborn mouse (NB). Immature Schwann cells (S**) communally envelop a group of axons of various diameters. Other cells (S) are at a very early stage of myelination, the promyelination stage, having formed a 1:1 relationship with large axons (A). S* has progressed slightly farther along the myelin lineage and is starting to form a compact sheath around the axon (A*). Part of a thin myelin sheath can be seen on the extreme right-hand side of the field (arrow). Note that, in contrast to the compact arrangement seen earlier in development, perinatal nerves contain considerable collagen (C) and extracellular space. Bar: 0.5 ␮m. Bottom, Mature Schwann cells in transverse section of adult rat sciatic nerve (AD). Left, A myelinating Schwann cell forms a compact multilayered sheath (M) around a single large-diameter axon (A*). Parts of other myelinated fibers are present. Bar: 1 ␮m. Right, Cross section through the nucleus of a nonmyelinating Schwann cell (N-M) that ensheathes 13 axons (e.g., A), each lying in a separate trough in the cell surface. Nonmyelinating Schwann cells can also ensheath singly. Myelin sheaths (M) of neighboring axons are visible, and the axon–Schwann cell units are surrounded by collagen-rich extracellular spaces (C). (Middle and bottom panels courtesy of R. M. King and P. K. Thomas.) Bar: 1 ␮m. (Modified from Jessen, K. R., and Mirsky, R.: Schwann cell development. In Lazzarini, R. A. [ed.]: Myelin Biology and Disorders, Vol. 1. San Diego, CA, Academic Press, p. 329, 2003.)

DEVELOPMENTAL POTENTIAL OF EARLY PERIPHERAL GLIA During normal development Schwann cell precursors are unlikely to generate cells other than Schwann cells, such as neurons or melanocytes. Nevertheless, it is clear that,

when cells from early nerves or DRGs of chick, rat, or mouse are exposed to growth factors in cell culture, they can be directed to generate other cell types. There could be two principal reasons for this (see below). Possibly, the early glia of peripheral nerves are not yet irreversibly committed to the glial lineage. If such

350

Neurobiology of the Peripheral Nervous System

cells are exposed in vitro to a signaling environment significantly different from that which they normally encounter in vivo, their potential to generate other crest derivatives could be revealed. The other possibility is that, in addition to specified glia, early nerves contain a small, undifferentiated cell population that remains similar to migrating neural crest cells in molecular phenotype and developmental potential. Studies on avian systems provide the earliest evidence that early nerve trunks can in vitro give rise to cells that are not normally generated in vivo. When cultured in complex media, nerve segments from E4-E6 chick embryos, presumed to consist largely of Schwann cells with a minimal number of fibroblasts, gave rise to pigmented cells of the melanocyte lineage.294 Melanocytes were not generated from nerves of older embryos in these studies, although recent experiments indicate that cells from adult mouse nerves can be diverted to melanocyte differentiation. Nerve injury triggers this melanogenesis, which is more extensive in mice containing a single copy of the neurofibromatosis type 1 (Nf1) gene than in wild-type mice.267 Experiments on early quail nerve segments treated with FGF-2 also showed that melanocytes can arise from cells that were P0 positive, and that were therefore presumed to be in the glial lineage. Thus, FGF-mediated signaling results in transdifferentiation or respecification of Schwann cell precursors that would normally give rise to Schwann cells.294,316 Two other factors, endothelin-3 and Steel factor, have also been implicated in the respecification.228 Cultures of sensory ganglia can also give rise to melanocytes.62,232 Taken together, these studies show that specified P0-positive avian Schwann cell precursors can be diverted to generate at least one other crest derivative (i.e., melanocytes), and that therefore they are not irreversibly committed to the Schwann cell lineage. Related observations have been made on early rat nerves.219 In these studies, cells dissociated from E14 rat nerves were divided into subpopulations by fluorescenceactivated cell sorting (FACS) prior to culturing, using antibodies to myelin protein P0 and p75NTR. Cells initially sorted as P0 positive and p75NTR positive were analyzed at clonal density after 2 weeks of culture in complex medium containing chick embryo extract and retinoic acid. Many of these cells had given rise to GFAP-positive Schwann cells, but also to neurons and/or cells that express smooth muscle actin. BMP2 promoted the development of neurons from many of the sorted cells, and neuregulin-1 blocked neurogenesis. These signals have similar effects on migrating crest cells.219 A comparable strategy was used to investigate the developmental potential of non-neuronal cells from E14 rat DRG. These cells coexpress p75NTR, P0, and PMP22 and are therefore likely to be cells that in normal unperturbed development would give rise to satellite glial cells. Nevertheless, in vitro they were able to generate non–glial crest derivatives, including neurons and smooth muscle

cells.115 Taken together, these experiments are consistent with the idea, derived from studies on avian nerves, that specified P0-positive early glial cells in the PNS retain some developmental plasticity, although there is some uncertainty about the FACS methodology used in the experiments of Morrison et al.219 (see below). As discussed previously, a distinct population of multipotent crest-like cells that could be diverted to nonglial lineages might be present in embryonic nerves, in addition to Schwann cell precursors. It has been suggested that these cells can be distinguished from Schwann cell precursors by absence of P0 expression and that this type of P0-negative cell accounts for approximately 15% of the total number of cells dissociated from E14 rat nerves.218,219,358 The evidence for this interesting possibility is not yet complete. Initially, it came from experiments in which cells from E14 rat nerves were fractionated using FACS (see above).155,218,219,358 Cells that were P0 negative and p75NTR positive, a molecular phenotype expected of migrating neural crest cells, were able to self-renew and generate neurons, glia, or “myofibroblasts” in response to BMPs, neuregulin-1, and transforming growth factor-␤ (TGF-␤), responses that are qualitatively similar to those of migrating crest cells. Subsequent studies demonstrated, however, that the E14 nerve-derived cells were about 10 times less sensitive to the neurogenic signal BMP2 than crest cells derived directly from neural tube cultures. Using an in vivo assay, the nerve-derived cells were also shown to be strongly biased toward generating glia, in comparison to tube-derived crest cells.358 Similarly, in vitro the P0-negative, p75NTR-positive E14 nerve-derived cells were more sensitive to the gliogenic actions of Delta-Notch signaling than migrating crest cells. This change may be caused by reduced expression of the putative Notch inhibitor NUMB and elevated expression of Notch in these cells.155 Thus cells that FACS sort from E14 nerves as P0 negative and p75NTR positive retain more than one developmental option, at least in vitro, but in comparison with migrating crest cells are strongly biased toward gliogenesis. As suggested by Morrison et al.,219 these cells may represent a distinct population of resident neural crest cell–like cells in embryonic nerves. Another interpretation of these results is possible, however, particularly since the methodology used for obtaining them raises some concern. These cells were FACS sorted as P0 negative and p75NTR positive using the P07 anti-P0 antibody described by Archelos et al.6 Two issues raise questions about relying on the P07 antibody to sort cells from embryonic nerves. Although it was made against an extracellular portion of P0, binding of this antibody to living, unfixed Schwann cells or their precursors cannot be detected immunohistochemically. Significant binding of this antibody to cell surfaces appears to require previous fixation. Second, another population of cells that can be sorted from

Molecular Signaling in Schwann Cell Development

the same tissue (i.e., P0-positive, p75NTR-negative cells) forms neither Schwann cells in the presence of neuregulin-1, as would be expected of cells expressing P0, nor neurons in the presence of BMP2/4.219 The absence of neuronal and glial differentiation from these cells in the presence of BMPs and neuregulin, respectively, as well as the population size (~20% of the total cell suspension derived from freshly dissected E14 nerves) and the lack of p75NTR expression, are consistent with the idea that these might be non-neural, mesenchymal cells adhering to dissected E14 nerves, as indicated by Morrison et al.219 In this case, the expression of P0 by these cells would be puzzling because such cells are not reported to express P0. The identity of the P0-negative, p75NTR-positive FACSsorted cells from E14 nerves as a distinct population of crest-like cells is therefore ambiguous. One alternative is that the cells in question represent Schwann cell precursors, a possibility that would rest on the presumption that FACS based on this P0 antibody is not meaningful. The observation that these cells can adopt more than one lineage in vitro does not illuminate the issue because, as discussed earlier (see Protein Zero above), plasticity may well be a property of both migrating crest cells and Schwann cell precursors. The likelihood that the P0-negative, p75NTR-positive cells sorted in these experiments are Schwann cell precursors would be consistent with their strong bias toward gliogenesis. It would also accord with the action of Notch activation in these cells, which is to induce irreversible commitment to the glial lineage.219 It is unlikely that such a commitment would occur as migrating crest cells generate Schwann cell precursors, because, as discussed above, precursors show lineage plasticity. Conversely, Schwann cells are much more lineage restricted.56,219 It seems more likely that the cells from E14 nerves that become irrevocably confined to glial differentiation on Notch activation are Schwann cell precursors. This implies that a function of Notch activation is to promote the generation of Schwann cells from Schwann cell precursors, an idea that has not been tested. To conclude, at least two plausible explanations exist for the phenotypic plasticity observed in cells from early embryonic nerves, and they are not mutually exclusive. First, although Schwann cell precursors are specified glial cells, they may not yet be determined (i.e., irreversibly committed) to a glial fate. Second, a minority population of cells with a developmental potential and phenotype similar to that of migrating crest cells might well be present in early nerves. New studies on stem cell development, respecification, and transdifferentiation of one cell type into another have challenged earlier views on commitment and phenotypic plasticity.93,339 Taken together, this work suggests that the differentiated state is less fixed than previously envisaged. The idea that specified precursor cells can be reprogrammed in vitro by a variety of external signals is no longer surprising

351

and is consistent with a large number of experiments. The oligodendrocyte lineage provides a parallel example to the Schwann cell precursor.153 This work shows that cells that are unambiguously specified as oligodendrocyte precursors, and are therefore in a glial lineage, can be converted into multipotential stem cells by extracellular signals. An important conclusion resulting from this and other work is that, during development, a significant time lag may occur between specification of a cell to a particular lineage and determination, a state that implies irreversible commitment. In Schwann cell development this is probably reflected in the observation that a broader spectrum of cells can be generated from rat nerves at E14/15, which contain Schwann cell precursors, than from E17 nerves, in which a majority of the cells are Schwann cells.219 Comparable observations have been made using chick DRGs and nerves.56

IMPORTANCE OF NEUREGULIN-1 IN SCHWANN CELL DEVELOPMENT Neuregulin-1 is involved in all stages of Schwann cell development from the neural crest stage through to fully differentiated Schwann cells and is the most important known cell-cell signal regulating the lineage.1,34,99,173 It has a long history of association with Schwann cells. In an impure form, it was initially isolated as one of the first mitogens for cultured rat Schwann cells, being later identified as the type II isoform of neuregulin-1, GGF.101,172,191 It is now considered to be the major axonally derived Schwann cell mitogen (see Cell Culture Studies on Schwann Cell Proliferation and Differentiation below). In recent years, three other members of the neuregulin gene family (neuregulins 2 through 4) have been identified.36,118,372 Neuregulin-2 and -3 are expressed in the nervous system, but their function in peripheral nerve development, if any, is unknown.1 As mentioned earlier (see Schwann Cell Precursors above), mice lacking either neuregulin-1 or the neuregulin receptors ErbB2 or ErbB3 have embryonic nerve trunks that are essentially devoid of Schwann cell precursors, although, interestingly, development of satellite glia within DRGs appears to be normal, as is the initial development of DRG neurons.99,204,216,266,361 Sympathetic ganglia, however, fail to develop normally in these animals.28 These striking phenotypes have provided essential clues for understanding the role of neuregulin signaling in embryonic development of the PNS. Because satellite glia within ganglia and Schwann cells along nerve trunks develop from neural crest cells, the normal occurrence of DRG satellite cells in neuregulin mutants suggests that neuregulin-1 signaling is not obligatory for glial differentiation from the neural crest in vivo. This accords with the observation that GFAP-positive glia develop readily from cultured neural crest cells with or without added

352

Neurobiology of the Peripheral Nervous System

neuregulin.291 Although neuregulin strongly represses the formation of neurons from cultured crest cells,291 the neuregulin mutants do not provide any obvious indication, such as overproduction of neurons, to support the idea that neuregulin-1 constrains neurogenesis in vivo. Nevertheless, it is possible that this effect of neuregulin is masked in vivo by alternative factors that are absent in the culture system. One important reason for the remarkable reduction or absence of Schwann cell precursors from nerve trunks in neuregulin mutants is likely to be connected to the role of neuregulin-1 as a survival factor for Schwann cell precursors and early Schwann cells.69,136,171 Unlike Schwann cells, Schwann cell precursors are unable to support their own survival by autocrine mechanisms and depend instead on paracrine signals from axons (see Regulation of the Precursor–Schwann Cell Transition, and Signals That Control Schwann Cell Survival, below).134,200 In developing nerves they are closely apposed to axons, and die rapidly when they are removed from axonal contact either by culture in the absence of axons or by nerve transection in vivo.57,69,134,360 Neuregulin-1 isoforms are found in vivo in embryonic DRGs and motor neurons and accumulate along axonal tracts.81,185,191,239 Furthermore, cell culture studies show that DRG neurons are a potent source of both axon-associated and secreted survival signals that have been identified as neuregulin-1 isoforms.69 Neuregulin-1 is therefore available at the right time and place to enable the survival of precursors in embryonic spinal nerves. This is further supported by observations on mutants lacking isoform III of neuregulin-1, the major isoform in peripheral nerves.362 In these mice, early Schwann cell precursors start populating nerves at E11, but by E14, a time when precursors are converting rapidly to Schwann cells, the number of cells in the nerves is severely depleted. This argues against the idea that defective migration of precursors into nerve trunks is the major reason for absence of precursors in the nerves of neuregulin mutants (see below). Neuregulin has also been found to support the survival of early glia in embryonic chick nerves.360 Taken together, these observations strongly support the view that a major function of neuregulin-1 in embryonic nerves is to ensure the survival of Schwann cell precursors.69,136 The other factor that may contribute to the lack of Schwann cell precursors in developing nerves of mice deficient in neuregulin signaling is reduced cell migration. Neuregulin-1 stimulates migration of postnatal Schwann cells in vitro and glial migration from newborn mice DRG explants.201,216 Furthermore, in mice lacking ErbB3 receptors, neural crest cell migration along the ventromedial pathway is impaired. This is thought to be a major factor in the failure to develop sympathetic ganglia.28 Finally, in cell culture, the migration of glia from E12 DRG of ErbB2mutant mice is reduced.216 However, neuregulin-1 has no

effect on migration of glia from E12 DRGs from normal mice in vitro, and neural crest cells migrate out normally from the neural tube, and from DRGs in appropriate locations, in mice lacking neuregulin-1 or the ErbB3 receptor. While the evidence relating to migration is therefore ambiguous, it is obviously possible that reduced migration is a contributory factor to the sparsity of precursors in nerves of mice with deficient neuregulin signaling. From studies on neuregulin-mutant mice, it has been concluded that neuregulin-1 isoform I (NDF, ARIA), isoform II (GGF), and, in particular, isoform III (SMDF, cysteine-rich domain [CRD]-NRG-1) are the most important isoforms for PNS and Schwann cell development.34,90,205 Animals with inactivation of the neuregulin-1 gene, and therefore of all three isoforms, or of the neuregulin receptors ErbB2 or ErbB3 have the fewest Schwann cell precursors in their peripheral nerve trunks. Mice with selective inactivation of the neuregulin-1 isoform III initially show a milder phenotype, although at later stages Schwann cell precursor numbers are severely depleted and Schwann cells are essentially absent from the distal portion of the nerves (see above).362 In contrast, normal glial development is seen in mice lacking neuregulin-1 isoforms I and II, although muscle spindle development is deficient.122,205 Therefore, of the three main neuregulin-1 isoforms, isoform III is most crucial for Schwann cell development. Isoforms I and II contain an immunoglobulin domain and bind heparin, and may therefore associate with cell surfaces,1 whereas isoform III is mostly expressed as a transmembrane protein.347 Membrane bound, but not soluble, isoform III induces S100 and Oct6 expression in P0-positive cultured non-neuronal cells from embryonic rat DRGs.171 This suggests, first, that membranebound forms may have functions not shared by soluble forms and, second, that one of the functions of neuregulin-1 is to accelerate the precursor–Schwann cell transition (see Regulation of the Precursor–Schwann Cell Transition).

NEUREGULIN-1 SIGNALING PATHWAYS IN SCHWANN CELLS Neuregulin signaling is important in cell types other than Schwann cells, and the intracellular signaling pathways that transduce neuregulin signals have been extensively studied. However, it is not known how many of the interacting proteins involved in the neuregulin signaling pathway identified in other cells are used by Schwann cells, as reviewed by several authors.33,99,165,175,234 In Schwann cells, neuregulin-1 induces proliferation, phosphorylation, and dimerization of the neuregulin receptors ErbB2 and ErbB3, processes potentiated by interaction with the transmembrane protein CD44, which also promotes neuregulin-1–induced cell survival.108,258,266,269,296,318,322,345,361 (CD44 is widely expressed. The extracellular matrix component hyaluronan binds to it, and it plays a role in migration, adhesion, and survival of

Molecular Signaling in Schwann Cell Development

some cell types.) In Schwann cells, as in other cell types, neuregulin-1 induces activation of phosphatidylinositol 3 (PI3) kinase and its target Akt, and mitogen-activated protein (MAP) kinase pp44/42. Inhibition of the PI3 kinase pathway inhibits DNA synthesis in Schwann cells in response to applied neuregulin-1 and contact with neurites.147,198 In these studies, inhibition of the MAP kinase pathway had no effect, suggesting that the PI3 kinase pathway is more important in the proliferative response. Schwann cell survival in response to neuregulin-1 also requires activation of PI3 kinase and its effector Akt and subsequent phosphorylation and inactivation of the proapoptotic protein BAD.70,179,198,200 In some studies, but not in all, inhibition of the MAP kinase pathway also caused considerable cell death.198,200,249 In Schwann cell precursors, inhibition of either the PI3 kinase or the MAP kinase pathway completely inhibits survival in neuregulin-1.70 Neuregulin-1 also regulates the Jun N-terminal kinase (JNK) and p38 stress kinase pathways, both of which are likely to be important in Schwann cell biology (see Transcription Factors Involved in Myelination below). In breast cancer cells and in Schwann cells, neuregulin-1 activates the p38 MAP kinase pathway, and inhibition of this pathway prevents neuregulin-1–induced proliferation230 (D. B. Parkinson, A. Bhaskaran, K. R. Jessen, and R. Mirsky, unpublished observations, 2003). The JNK pathway is also activated by neuregulin-1. Phosphorylation (activation) of c-Jun is required for Schwann cell proliferation in response to neuregulin,249 and is also required for TGF-␤–induced cell death (see Signals That Control Schwann Cell Survival below).248 Elevation of intracellular cAMP levels and protein kinase A (PKA) activation have profound effects on both proliferation and differentiation of Schwann cells in vitro (see Transcription Factors Involved in Myelination below). Neuregulin-induced Schwann cell proliferation involves and requires a delayed rise in intracellular cAMP levels and PKA activation.147 This response may account for the observation that, unlike most other growth factors, neuregulin-1 promotes DNA synthesis in the absence of other agents that stimulate cAMP pathways, although both forskolin (which elevates cAMP levels) and IGF potentiate the proliferative response to neuregulin-1.61,310 In Schwann cells, neuregulin-1 also activates pp70s6 kinase (a downstream effector of PI3 kinase; see below), PAK65 (a component of stress-activated signaling pathways), p95RSK2 (a cAMP response element [CREB] kinase activated by MAP kinase), and the transcription factor CREB by phosphorylation on serine 133. It also stimulates cyclin D1 expression and phosphorylation of retinoblastoma protein, both enhanced by addition of forskolin.258,324 (Cyclin D1 and retinoblastoma protein are both involved in regulation of the cell cycle.) There is also evidence that neuregulin-1 signaling

353

involves protein kinase C, another potential effector of PI3 kinase.276,369 Both neuregulin and autocrine-mediated survival of Schwann cells (see Signals That Control Schwann Cell Survival below) depend on the activity of Ets transcription factors, several of which are made by Schwann cells.249 In a related system, neuregulin-mediated stimulation of acetylcholine receptor synthesis at the neuromuscular junction also depends on an Ets transcription factor binding site.94 (Ets transcription factors bind to a specific DNA core sequence [GGAA/T] and cooperate with other transcription factors and cofactors to regulate proliferation, apoptosis, and development and differentiation of specific cell lineages.)

REGULATION OF THE PRECURSOR–SCHWANN CELL TRANSITION When Schwann cells are generated from Schwann cell precursors in vivo, a coordinated change in a number of disparate and apparently unrelated phenotypic features occurs. There are modifications in antigenic expression (see Fig. 16–2), survival regulation, response to mitogens, motility, and cell-cell interactions (see Schwann Cell Precursors above). Similar changes can be accomplished in vitro by culture of precursors from rat E14 nerves in the absence of neurons in chemically defined medium containing neuregulin-1 for 4 days (E14 ⫹ 4 ⫽ E18). When these cells are dissociated and re-plated, over 80% of them have a Schwann cell rather than precursor phenotype with respect to the presence of autocrine survival loops (see Signals That Control Schwann Cell Survival below), antigenic phenotype, and mitogenic responses.69 This demonstrates that neuregulin-1 alone is sufficient to support not only precursor survival but also lineage progression. A minority of precursors fail to convert to Schwann cells in these simplified conditions, and it is probable that the appropriate rate of Schwann cell generation in vivo involves multiple signals.26,171 Endothelin, acting through the endothelin B receptor, is likely to be one of these signals.26 The survival of Schwann cell precursors in vitro is supported by endothelins 1 through 3, and endothelins and endothelin receptors are present in embryonic nerves. Unlike neuregulin, endothelin does not induce precursor proliferation, and these two factors also affect lineage progression differently. In precursor cultures exposed to endothelin alone, Schwann cells are generated very slowly, but in the combined presence of endothelin and neuregulin-1, the rate of Schwann cell generation increases and is intermediate between that seen in endothelin alone or neuregulin-1 alone. This result suggests that neuregulin actively promotes the transition of precursors to Schwann cells. It

354

Neurobiology of the Peripheral Nervous System

also indicates that endothelins act to slow down Schwann cell generation, because Schwann cells are generated more slowly in endothelin plus neuregulin-1 than in neuregulin-1 alone. In vivo experiments on the spotting lethal rat confirm this idea. In these animals, endothelin B receptors are not functional and Schwann cell generation occurs ahead of schedule, as expected if, in normal nerves, endothelin B receptor activation acts to brake the rate of lineage progression.26 It is possible that FGF-2 may be a third regulatory signal in this system, because addition of FGF-2 accelerates Schwann cell generation in vitro in comparison with that seen in neuregulin-1 alone.70 This remains to be confirmed in vivo. The transition from precursors to Schwann cells involves systematic alterations in gene expression, implying, in turn, the existence of a coordinating gene control mechanism. At present, however, only one transcription factor, AP2␣, has been linked to this process.311 At the precursor–Schwann cell transition in vivo, both in rats and in mice, AP2␣ expression is sharply downregulated, and in vitro, enforced expression of AP2␣ in precursors delays their conversion to Schwann cells, suggesting that the rate of AP2␣ downregulation may be one of the factors that determines the rate of Schwann cell generation.311 There is evidence that the time course of differentiation of a related cell type, the oligodendrocyte progenitor cell, is controlled by a member of the Id family of HLH genes, Id4, the levels of which drop as the cell differentiates into an oligodendrocyte.152,348 Although Schwann cell precursors and Schwann cells coexpress Id1 through Id4, these factors are not obviously regulated during embryonic nerve development, and at present there is no evidence that Id proteins are involved in controlling the speed of the precursor–Schwann cell transition.

SIGNALS THAT CONTROL SCHWANN CELL SURVIVAL Axon-associated neuregulin-169,108,322,340 and autocrine Schwann cell signals51,200 are two signals that play a major role in promoting the survival of developing Schwann cells. Transection of neonatal sciatic nerves results in the death of all teloglia at the neuromuscular junction.340 Although most Schwann cells within the main nerve trunk survive, there is also increased Schwann cell death within the nerve, indicating a partial dependence on axonal signals. Since the cells can be rescued by application of exogenous neuregulin-1, it is likely that the death is caused, at least in part, by loss of contact with axon-associated neuregulin as axons degenerate after transection.108,322,340 After the first postnatal week, there is no comparable Schwann cell death immediately following nerve transection in rats, and in adult animals Schwann cells in the distal stumps survive for several months after nerve transection, although their number slowly decreases and they respond less to extrinsic signals.108,178,319,340 Significant Schwann cell death is, however, seen in regenerating adult nerves at 3 weeks postcrush at a time when there is a requirement to match the number of axons and Schwann cells.89 Schwann cell survival in the absence of axons is crucial for axonal regeneration after nerve injury, because Schwann cells provide both adhesive substrates and trophic factors that promote axonal growth. A major reason for Schwann cell survival in these conditions is likely to be the presence of autocrine circuits, which enable Schwann cells to support their own survival51,200 (Fig. 16–4). Schwann cell precursors, which fail to survive without axons, do not possess autocrine survival circuits and appear to be wholly dependent on axonal neuregulin-1 signaling for survival.70,200,360 Despite the reports that Schwann cells can make low levels of neuregulin-1,37,269 the

FIGURE 16–4 Survival regulation in the Schwann cell lineage. Both axon-derived and autocrine signals regulate survival of cells in the Schwann cell lineage. Schwann cell precursors rely exclusively on axonal signals, principally neuregulin-1, for survival, but as development proceeds and precursors generate Schwann cells, there is a shift to the establishment of autocrine loops, the main components of which are likely to include IGF-2, PDGF-BB, NT-3, LIF, and a LPA-like activity. (From Jessen, K. R., and Mirsky, R.: Schwann cell development. In Lazzarini, R. A. [ed.]: Myelin Biology and Disorders, Vol. 1. San Diego, CA, Academic Press, p. 329, 2003, with permission.)

Molecular Signaling in Schwann Cell Development

autocrine Schwann cell survival signal is unlikely to be neuregulin-1 because the autocrine signal is not mitogenic for precursors and does not support their survival.51,200 Major components of the autocrine loop have been identified as IGF-2, combined with neurotrophin-3 (NT-3) and plateletderived growth factor-BB (PDGF-BB).200 These factors support survival when applied together in very low concentrations to Schwann cells that are plated so sparsely that they would die without added survival signals. Schwann cells also express receptors for these factors in vivo and in vitro. Furthermore, using this assay, the survival activity present in Schwann cell–conditioned medium is blocked by antibodies to these proteins. IGF is also a potent Schwann cell survival factor in an alternative survival assay in which Schwann cell death is induced by abrupt serum deprivation. In these experiments, IGF acts via PI3 kinase and Akt to inhibit activation of JNK and prevent caspase-mediated apoptosis.35,49,67,323 Leukemia inhibitory factor (LIF) and lysophosphatidic acid (LPA) are other potential autocrine Schwann cell survival factors. Denervated Schwann cells secrete LIF, which can promote Schwann cell survival in the presence of other growth factors.71 LPA also promotes survival in the serum deprivation assay, and densely plated Schwann cells secrete an LPA-like activity into the medium. LPA signals through the Gi protein–coupled lpA1 receptor, which in turn activates the PI3 kinase pathway and Akt.354,355 Longer term survival in the presence of autocrine signals is enhanced by culture on a laminin substrate, although laminin alone does not support survival.200 As mentioned earlier, the activity of Ets transcription factors is required for Schwann cell survival mediated by neuregulin-1 or by Schwann cell–conditioned medium, whereas it is not needed for LPA-induced survival.249 In summary, the survival of Schwann cell precursors is acutely dependent on axonal neuregulin-1, and they do not possess autocrine survival circuits, whereas Schwann cells can save themselves from death by secretion of several survival factors that are likely to include IGF-2, PDGF-BB, NT-3, LIF, and LPA. An important consequence of this is that, in older nerves, Schwann cells can survive in the absence of axons for prolonged periods. In the perinatal period, axonal neuregulin-1 is likely to contribute additional survival support for Schwann cells in parallel with that provided by autocrine signals. In addition to positive signals that promote survival and proliferation, factors that actively promote apoptosis may also contribute to the regulation of Schwann cell numbers in developing nerves. There is evidence that two such factors, nerve growth factor (NGF) and TGF-␤, can act in this way.248,301 NGF, acting via p75NTR, promotes cell death in several systems, including Schwann cells. However, the effect is complex because, depending on conditions, NGF can promote either Schwann cell death or survival. This underlines the importance of the total cellular and signaling context in the balance between survival and death.145 There

355

is less Schwann cell death in nerves of p75NTR-null mice than in normal mice following neonatal nerve transection or in regenerating adult nerves 3 weeks after nerve crush.89,321 When deprived of serum and growth factors, cultured expanded neonatal Schwann cells from these mice survive better than normal Schwann cells,91,301,321 and antibodies to p75NTR block apoptosis in Schwann cells cultured from transected adult rat nerve distal stumps.123 Studies on a protein that interacts specifically with p75NTR, receptor-interacting protein (RIP)-2, highlight the complexity of the Schwann cell response to p75NTR activation. In the presence of RIP-2, Schwann cell death in response to NGF is prevented through activation of the transcription factor NF-␬B. RIP-2, a protein kinase that contains a caspase recruitment domain (CARD), is expressed by freshly isolated Schwann cells, but not by Schwann cells that have been cultured for a prolonged period. Schwann cells that do not express RIP-2 are sensitive to NGF-induced cell death, and transfection of constitutively active RIP-2 confers protection against death. Conversely, expression of dominant-negative RIP-2 in freshly cultured RIP-2–expressing Schwann cells confers sensitivity to NGF-induced cell death.145 Tumor necrosis factor (TNF) receptor–associated factor (TRAF)-6 is another factor that associates with p75NTR. This activates the JNK pathway in Schwann cells (see below), and may be responsible for the NGF-induced death that occurs in the absence of RIP-2.145,146 Expression of other p75NTR-associated proteins has been correlated with cell death in immortalized cell lines, but their involvement in Schwann cell death has not been demonstrated. Among these proteins are the zinc finger proteins neurotrophin receptor interacting factor (NRIF)-1 and -2, p75NTR-associated cell death executor (NADE), and neurotrophin receptor–interacting tumor suppressor MAGE (NRAGE), which also mediates cell cycle arrest in sympathetic neuroblasts.14,39,144,221,222,275 In addition, two other proteins that do not appear to be involved in cell death have been shown to bind to p75NTR. These are the zinc finger protein Schwann cell factor-1, which promotes growth arrest in Schwann cells, and Rho(A) GTPase, which promotes neurite elongation in PC12 cells, and which has been implicated in actin-based movement in many cell types.54,367 TGF-␤s, like NGF, can induce apoptotic cell death in neonatal Schwann cells both in vitro and in vivo.248 Freshly isolated neonatal Schwann cells cultured in serum-free media die by apoptosis in response to TGF-␤1 application, and increased cell death is seen when TGF-␤ is injected into transected, but not normal, neonatal nerves. In vitro, a combination of neuregulin-1 and the autocrine signal cocktail of IGF-2, PDGF-BB, and NT-3 completely blocks apoptosis. TGF-␤1 signals via JNK and phosphorylation of c-Jun, since adenoviral expression of dominant-negative c-Jun inhibits TGF-␤ induced apoptosis, while retroviral expression of constitutively active Jun (v-Jun) promotes cell death.

356

Neurobiology of the Peripheral Nervous System

Myelinating Schwann cells isolated from postnatal day 4 nerves are resistant to TGF-␤–induced cell death, and are unable to phosphorylate c-Jun in response to TGF-␤.248 In vitro, enforced expression of the myelin-related transcription factor Krox-20 is sufficient to make Schwann cells resistant to TGF-␤–induced death246,247 (see Transcription Factors Involved in Myelination below). In the presence of serum, a combination of TGF-␤ and TNF-␣, but neither alone, induces Schwann cell death. This combination also kills freshly isolated Schwann cells more effectively than TGF-␤ alone in serum-free conditions.245,299 In immortalized Schwann cells, TNF-␣ in combination with interferon-␥ also induces death via production of nitric oxide and ceramide.224 The actions of TGF-␤ are strikingly context dependent. In addition to inducing apoptosis in defined medium, it can promote Schwann cell DNA synthesis in the presence of serum or cAMP-elevating agents or both.75,264 Another potential of TGF-␤ is revealed in neuron–Schwann cell co-cultures, wherein TGF-␤ inhibits the Schwann cell proliferation induced by contact with neurites and blocks myelination.76,110 TGF-␤ also suppresses myelin gene expression in purified Schwann cell cultures.215 As described above, in perinatal nerves, the population of immature Schwann cells is exposed to survival signals (neuregulin-1 and autocrine factors), proliferative signals (principally neuregulin-1), and death signals (NGF and TGF-␤). These signals in combination are likely to provide a mechanism for matching the number of Schwann cells to axons. Two of the signals (TGF-␤ and neuregulin-1) can also suppress myelination (see Transcription Factors Involved in Myelination below) and could also play a role in timing the length of the promyelination phase of Schwann cell development.

CELL CULTURE STUDIES ON SCHWANN CELL PROLIFERATION AND DIFFERENTIATION Although neuronally derived neuregulin-1 (see Importance of Neuregulin-1 in Schwann Cell Development, and Neuregulin-1 Signaling Pathways in Schwann Cells, above) is usually thought to be the major mitogen in Schwann cell development, a variety of other growth factors present in neurons cause Schwann cell proliferation. These factors may therefore help to regulate proliferation in vivo.210 Two in vitro models have been employed to investigate proliferation and myelin-related differentiation in Schwann cells. The first model involves co-cultures of DRG neurons and Schwann cells. In this system, axon-induced proliferation and myelination occur sequentially in defined media with or without serum, and the effects of antibodies, growth factors, genetically engineered constructs, or intracellular pathway inhibitors (see Transcription Factors

Involved in Myelination below) can be studied. In particular, this model has established that neuregulin-1 is the primary axonal mitogen for Schwann cells and has established the importance of the basal lamina in myelination.31,32,78,177,220 In the second system, intracellular cAMP levels are elevated in purified cultures of Schwann cells, grown in media with or without identified growth factors or serum, to mimic the effects of axon-induced proliferation and differentiation. In this model, elevation of cAMP levels in the absence of growth factors (see below) strongly stimulates expression of myelin-related proteins and lipids, including P0, PMP22, periaxin, and galactocerebroside, without inducing DNA synthesis, thus mimicking events that occur when myelination is induced by axons. Alternatively, when cAMP levels are elevated in the presence of growth factors or serum, cell division is stimulated, and elevation of proteins such as P0 in the population as a whole is much reduced (reviewed by Jessen and Mirsky138,139,210). Results using these two models largely complement one another. Because so many different protocols have been used to study proliferation in purified Schwann cell cultures, it is difficult to strictly compare results from different laboratories, particularly when activation of intracellular signaling pathways is investigated. It is important to note that cells prepared by different methods, including freshly isolated immunopanned cells, serum-purified cells, and cells expanded in growth factors and forskolin for short or long periods, may show different responses to applied reagents. In addition, Schwann cells in co-cultures with neurons may also respond differently from purified Schwann cells cultured alone. Studies on the regulation of DNA synthesis using serumpurified Schwann cell cultures have identified several growth factors, including FGF-1 and -2, PDGF-BB, TGF-␤s, and Reg-2, that induce Schwann cell DNA synthesis in the presence or absence of serum, provided cAMP levels are elevated and type 1 IGF receptors are activated. None of these factors induces significant Schwann cell proliferation when used alone in the absence of cAMP elevation48,181,310 (for review, see Eccleston74 and Mirsky and Jessen210). As mentioned above, neuregulin-1 induces Schwann cell proliferation without the obligatory presence of cAMP-elevating agents, presumably as a result of the ability of neuregulin-1 to stimulate cAMP pathways (see Neuregulin-1 Signaling Pathways in Schwann Cells above). IGFs act in Schwann cells and other cells as progression factors and potentiate the effects of most if not all Schwann cell mitogens, including neuregulin-1.47,48,61,181,310,315 While neuregulin-1 and IGFs both promote cell cycle progression in Schwann cells, IGFs, but not neuregulin-1, also increase cell size in the G1 phase of the cell cycle.60,61 Uniquely, the mitogenic effect of hepatocyte growth factor (HGF), which stimulates Schwann cell proliferation in the presence of low levels of serum, is inhibited by cAMP elevation.154

Molecular Signaling in Schwann Cell Development

At present, no comprehensive picture of all the pathways involved in the cAMP-dependent stimulation of proliferation exists, but three different sets of experiments suggest some of the mechanisms that may be used. Early studies suggested that a major function of cAMP in Schwann cell proliferation was to upregulate synthesis of receptors for Schwann cell mitogens such as PDGF.356 Recent studies have confirmed this observation, but indicate that other mechanisms are involved at earlier stages in the response.148,149 Using the interaction between cAMP and PDGF-BB as a model, they show that PDGF acts upstream of cAMP as a competence factor in the G1 phase of the cell cycle. The primary effect of cAMP is to induce sustained elevation of the G1 phase–specific protein cyclin D1, which is induced only transiently by PDGF alone. Ectopic expression of cyclin D1 alleviates the G1 phase requirement for cAMP, while the tumor suppressor gene Nf1 antagonizes accumulation of cAMP and expression of cyclin D1. cAMP is likely to act in similar ways in concert with other Schwann cell mitogens. In cyclin D1– or cyclin D2–null mice, Schwann cell proliferation proceeds normally in the early postnatal period. In contrast, little proliferation is seen in serumpurified cultured cyclin D1–null Schwann cells in response to either PDGF plus forskolin or neuregulin-1 or after nerve transection in cyclin D1–null mice. Thus cyclin D1 is required for proliferation in the absence of axons, either in neonatal Schwann cell cultures or in transected adult nerves, whereas it is not required for axonally driven proliferation during development.8,148 Experiments on Schwann cells and thyroid follicular cells have implicated the p70 ribosomal S6 kinase (p70s6k) in cAMP-dependent proliferation, because cAMP-dependent proliferation in both cell types is inhibited by rapamycin, a specific blocker of this kinase.40 Further investigation of this mechanism in thyroid follicular cells suggests that cAMP-dependent proliferation is dependent on the convergence of two separate cAMP-activated pathways on p70s6k. One pathway involves PKA-dependent activation of p70s6k while the other involves PI3 kinase–dependent, PKA-independent, activation of p70s6k.41,42 Interestingly, in Schwann cell–neuron co-cultures, where proliferation is driven by contact with neurites, retroviral inhibition of cAMP-dependent PKA does not inhibit mitosis, whereas it does so in purified Schwann cells exposed to cAMP-elevating agents and mitogens. Neurite-induced proliferation is nevertheless inhibited by PKA inhibitors such as H89 and KT5780, which are presumably less specific than the retrovirally mediated vectors. This suggests that, in development, neurite-induced proliferation may involve pathways other than those activated by PKA127,147 (see below). The cyclin-dependent kinases (Cdks) 2, 4, and 6 are involved in cell cycle progression, and levels are high in Schwann cells cultured under proliferating conditions.195

357

Cdk2 levels, which are controlled transcriptionally, are low in growth-arrested cells in vitro, high in early postnatal nerves, and low in adult nerves.330 Levels of p27, a cell cycle–inhibitory protein, are higher in arrested Schwann cells than in proliferating cells,330 and p27 is elevated by enforced expression of the transcription factor Krox20, which also induces growth arrest.245 Schwann cell growth arrest can be induced by constitutive activation of Ras or Raf kinase, which results in activation of the MAP kinase pathway. Arrest occurs in the G1 phase of the cell cycle via the induction of p21Cip1 (but not p27Kip1), with subsequent inhibition of cyclin/Cdk activity.182,265

TRANSCRIPTION FACTORS INVOLVED IN MYELINATION Transcription factors that are expressed by Schwann cells or their precursors are summarized in Table 16–1. Those most relevant to myelination are discussed here. Myelination is the final stage of Schwann cell differentiation. On myelination, Schwann cells radically transform their phenotype in response to signals from the larger axons. This response represents one of the most striking examples of cell-cell interaction that is known, and results in the generation of one of the most highly specialized cell types in the body, the myelinating Schwann cell. This differentiation is accompanied by reciprocal changes in axonal ion channels and cytoskeleton that enable saltatory conduction (Fig. 16–5; see also Fig. 16–3). Although many experiments indicate that myelination and myelin maintenance depend on the presence of axons, and on close contact between axons and Schwann cells, no myelination-inducing signal from axons has been identified, nor have receptors that signal myelin formation been described in Schwann cells. A clear demonstration at the molecular level of cell-cell signaling leading to myelination is therefore lacking. Despite this, several other key steps in the process have been clarified. The importance of the basal lamina in enabling myelination is well established, and the role of extracellular matrix proteins and their receptors in myelination is discussed below (see Extracellular Matrix, Matrix Receptors, and the Cytoskeleton in Schwann Cell Development) and has been reviewed by several authors.32,53,256 Significant progress has also been made in the analysis of transcriptional mechanisms that control myelination. The crucial importance of the transcription factor Sox10 in early development of the glial lineage has already been outlined (see Neural Crest Origins of Schwann Cells above), but this factor also participates in later lineage events (see below). Two other transcription factors, the POU domain factor Oct6 and the zinc finger protein Krox20, have important roles in myelination15,133,175,213,337 (for reviews of these and other transcription factors in the Schwann cell lineage, see Mirsky

358

Neurobiology of the Peripheral Nervous System

Table 16–1. Transcription Factors Associated with Schwann Cell Precursors and/or Schwann Cells* Factor

DNA-Binding Domain

bZIP Factors AP2 CREB C/EBPa E12/47 c-Fos c-Jun Jun D Jun B

bZIP bZIP bZIP bZIP bZIP bZIP bZIP bZIP

Ets Factors ER81 ERM† Ets2 GABP-␣ Net

Ets Ets Ets Ets Ets winged HLH

HLH Factors Hey1 Id1, -2, -3, -4 REB

bHLH bHLH bHLH

POU Homeodomain Factors Brn2 Brn5 Oct1 Oct6 Zinc Finger Factors Krox20 Krox24 (NGF-1A/Egr1) KZF-1 SCp55 ZFP57 Other Factors Elongin A fkd6‡ NF-␬B NT3R Pax3 Sox10 Steroidogenic acute regulatory protein TFIIA large subunit

In Vivo

In Vitro

⫹ ⫹ ⫹ ⫹

⫹ ⫹ ⫹ ⫹

⫹ ⫹ /⫺ ⫹ /⫺

⫹ ⫹ ⫹

312 147, 170, 309, 324 18, 128 314 226 66, 213, 309 309 309

114, 242, 249

⫹ ⫹

⫹ ⫹

POU homeodomain POU homeodomain POU homeodomain POU homeodomain

⫹

⫹

⫹ ⫹

⫹ ⫹

Zn finger Zn finger Zn finger Zn finger Zn finger with KRAB domain

⫹ ⫹

⫹ ⫹

forkhead domain NF-␬B/Rel family Nuclear hormone receptor Paired box domain HMG domain START domain

⫹ ? ⫹ ⫹ ⫹

? ⫹ ? ⫹ ⫹

Four-helix bundle (FHB)

References

225 314, 328 314 132, 226, 298 366 22 15, 120, 133, 211–213, 357 22, 223, 337, 375, 376 158, 194, 226, 236, 335 226 161 72, 238 226 143 38, 233 12 22, 150 27, 157, 303 226 226

bZIP ⫽ basic leucine zipper motif; NF ⫽ nuclear factor; NT3R ⫽ nuclear triiodothyronine (T3) receptor; TF ⫽ transcription factor. *This list is almost certainly not comprehensive. The many gene array studies now underway using Schwann cells are likely to reveal additional important Schwann cell transcription factors in the coming years. † Note that this transcription factor is more highly expressed in neurons and satellite cells than in Schwann cell precursors or Schwann cells.114 It can nevertheless be detected by reverse transcription–polymerase chain reaction in Schwann cells.249 ‡ So far described only in zebra fish Schwann cells and satellite cells.143

and Jessen209 and Topilko and Meijer336). Myelination is severely delayed in Oct6-null mice,15,133 while Schwann cell myelination fails completely in Krox20-null mice.336,337 In both these mice, Schwann cells successfully achieve a 1:1 relationship with the larger axons and wrap

around them up to one and a half times before myelination is arrested temporarily (Oct6 null) or permanently (Krox20 null). In humans, Egr2 (Krox20) mutations are found associated with Charcot-Marie-Tooth disease, Déjérine-Sottas disease, and hereditary sensory and motor

Molecular Signaling in Schwann Cell Development

359

FIGURE 16–5 The main changes in protein and lipid expression that take place as the population of immature Schwann cells diverges to generate myelinating and nonmyelinating cells during development. Myelination involves a combination of downregulation and upregulation of key molecules, most of which are associated with formation of the myelin sheath. Dark gray box: Molecules that are strongly upregulated on myelination. Light gray box: Molecules associated with immature Schwann cells that are downregulated when cells myelinate. The generation of mature nonmyelinating cells from immature cells involves many fewer molecular changes, and most of the molecules typical of immature Schwann cells are also expressed by nonmyelinating cells, as indicated. Galactocerebroside (GalC) is expressed by both myelinating and nonmyelinating Schwann cells, but not by immature cells, while ␣1␤1 integrin is upregulated specifically on nonmyelinating cells as they mature. 04 antigen and S100 are expressed by all Schwann cells in peripheral nerves, although 04 is downregulated in the absence of axons. Hereditary demyelinating neuropathies that result from abnormal expression of some of the myelinassociated proteins are indicated. CMT ⫽ Charcot-Marie-Tooth neuropathy; PMD ⫽ Pelizaeus-Merzbacher disease. (From Jessen, K. R., and Mirsky, R.: The Schwann cell lineage. In Kettenmann, H., and Ransom, B. R. [eds.]: Neuroglia, 2nd ed. Oxford, UK, Oxford University Press, 2004, with permission.)

neuropathies, emphasizing the crucial importance of Krox20 in myelination.13,23,244,331,350,351,368 Schwann cells that have been signaled to myelinate express Krox20 selectively and at high levels (in the mouse from E16 onward).247,375,376 In contrast, Oct6 can be detected at low levels in late precursors/early Schwann cells from about E13 (mouse)/E15 (rat) onward. It is expressed by all Schwann cells in late embryogenesis and the early postnatal period, with highest expression in promyelin cells, declining to very low levels in adult

nerves.7,22 Schwann cells in peripheral nerves of Oct6-null mice do not express Krox20 at the appropriate time but do so after developmental delay of about 2 weeks just before they myelinate. The ability of Oct6-null Schwann cells to myelinate eventually is likely to be due to the activity of a related POU domain transcription factor, Brn2.132 Brn2 and Oct6 are expressed independently of each other but with a similar time course in developing nerves. While myelination is unaffected when Brn2 alone is knocked out, the myelination delay seen in Oct6-null nerves is much

360

Neurobiology of the Peripheral Nervous System

more severe when Brn2 is also missing. Conversely, overexpression of Brn2, achieved by driving Brn2 expression under the control of Oct6 Schwann cell enhancer (SCE) (see below), partially rescues the Oct6-null phenotype. Thus Brn2 can substitute for Oct6 in the initiation of myelination. It is intriguing that, even in the combined absence of Brn2 and Oct6, myelination is not permanently blocked, and that a large number of fibers eventually myelinate.104,132,298,336 Schwann cells in nerves of Krox20-null mice express Oct6 on time, but in the early postnatal period maintain expression at a higher than normal level. Thus Krox20 may participate in downregulating levels of Oct6. It may also regulate cell proliferation and death, because Schwann cells in these mice have elevated DNA synthesis and death rates.104,336,376 On nerve transection, Krox20 levels drop rapidly and both Oct6 and Krox20 are reexpressed in regenerating nerves and in cultured Schwann cells on elevation of cAMP, indicating that they are regulated by axons, like myelin protein genes.223,376 Gene array technology has been used to show that enforced expression of Krox20 in cultured Schwann cells is sufficient to induce a large set of diverse genes, including those involved in myelin protein and lipid synthesis, together with many other genes of unknown function in the Schwann cell.226 These and other studies245–247 demonstrate that enforced expression of Krox20 is sufficient to organize a large number of phenotypic changes in cultured Schwann cells, all of which, in vivo, are linked to the transition from immature proliferating cells to quiescent myelinating cells. Thus, in Schwann cells expressing Krox20, proliferation induced by the major axonal mitogen neuregulin-1 is blocked and cell death induced by TGF-␤ is inhibited. 137,245 The myelin proteins periaxin, myelin-associated glycoprotein, and P0 are activated and L1, a marker of nonmyelinating and immature Schwann cells, is downregulated, as occurs on myelination in vivo.192 Furthermore, Krox20 induces expression of periaxin and P0 and inhibits DNA synthesis and apoptosis in an unrelated fibroblastic cell type, the 3T3 cell,245 behavior that is reminiscent of the action of master regulatory genes such as MyoD, neural bHLH factors, or peroxisome proliferator–activated receptor (PPAR)-␥.65,167,183,272,334,376 (PPARs are involved in the regulation of lipid metabolism.) Examination of the molecular mechanisms used by Krox20 in Schwann cells has revealed an important interaction between Krox20 signaling and the JNK/c-Jun pathway.245 These studies indicate, first, that the JNK/c-Jun activity acts as a brake on myelin differentiation and promotes proliferation and death, and, second, that Krox20 expression is sufficient to inactivate this pathway. This would provide Krox20 with a mechanism for coordinated control of myelination, proliferation, and cell death in developing nerves.

The gene promoters of both the Oct6 and Krox20 genes contain SCE elements, which are controlled by axonally regulated transcription factors. (SCE sequences function to specifically regulate gene expression levels in Schwann cells.) The Oct6 SCE, which resides within a 4.3-kb sequence 12 kb downstream of the transcription initiation site, is sufficient to drive spatially and temporally correct expression of the gene both developmentally and during regeneration.104,190 The Krox20 promoter contains two widely separate elements. One of these, the immature SCE, situated upstream of the transcription start site, is active in immature Schwann cells from E15, presumably in promyelinating cells but not in actively myelinating cells. The other, the myelinating SCE, is active from E18 onward in myelinating cells. It is located within a 1.3-kb sequence positioned 35 kb downstream of the Krox20 open reading frame. It is dependent on Oct6 for activation and contains multiple Oct6 binding sites. In regenerating nerves sequential reexpression of both enhancers is observed.105 In addition to its function in promoting gliogenesis from the neural crest early in development (see Neural Crest Origins of Schwann Cells above), evidence suggests that Sox10 may also regulate later steps in the Schwann cell lineage. Sox10 can function synergistically with Oct6 to promote gene expression of reporter constructs containing recognition sites for both Sox10 and Oct6. Sox10 also interacts positively with Pax3 in studies using reporter constructs with Sox10 and Pax3 binding sites, but shows a negative interaction with Krox20 in comparable experiments.156 It positively regulates a P0 promoter construct in N2A neuroblastoma cells, and it functions synergistically with Krox20 to activate a gap junction connexin 32 (Cx32) promoter construct in N2A neuroblastoma cells and HeLa cells.24,253 Cx32 is expressed by myelinating Schwann cells and mutations in the Krox20/Sox10-binding region of the human Cx32 promoter are among the more than 80 mutations that result in X-linked Charcot-Marie-Tooth neuropathies. Furthermore, mutated forms of Sox10 and Krox20 identified in patients with Charcot-Marie-Tooth disease fail to transactivate the Cx32 promoter.24 Thus Sox10 has important functions in both the early and later stages of peripheral glial development. It should be noted that Sox10 is also required for terminal differentiation of oligodendrocytes, the myelinating cells of the CNS.317 Recently, a role for the transcription factor NF-␬B in enabling myelination has been shown.233 NF-␬B is highly expressed in premyelinating Schwann cells in sciatic nerves, but like Oct6 is downregulated in adult nerves. In myelinating neuron–Schwann cell co-cultures, NF-␬B is activated prior to Oct6 upregulation. Treatment with an inhibitor of NF-␬B or infection with a super-suppressor form of the NF-␬B inhibitory protein I␬B␣ prevents myelination, and Schwann cells fail to ensheath axons. Likewise, when Schwann cells null for the p65 subunit of

Molecular Signaling in Schwann Cell Development

NF-␬B are co-cultured with DRG neurons, myelination is severely reduced and cells fail to elongate properly along axons, making it likely that NF-␬B is required for Schwann cells to enter the promyelinating stage.233 HLH factors regulate both neuronal and muscle differentiation, and roles for HLH factors in regulating Schwann cell myelination have been suggested but not proved. The HLH genes expressed in Schwann cells include the B class HLH factor Mash2, which is expressed in a subset of Krox20-positive myelinating Schwann cells in adult nerves. (Mash 2 is required for development of the placenta.) Mash2 inhibits Schwann cell proliferation in culture and is expressed at lower levels after nerve transection. It regulates the expression of Krox24, the chemokine Mob1, and the chemokine receptor CXCR4 genes in vitro.160 In addition, the four HLH Id genes, which inhibit binding of B class HLH proteins to their A class binding partners such as REB and E12/47, are found in the Schwann cell lineage from the precursor stage onward.314 Id1 and -3 mRNAs are expressed at lower levels when peak myelination is occurring in 10-day nerves than in adult nerves. Both Id1 and -3 strongly repress myelin gene promoter activity, suggesting that the downregulation of myelin gene expression that occurs in adult nerves may involve these factors.328 The zinc finger transcription factor Krox24, which is closely related to Krox20 and binds to some of the same DNA sequences (see Topilko et al.335 and references therein), is expressed in mouse Schwann cell precursors from E11/12 and persists in nonmyelinating cells in adult nerves, being upregulated after nerve cut335 (but see Kury et al.160). In peripheral nerves of Krox24-null mice, myelination and regeneration appear to be normal, but after transection of neonatal nerves there is increased cell death compared with wild-type nerves.119,338

SIGNALING PATHWAYS AND GROWTH FACTORS THAT REGULATE MYELINATION As mentioned above, cell culture studies indicate that cAMP pathways are important in myelination. In neuron–Schwann cell co-cultures, retroviral inhibition of PKA inhibits myelination,127 and in purified Schwann cell cultures elevation of intracellular cAMP, particularly in nonproliferating conditions in the absence of growth factors, stimulates the expression of myelin-related molecules such as Oct6, Krox20, and myelin proteins and lipids. Thus elevation of cAMP partially mimics the axonal signals that initiate myelination in vivo.141,174,203,214,215,223,247,280,300 Furthermore, Schwann cells that express high levels of P0 in response to cAMP elevation also downregulate expression of GFAP, neural cell adhesion molecule, and p75NTR, all of which are downregulated in myelinating cells in vivo214 (see Fig. 16–4).

361

Proliferation is induced in Schwann cell cultures when cAMP-elevating agents and growth factors are present together, as discussed above (see Cell Culture Studies on Schwann Cell Proliferation and Differentiation). DNA synthesis is not induced, however, when cAMP pathways are activated in the absence of growth factors or serum.214 When cAMP-induced P0 expression in the presence (proliferating conditions) and absence (nonproliferating conditions) of growth factors is compared, much higher P0 levels are seen in the absence of growth factors.214 Furthermore, in individual cells, high P0 protein and mRNA levels are not seen in cells undergoing DNA synthesis.214,215,313 Therefore, cAMPinduced myelin-related differentiation in vitro is incompatible with proliferation, as is axonally induced myelination in vivo. Because cAMP elevates Krox20 levels in these cultures, both the upregulation of myelin proteins and cessation of proliferation are likely to be mediated by this transcription factor (see above). As mentioned before, IGFs promote cAMP-dependent Schwann cell proliferation in the presence of other growth factors (see Cell Culture Studies on Schwann Cell Proliferation and Differentiation above). In nonproliferating Schwann cell cultures, IGFs promote cAMP-induced P0 expression in the absence of growth factors310 and myelination in Schwann cell–DRG neuron co-cultures.47,48,271 They also promote the attachment and ensheathment of Schwann cells to axons, perhaps by enhancing Schwann cell motility.48,50 IGFs therefore have the potential to promote Schwann cell motility, proliferation, and differentiation. There is evidence that the PI3 kinase and Akt pathway, which is stimulated by IGFs, may be involved in early events of myelination in neuron–Schwann cell co-cultures, because pharmacologic block of PI3 kinase, but not MAP kinase, reversibly inhibits myelination, but not myelin maintenance, perhaps by interfering with Schwann cell–axon interactions that occur early in myelination.198 The rate of myelination can also be regulated by the neurotrophins brain-derived neurotrophic factor (BDNF) and NT-3. Reduction of BDNF signaling retards myelination during peripheral nerve regeneration and in neuron– Schwann cell co-cultures, while increasing BDNF levels enhances myelination.43,373 NT-3 acts in the opposite way and inhibits myelination.43 Progesterone increases the rate of myelination both in vivo and in vitro, although the evidence suggests that it does not affect the eventual extent of myelination44,151,237 (for reviews see Magnaghi et al.189 and Schumacher et al.285). In addition to availability from blood, both neurons and Schwann cells can synthesize progesterone from cholesterol, and there is evidence that both neurons and Schwann cells possess classic and nonclassic progesterone receptors, with higher expression in neurons45 (for review see Magnaghi et al.189 and Schumacher et al.285). Several studies suggest that exposure to progesterone elevates P0 and PMP22 mRNA in cultured Schwann cells and causes transient elevation of Krox20

362

Neurobiology of the Peripheral Nervous System

mRNA and protein.68,111,189,237,285 These experiments imply a direct action of progesterone on Schwann cells. Conversely, Chan and colleagues, using Schwann cell–DRG neuron co-cultures, provided evidence that progesterone synthesized by Schwann cells acts on progesterone receptors in neurons to activate neuronal target genes, which might in turn regulate Schwann cell myelin genes.45 Another growth factor, glial cell line–derived neurotrophic factor (GDNF), induces Schwann cell proliferation and promotes myelination of small-diameter axons that would normally not myelinate when it is administered exogenously to adult rats.125 It is argued that the direct effect of GDNF is to increase the size and axon diameter of the smaller neurons and that its effects on proliferation and myelination of Schwann cells are indirect. In the absence of axonal signals, growth-arrested Schwann cells do not myelinate, indicating that growth arrest alone is insufficient to trigger myelination. Conversely, myelinrelated differentiation can be inhibited without simultaneously stimulating DNA synthesis, as seen in experiments involving TGF-␤. This factor suppresses cAMP-induced P0, 04 antigen, and galactocerebroside induction in Schwann cell cultures, even at concentrations too low to induce Schwann cell proliferation,203,215,308 and has similar effects in myelinating neuron–Schwann cell cultures, suppressing DNA synthesis, P0 and galactocerebroside induction, and myelination.76,109 Thus TGF-␤s can override the myelin differentiation pathways activated either by axon-associated myelination signals or by elevation of intracellular cAMP, again suggesting that cAMP and axonal myelin signals activate common intracellular signaling mechanisms. Neuregulin-1 also inhibits myelination in DRG neuron– Schwann cell co-cultures and, unlike TGF-␤s, induces demyelination in cultures in which myelination has already taken place.370 In contrast, in transgenic mice, when neuregulin signaling is disrupted by controlled excision of ErbB2 driven by Krox20-Cre, hypomyelination is seen in peripheral nerves, suggesting that in vivo neuregulin signaling promotes rather than inhibits myelination.99,100,370

NEURAL ACTIVITY AND SCHWANN CELL DEVELOPMENT Neural activity may play a part in controlling glial development in peripheral nerves. Recent studies suggest that the phenotype of immature Schwann cells may be regulated by neuron-derived ATP, acting via P2-type purinergic receptors.3,59,187,188,199,306,307 (Purinergic receptors, which signal in response to extracellular adenine and uracil nucleotides, can be divided into the P2X family of ligandgated ion channel receptors and P2Y metabotropic G protein–coupled receptors.) In DRG neuron–Schwann cell co-cultures, electrical stimulation of neurons causes an

increase in Ca2⫹ concentration in the soma of individual neurons that is followed by delayed calcium elevation in Schwann cells associated with neurites of the same cell. This is caused by nonsynaptic ATP release from the neurons acting via P2 receptors.306 In response to ATP, cell proliferation is inhibited at the same time as Schwann cell differentiation, measured by 04 antigen expression and myelination, is delayed.306,307 Primary Schwann cells express purinergic receptors of the P2Y and P2X(7) subtypes,59,187 and P2U receptors are expressed by immortalized cells.17 The effect of ATP on proliferation in the co-culture system is probably caused by depression of adenylate cyclase activity induced by the action of P2Y receptors. As discussed earlier, in normal nerve development inhibition of proliferation is associated with the onset of myelination,30,313 and takes place well after 04 antigen is first expressed.70,207 Further studies will therefore be needed to determine how the effects of ATP signaling seen in these in vitro experiments integrate with other signals to generate the coordinated pattern of Schwann cell maturation, proliferation, and myelination seen in vivo. Related experiments using pure neuronal DRG cultures indicate that complicated patterns of neural activity lower neuronal expression of adhesion molecules such as N-cadherin and L1, which could affect interaction with Schwann cells.129,130 Low-frequency stimulation of neurons in DRG–Schwann cell co-cultures, which reduces axonal but not Schwann cell expression of L1, reduces myelination to one third of normal levels. Decreased axonal L1 also reduces Schwann cell adhesion to neurites in the cultures in a short-term assay.306,307 Some other studies also provide evidence for a role for L1 prior to myelination. Decreased L1 expression increases defasciculation both in vivo and in vitro,126,163,180 and function-blocking L1 antibodies in the neuron–Schwann cell co-culture system prevent Schwann cell alignment along axons, which in turn results in failure to myelinate.287,288,363 Despite this, in L1-null mice no major disturbance in development of either myelinated or unmyelinated fibers has been described, although disturbances in the maintenance of axon–Schwann cell relationships in sensory fibers occur in adult nerves, with subsequent axonal loss.64,117

EXTRACELLULAR MATRIX, MATRIX RECEPTORS, AND THE CYTOSKELETON IN SCHWANN CELL DEVELOPMENT Laminin Signals from the basal lamina and other extracellular matrix molecules are essential for the polarization and differentiation of many non–connective tissue cell types, including Schwann cells. In Schwann cells basal lamina forms just before myelination.31,193 The importance of this event for

Molecular Signaling in Schwann Cell Development

the myelination process was first demonstrated by Richard and Mary Bunge and their colleagues,31 and more recent experiments have built on this earlier work to identify the molecular interactions that occur between the major Schwann cell extracellular matrix molecules, their membrane receptors, and the intracellular signaling systems to which they are linked. The most important component of the basal lamina in this respect is laminin. Laminin exists in 11 different isoforms, each of which consists of different combinations of three subunits, designated ␣ (1 through 5), ␤ (1 through 3), and ␥ (1 and 2).256 The major isoform in the Schwann cell basal lamina is laminin 2 (merosin), consisting of ␣2, ␤1, and ␥1 chains, although smaller amounts of laminins 4 and 1 and other laminins are also present.131,176,227 In the dy/dy mouse there are mutations in the ␣2 chain, which is a constituent of laminins 2 and 4. These lead to defective laminin polymer formation and incomplete formation of the basal lamina. In addition to skeletal muscle atrophy, this mouse shows complex pathology of peripheral motor nerves, as reviewed by several authors.25,58,197,227 The most severe Schwann cell abnormalities are seen in the dorsal and ventral spinal roots, where undifferentiated Schwann cells surround but fail to penetrate naked axon bundles. More general defects seen not only in the roots but also within peripheral nerve bundles include a patchy basal lamina surrounding individual nerve fibers, hypomyelination, excessively wide nodes of Ranvier, and overlapping Schwann cells.2,343 The Schwann cell plasma membrane contains three major receptors for laminin 2. These are integrins ␣6 /␤1 and ␣6 /␤4, and ␣-dystroglycan, a member of the dystrophin complex of proteins.79,80,116,193,196 (Integrins are a large family of plasma membrane receptor molecules. They bind extracellular matrix components externally and cytoskeletal molecules internally, and thus link the extracellular matrix with the cytoskeleton. They consist of two subunits, ␣ and ␤.)

Integrin Signaling in Schwann Cells and Interactions with the Actin Cytoskeleton Schwann cells in vivo or in culture express a variety of integrins, including ␣1/␤1, ␣2 /␤1, ␣5 /␤1, ␣6 /␤1, ␣6 /␤4, an unidentified integrin (␣150 /␤1), ␤8 integrin, ␣v /␤3, and ␣v /␤1, plus very low levels of ␣4 and ␣5.76,87,206,314 The combinations of integrin receptors that are expressed by axons, Schwann cell precursors, and Schwann cells vary with the stage of peripheral nerve development and culture conditions (reviewed by Previtali et al.255). Conditional disruption of ␤1 integrin has demonstrated the importance of ␤1 integrin signaling in late embryonic and subsequent phases of peripheral nerve development in Schwann cells.83 Mice carrying a P0Cre transgene were crossed with mice carrying a floxed allele of ␤1 integrin,

363

resulting in excision of ␤1 integrin specifically from Schwann cells in peripheral nerves, starting between E13.5 and E14.5. Thus peripheral nerves can be populated by precursors and generate Schwann cells just before the ␤1 integrin is excised. The null Schwann cells survive and proliferate normally, but their relationships with axons are severely abnormal. Axons fail to segregate properly in many cases and fail to achieve the 1:1 relationship with the Schwann cell required for myelination. Despite this, a few axons are myelinated in adult mice, but even around these axons myelination is developmentally delayed. Aspects of the phenotype seen resemble those observed in the nerve roots of the laminin 2–deficient dy/dy mouse, and on this basis it is likely that the main partner for the ␤1 integrin is laminin 2, acting via the ␣6 /␤1 integrin receptor, which is highly expressed in embryonic Schwann cells.83,255 Absence of ␤1 integrin would then result in improper linkage between the laminin in the basal lamina with the Schwann cell cytoskeleton, a process that is required for axonal ensheathment.83 In addition to interactions between laminin and integrin receptors, there is considerable evidence that interactions between integrins and the cytoskeleton are important in Schwann cells. In myelinating Schwann cell–neuron cocultures, ␣6 /␤1 and ␣1/␤1 integrins are expressed by Schwann cells prior to myelination and antibodies to ␤1 block myelination, whereas antibodies to ␣1/␤1 do not, implying that ␣6 /␤1 is the receptor required for the onset of myelination.87,124 A block of myelination also occurs in response to drug-induced disruption of the actin cytoskeleton.86 Furthermore, in differentiating Schwann cell–neuron cocultures but not in pure Schwann cell cultures, ␤1 integrin co-precipitates in pull-down assays with focal adhesion kinase (FAK) and the actin-linked protein paxillin. (FAK is part of the signaling complex that forms when extracellular matrix components bind to integrins. This leads to clustering of the receptors, recruitment of actin filaments and signaling molecules to the cytoplasmic domain of the receptors, and attachment to the substrate.) As Schwann cells in co-cultures form basal lamina and differentiate, tyrosine phosphorylation of both FAK and paxillin also increases, suggesting that ␤1-mediated intracellular signaling is involved in the differentiation process.46 IGF-1, acting via a PI3 kinase pathway, activates FAK and promotes Schwann cell process extension and motility.50 Whether integrins take part in this process is not known, although it is likely given the co-localization of FAK, paxillin, and ␤1 integrin in filopodia.327 Another important Schwann cell protein, the tumor suppressor protein merlin/Schwannomin, also binds to paxillin, enabling it to anchor to the plasma membrane and associate with ␤1 integrin and ErbB2.88 Merlin is the product of the neurofibromatosis type 2 gene, and is a member of the ezrin/radixin/moesin family of proteins that links membrane proteins to the actin cytoskeleton in epithelia and other cell types.270,341 Mutations in merlin lead to an increased frequency of schwannomas, most sporadic schwannomas

364

Neurobiology of the Peripheral Nervous System

and meningiomas carry double mutations in merlin, and enforced expression of merlin in merlin-null cell lines inhibits cell proliferation, indicating a role for this gene in regulation of Schwann cell proliferation.284 In myelinating Schwann cells, it co-localizes at paranodal membranes with Rho(A), a protein heavily implicated in actin-based movement in other cell types, and interacts directly with the actin-binding protein ␤II spectrin and several other proteins, including HGF-regulated tyrosine kinase substrate, through which it regulates cell proliferation in schwannoma cell lines.278,286,320 Human schwannoma cells deficient in merlin have increased levels of activated Rac.140 (Rac is a Rho-GTPase associated with formation of lamellipodia in fibroblasts and other cell types.) This results in increased levels of activated JNK in the nucleus compared with normal human Schwann cells, although increased levels of phospho–c-Jun were not detected. This accords with earlier studies showing that merlin-deficient cells have increased membrane ruffling.292 While these observations indicate that merlin modifies the activity of Rac, a number of studies also showed that Rac controls the function of merlin.292,295 These effects of merlin loss are the converse of the effect of Krox20 expression in normal Schwann cells, namely downregulation of JNK, and of c-Jun and c-Jun activation (see Transcription Factors Involved in Myelination above). Perhaps related to this, merlin also interacts with the cytoplasmic tail of CD44 (see Importance of Neuregulin-1 in Schwann Cell Development above) to mediate contact inhibition of growth, while loss of merlin destabilizes adherens junctions.162,217,274 The phospholipid signaling molecule LPA (see Signals That Control Schwann Cell Survival above) also causes focal adhesion assembly of paxillin and vinculin, acting via the Rho/p160 Rho-associated kinase (ROCK) pathway; the rearrangement of the actin cytoskeleton into wreath-like structures; and induction of N-cadherin/catenin–mediated cell-cell contacts between adjacent Schwann cells.355 (Vinculin is a component of the FAK that binds actin. Catenins are multifunctional proteins that link cadherins to the actin cytoskeleton. They can also function as transcription factors.) Consistent with its distribution in vivo, N-cadherin mediates axon-aligned process growth of Schwann cells and Schwann cell–Schwann cell contacts in culture.297,349 Another actin-binding protein, dystonin, may also be involved in Schwann cell myelination. Schwann cells from mice with dystonin mutations have a severely disorganized cytoskeleton, fail to attach normally to laminin, and show abnormal myelination.16 Schwann cell migration, which is of importance in embryonic nerves and regeneration, is also likely to involve integrins. Antibodies to ␤1 integrin block Schwann cell migration on laminins 1 and 2 in response to neuregulin and forskolin. Antibodies to ␣6 /␤1 block migration on laminin 1 but not laminin 2, while RGD peptides block

migration on fibronectin, suggesting involvement of ␣v integrins.206 The MAP kinase pathway appears to be an important intracellular pathway used in migration on laminin,201 and integrins ␣3, ␣6, and ␤1 also associate with the tetraspan protein CD9, which has been implicated in Schwann cell adhesion, proliferation, and migration.5,113 In contrast to ␣6 /␤1, which is expressed earlier in development, the second main laminin receptor in Schwann cells, ␣6 /␤4 integrin, is strongly expressed in differentiated Schwann cells, including myelinating Schwann cells, although no peripheral nerve phenotype has been reported in ␤4 knockout mice77,83,84,87,235,344 (see Mirsky and Jessen208 for discussion of this point). Equally, strong ␤8 integrin immunolabeling is seen around myelin sheaths in mouse peripheral nerve, suggesting this protein might participate in signaling during myelination, but this has not been tested.206 Possible roles for another integrin, ␣1/␤1, are suggested by its in vivo distribution. In mature nerves, it is absent from myelinating cells but is found on mature nonmyelinating Schwann cells. It is expressed by all Schwann cells in adult nerve after transection, although it is not strongly expressed by Schwann cell precursors or Schwann cells in developing nerves. It might therefore be involved in the interaction of small-diameter axons with their enveloping Schwann cells, in the interaction of regenerating axons with Schwann cells, or in Schwann cell migration.314 Mice lacking ␣4 or ␣5 integrins provide evidence that these integrins play a role in the early embryonic development of the Schwann cell lineage.112 These experiments implicate ␣4 integrin in the survival of early developing glia, while ␣5 integrin is needed for early glial proliferation.

Dystroglycan Signaling in Schwann Cells The third main Schwann cell laminin receptor, dystroglycan, is involved in the assembly and maintenance of the basal lamina. When laminin binds to dystroglycan, it provides a scaffold for the assembly of a nascent basal lamina, which can then incorporate type IV collagen. In contrast, laminin interaction with ␤1 integrin induces a fibrillar matrix.342 In Schwann cells, ␣-dystroglycan binds to the transmembrane protein ␤-dystroglycan, which, in turn, binds to dystrophinrelated protein (DRP)-116 (an alternatively spliced form of dystrophin that is specific to Schwann cells), which lacks an actin-binding domain, or to utrophin, which has an actinbinding domain.273,293 In myelinating Schwann cells, dystroglycan can also bind to the recently described Schwann cell dystrophin isoform DRP-2. This in turn binds to the Schwann cell cytoskeletal PDZ-domain protein periaxin.293 (Postsynaptic density protein 95, Drosophila discs large tumor suppressor, zonula occludens-1 [PDZ] domains are commonly found in proteins that participate in signaling complexes.) Postnatally, periaxin is clearly restricted to myelinating cells, and in developing nerves, initial activation of the periaxin gene is seen in a subpopulation of early

Molecular Signaling in Schwann Cell Development

immature Schwann cells in embryonic nerves, where it appears to be an early indicator of myelin differentiation.247,281 Periaxin exists in two alternatively spliced forms, both of which contain an N-terminal PDZ domain. The large form of periaxin is associated with the plasma membrane in myelinating Schwann cells, where it forms clustered patched complexes with DRP-2, which are seen in close association with portions of the outermost myelin lamella. These clustered complexes are not formed in periaxin-null mice, although DRP-2 protein levels are normal.73,106,107,281,293 Mutations of periaxin result in severe hypomyelinating sensory neuropathy in mice and Charcot-Marie-Tooth neuropathy type 4F in humans.107,293,325,359 Although periaxin-null mice myelinate normally, it is clear that these complexes are essential to the ongoing stability of mature myelin.107,293 Interaction of Schwann cell–derived laminin 2 with a pathogenic bacterium is seen. In leprosy, dystroglycanlaminin interactions are perturbed. This disease, caused by Mycobacterium leprae, is characterized by infiltration and infection of the Schwann cells of sensory nerves of the skin and elsewhere. The infective agent in leprosy, the phenolic glycolipid-1 of the M. leprae cell wall, interacts specifically with the G domain of the laminin ␣2 chain of laminin 2. This interferes with the normal laminin interaction with ␣-dystroglycan.29,231,260,261,293 A recent study, using Schwann cell–DRG neuron co-cultures, demonstrated that M. leprae induces different responses in myelinating and nonmyelinating Schwann cells.262 It has been suggested that different signaling cascades might be initiated in the two Schwann cell types because myelinating Schwann cells express the DRP-2 dystrophin isoform, in addition to DRP-116 and utrophin expressed by both myelinating and nonmyelinating cells.29 In nonmyelinating cells the bacteria are taken up and sequestered with the Schwann cells, while in myelinating cells binding of the phenolic glycolipid induces rapid demyelination with relatively little uptake of bacteria.262 Whether this is relevant to the pathology of leprosy remains to be determined.

ACKNOWLEDGMENTS The work from our laboratory quoted in this review was supported by the Wellcome Trust. We thank Mrs. D. Bartram for extensive editorial assistance.

REFERENCES 1. Adlkofer, K., and Lai, C.: Role of neuregulins in glial cell development. Glia 29:104, 2000. 2. Aguayo, A. J., and Bray, G. M.: Developmental disorders of myelination in mouse mutants. In Sears, T. A. (ed.): Neuronal-Glial Cell Interrelationships. Berlin, SpringerVerlag, p. 57, 1982.

365

3. Amedee, T., and Despeyroux, S.: ATP activates cationic and anionic conductances in Schwann cells cultured from dorsal root ganglia of the mouse. Proc. R. Soc. Lond. B Biol. Sci. 259:277, 1995. 4. Anderson, D. J.: Genes, lineages and the neural crest: a speculative review. Philos. Trans. R. Soc. Lond. B Biol. Sci. 355:953, 2000. 5. Anton, E. S., Hadjiargyrou, M., Patterson, P. H., and Matthew, W. D.: CD9 plays a role in Schwann cell migration in vitro. J. Neurosci. 15:584, 1995. 6. Archelos, J. J., Roggenbuck, K., Schneider-Schaulies, J., et al.: Production and characterization of monoclonal antibodies to the extracellular domain of P0. J. Neurosci. Res. 35:46, 1993. 7. Arroyo, E. J., Bermingham, J. R. Jr., Rosenfeld, M. G., and Scherer, S. S.: Promyelinating Schwann cells express Tst-1/SCIP/Oct-6. J. Neurosci. 18:7891, 1998. 8. Atanasoski, S., Shumas, S., Dickson, C., et al.: Differential cyclin D1 requirements of proliferating Schwann cells during development and after injury. Mol. Cell. Neurosci. 18:581, 2001. 9. Baechner, D., Liehr, T., Hameister, H., et al.: Widespread expression of the peripheral myelin protein-22 gene (PMP22) in neural and non-neural tissues during murine development. J. Neurosci. Res. 15:733, 1995. 10. Banerjee, S. A., and Patterson, P. H.: Schwann cell CD9 expression is regulated by axons. Mol. Cell. Neurosci. 6:462, 1995. 11. Bao, J., Wolpowitz, D., Role, L. W., and Talmage, D. A.: Back signaling by the Nrg-1 intracellular domain. J. Cell Biol. 161:1133, 2003. 12. Barakat-Walter, I., Duc, C., and Puymirat, J.: Changes in the nuclear 3,5,3'-triiodothyronine receptor expression in the rat dorsal root ganglia and sciatic nerve during development: comparison with regeneration. Eur. J. Neurosci. 5:319, 1993. 13. Bellone, E., Di Maria, E., Soriani, S., et al.: A novel mutation (D305V) in the early growth response 2 gene is associated with severe Charcot-Marie-Tooth type 1 disease. Hum. Mutat. 14:353, 1999. 14. Benzel, I., Barde, Y. A., and Casademunt, E.: Strain-specific complementation between NRIF1 and NRIF2, two zinc finger proteins sharing structural and biochemical properties. Gene 281:19, 2001. 15. Bermingham, J. R., Scherer, S. S., O’Connell, S., et al.: Tst-1/Oct-6/SCIP regulates a unique step in peripheral myelination and is required for normal respiration. Genes Dev. 10:1751, 1996. 16. Bernier, G., De Repentigny, Y., Mathieu, M., et al.: Dystonin is an essential component of the Schwann cell cytoskeleton at the time of myelination. Development 125:2135, 1998. 17. Berti-Mattera, L. N., Wilkins, P. L., Madhun, Z., and Suchovsky, D.: P2-purigenic receptors regulate phospholipase C and adenylate cyclase activities in immortalized Schwann cells. Biochem. J. 314:555, 1996. 18. Bharucha, V. A., Peden, K. W., Subach, B. R., et al.: Characterization of the cis-acting elements of the mouse myelin P2 promoter. J. Neurosci. Res. 36:508, 1993. 19. Bhattacharyya, A., Frank, E., Ratner, N., and Brackenbury, R.: P0 is an early marker of the Schwann cell lineage in chickens. Neuron 7:831, 1991.

366

Neurobiology of the Peripheral Nervous System

20. Bitgood, M. J., and McMahon, A. P.: Hedgehog and Bmp genes are coexpressed at many diverse sites of cell-cell interaction in the mouse embryo. Dev. Biol. 172:126, 1995. 21. Bixby, S., Kruger, G. M., Mosher, J. T., et al.: Cell-intrinsic differences between stem cells from different regions of the peripheral nervous system regulate the generation of neural diversity. Neuron 35:643, 2002. 22. Blanchard, A. D., Sinanan, A., Parmantier, E., et al.: Oct-6 (SCIP/Tst-1) is expressed in Schwann cell precursors, embryonic Schwann cells, and postnatal myelinating Schwann cells: comparison with Oct-1, Krox-20 and Pax-3. J. Neurosci. Res. 46:630, 1996. 23. Boerkoel, C. F., Takashima, H., Bacino, C. A., et al.: EGR2 mutation R359W causes a spectrum of Dejerine-Sottas neuropathy. Neurogenetics 3:153, 2001. 24. Bondurand, N., Girard, M., Pingault, V., et al.: Human connexin 32, a gap junction protein altered in the X-linked form of Charcot-Marie-Tooth disease, is directly regulated by the transcription factor SOX10. Hum. Mol. Genet. 10:2783, 2001. 25. Bönnemann, C. G., McNally, E. M., and Kunkel, L. M.: Beyond dystrophin: current progress in the muscular dystrophies. Curr. Opin. Pediatr. 8:569, 1996. 26. Brennan, A., Dean, C. H., Zhang, A. L., et al.: Endothelins control the timing of Schwann cell generation in vitro and in vivo. Dev. Biol. 227:545, 2000. 27. Britsch, S., Goerich, D. E., Riethmacher, D., et al.: The transcription factor Sox10 is a key regulator of peripheral glial development. Genes Dev. 15:66, 2001. 28. Britsch, S., Li, L., Kirchhoff, S., et al.: The ErbB2 and ErbB3 receptors and their ligand, neuregulin-1, are essential for development of the sympathetic nervous system. Genes Dev. 12:1825, 1998. 29. Brophy, P. J.: Microbiology: subversion of Schwann cells and the leper’s bell. Science 296:862, 2002. 30. Brown, M. J., and Asbury, A. K.: Schwann cell proliferation in the postnatal mouse: timing and topography. Exp. Neurol. 74:170, 1981. 31. Bunge, R. P.: Expanding roles for the Schwann cell: ensheathment, myelination, trophism and regeneration. Curr. Opin. Neurobiol. 3:805, 1993. 32. Bunge, R. P., Bunge, M. B., and Eldridge, C. F.: Linkage between axonal ensheathment and basal lamina production by Schwann cells. Annu. Rev. Neurosci. 9:305, 1986. 33. Buonanno, A., and Fischbach, G. D.: Neuregulin and ErbB receptor signaling pathways in the nervous system. Curr. Opin. Neurobiol. 11:287, 2001. 34. Burden, S., and Yarden, Y.: Neuregulins and their receptors: a versatile signaling module in organogenesis and oncogenesis. Neuron 18:847, 1997. 35. Campana, W. M., Darin, S. J., and O’Brien, J. S.: Phosphatidylinositol 3-kinase and Akt protein kinase mediate IGF-I- and prosaptide-induced survival in Schwann cells. J. Neurosci. Res. 57:332, 1999. 36. Carraway, K. L. 3rd, Weber, J. L., Unger, M. J., et al.: Neuregulin-2, a new ligand of ErbB3/ErbB4-receptor tyrosine kinases. Nature 387:512, 1997. 37. Carroll, S. L., Miller, M. L., Frohnert, P. W., et al.: Expression of neuregulins and their putative receptors,

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

ErbB2 and ErbB3, is induced during Wallerian degeneration. J. Neurosci. 17:1642, 1997. Carter, B. D., Kaltschmidt, C., Kaltschmidt, B., et al.: Selective activation of NF-kappa B by nerve growth factor through the neurotrophin receptor p75. Science 272:542, 1996. Casademunt, E., Carter, B. D., Benzel, I., et al.: The zinc finger protein NRIF interacts with the neurotrophin receptor p75(NTR) and participates in programmed cell death. EMBO J. 18:6050, 1999. Cass, L. A., and Meinkoth, J. L.: Differential effects of cyclic adenosine 3⬘,5⬘-monophosphate on p70 ribosomal S6 kinase. Endocrinology 139:1991, 1998. Cass, L. A., and Meinkoth, J. L.: Ras signaling through PI3K confers hormone-independent proliferation that is compatible with differentiation. Oncogene 19:924, 2000. Cass, L. A., Summers, S. A., Prendergast, G. V., et al.: Protein kinase A-dependent and -independent signaling pathways contribute to cyclic AMP-stimulated proliferation. Mol. Cell. Biol. 19:5882, 1999. Chan, J. R., Cosgaya, J. M., Wu, Y. J., and Shooter, E. M.: Neurotrophins are key mediators of the myelination program in the peripheral nervous system. Proc. Natl. Acad. Sci. U.S.A. 98:14661, 2001. Chan, J. R., Phillips, L. J. 2nd, and Glaser, M.: Glucocorticoids and progestins signal the initiation and enhance the rate of myelin formation. Proc. Natl. Acad. Sci. U.S.A. 95:10459, 1998. Chan, J. R., Rodriguez-Waitkus, P. M., Ng, B. K., et al.: Progesterone synthesized by Schwann cells during myelin formation regulates neuronal gene expression. Mol. Biol. Cell 11:2283, 2000. Chen, L. M., Bailey, D., and Fernandez-Vallé, C.: Association of beta 1 integrin with focal adhesion kinase and paxillin in differentiating Schwann cells. J. Neurosci. 20:3776, 2000. Cheng, H. L., and Feldman, E. L.: Insulin-like growth factor-I (IGF-I) and IGF binding protein-5 in Schwann cell differentiation. J. Cell. Physiol. 171:161, 1997. Cheng, H. L., Russell, J. W., and Feldman, E. L.: IGF-I promotes peripheral nervous system myelination. Ann. N. Y. Acad. Sci. 883:124, 1999. Cheng, H. L., Steinway, M., Delaney, C. L., et al.: IGF-I promotes Schwann cell motility and survival via activation of Akt. Mol. Cell. Endocrinol. 170:211, 2000. Cheng, H. L., Steinway, M. L., Russell, J. W., and Feldman, E. L.: GTPases and phosphatidylinositol 3-kinase are critical for insulin-like growth factor-I-mediated Schwann cell motility. J. Biol. Chem. 275:27197, 2000. Cheng, L., Esch, F. S., Marchionni, M. A., and Mudge, A. W.: Control of Schwann cell survival and proliferation: autocrine factors and neuregulins. Mol. Cell. Neurosci. 12:141, 1998. Cheng, L., and Mudge, A. W.: Cultured Schwann cells constitutively express the myelin protein P0. Neuron 16:309, 1996. Chernousov, M. A., and Carey, D. J.: Schwann cell extracellular matrix molecules and their receptors. Histol. Histopathol. 15:593, 2000. Chittka, A., and Chao, M. V.: Identification of a zinc finger protein whose subcellular distribution is regulated by serum

Molecular Signaling in Schwann Cell Development

55. 56.

57.

58.

59.

60.

61. 62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

and nerve growth factor. Proc. Natl. Acad. Sci. U.S.A. 96:10705, 1999. Chuah, M. I., and West, A. K.: Cellular and molecular biology of ensheathing cells. Microsc. Res. Tech. 58:216, 2002. Ciment, G.: The melanocyte Schwann cell progenitor: a bipotent intermediate in the neural crest lineage. Comments Dev. Neurobiol. 1:207, 1990. Ciutat, D., Calderó, J., Oppenheim, R. W., and Esquerda, J. E.: Schwann cell apoptosis during normal development and after axonal degeneration induced by neurotoxins in the chick embryo. J. Neurosci. 16:3979, 1996. Colognato, H., and Yurchenco, P. D.: The laminin alpha2 expressed by dystrophic dy(2J) mice is defective in its ability to form polymers. Curr. Biol. 9:1327, 1999. Colomar, A., and Amedee, T.: ATP stimulation of P2X(7) receptors activates three different ionic conductances on cultured mouse Schwann cells. Eur. J. Neurosci. 14:927, 2001. Conlon, I., and Raff, M.: Differences in the way a mammalian cell and yeast cells coordinate cell growth and cell-cycle progression. J. Biol. 2:7, 2003. Conlon, I. J., Dunn, G. A., Mudge, A. W., and Raff, M. C.: Extracellular control of cell size. Nat. Cell Biol. 3:918, 2001. Cowell, L. A., and Weston, J. A.: An analysis of melanogenesis in cultured chick embryo spinal ganglia. Dev. Biol. 22:670, 1970. Curtis, R., Stewart, H. J., Hall, S. M., et al.: GAP-43 is expressed by nonmyelin-forming Schwann cells of the peripheral nervous system. J. Cell Biol. 116:1455, 1992. Dahme, M., Bartsch, U., Martini, R., et al.: Disruption of the mouse L1 gene leads to malformations of the nervous system. Nat. Genet. 17:346, 1997. Davis, R. L., Weintraub, H., and Lassar, A. B.: Expression of a single transfected cDNA converts fibroblasts to myoblasts. Cell 51:987, 1987. De Felipe, C., and Hunt, S. P.: The differential control of c-Jun expression in regenerating sensory neurons and their associated glial cells. J. Neurosci. 14:2911, 1994. Delaney, C. L., Cheng, H. L., and Feldman, E. L.: Insulin-like growth factor-I prevents caspase-mediated apoptosis in Schwann cells. J. Neurobiol. 41:540, 1999. Desarnaud, F., Do Thi, A. N., Brown, A. M., et al.: Progesterone stimulates the activity of the promoters of peripheral myelin protein-22 and protein zero genes in Schwann cells. J. Neurochem. 71:1765, 1998. Dong, Z., Brennan, A., Liu, N., et al.: NDF is a neuron-glia signal and regulates survival, proliferation, and maturation of rat Schwann cell precursors. Neuron 15:585, 1995. Dong, Z., Sinanan, A., Parkinson, D., et al.: Schwann cell development in embryonic mouse nerves. J. Neurosci. Res. 56:334, 1999. Dowsing, B. J., Morrison, W. A., Nicola, N. A., et al.: Leukemia inhibitory factor is an autocrine survival factor for Schwann cells. J. Neurochem. 73:96, 1999. Durán Alonso, M. B., Zoidl, G., Taveggia, C., et al.: Identification and characterization of ZFP-57, a novel zinc finger transcription factor in the mammalian peripheral nervous system. J. Biol. Chem. 2004 in press. Dytrych, L., Sherman, D. L., Gillespie, C. S., and Brophy, P. J.: Two PDZ domain proteins encoded by the murine periaxin

74.

75.

76.

77.

78.

79.

80.

81.

82. 83.

84.

85.

86.

87.

88.

89.

90. 91.

367

gene are the result of alternative intron retention and are differentially targeted in Schwann cells. J. Biol. Chem. 273:5794, 1998. Eccleston, P. A.: Regulation of Schwann cell proliferation: mechanisms involved in peripheral nerve development. Exp. Cell Res. 199:1, 1992. Eccleston, P. A., Jessen, K. R., and Mirsky, R.: Transforming growth factor-beta and gamma-interferon have dual effects on growth of peripheral glia. J. Neurosci. Res. 24:524, 1989. Einheber, S., Hannocks, M.-J., Metz, C. N., et al.: Transforming growth factor-beta 1 regulates axon-Schwann cell interactions. J. Cell Biol. 129:443, 1995. Einheber, S., Milner, T., Giancotti, F., and Salzer, J.: Axonal regulation of Schwann cell integrin expression suggests a role for alpha6 beta4 in myelination. J. Cell Biol. 123:1223, 1993. Eldridge, C. F., Bunge, M. B., and Bunge, R. P.: Differentiation of axon-related Schwann cells in vitro: II. Control of myelin formation by basal lamina. J. Neurosci. 9:625, 1989. Engvall, E., Earwicker, D., Day, A., et al.: Merosin promotes cell attachment and neurite outgrowth and is a component of the neurite-promoting factor of RN22 schwannoma cells. Exp. Cell Res. 198:115, 1992. Ervasti, J. M., and Campbell, K. P.: A role for the dystrophinglycoprotein complex as a transmembrane linker between laminin and actin. J. Cell Biol. 122:809, 1993. Falls, D. L., Rosen, K. M., Corfas, G., et al.: ARIA, a protein that stimulates acetylcholine receptor synthesis, is a member of the neu ligand family. Cell 72:801, 1993. Fawcett, J. W., and Keynes, R. J.: Peripheral nerve regeneration. Annu. Rev. Neurosci. 13:43, 1990. Feltri, M. L., Graus Porta, D., Previtali, S. C., et al.: Conditional disruption of beta 1 integrin in Schwann cells impedes interactions with axons. J. Cell Biol. 156:199, 2002. Feltri, M. L., Scherer, S. S., Nemni, R., et al.: ␤4 integrin expression in myelinating Schwann cells is polarized, developmentally regulated and axonally dependent. Development 120:1287, 1994. Feng, L., Hatten, M. E., and Heintz, N.: Brain lipid-binding protein (BLBP): a novel signaling system in the developing mammalian CNS. Neuron 12:895, 1994. Fernandez-Vallé, C., Gorman, D., Gomez, A. M., and Bunge, M. B.: Actin plays a role in both changes in cell shape and gene-expression associated with Schwann cell myelination. J. Neurosci. 17:241, 1997. Fernandez-Vallé, C., Gwynn, L., Wood, P. M., et al.: Anti-␤1 integrin antibody inhibits Schwann cell myelination. J. Neurobiol. 25:1207, 1994. Fernandez-Vallé, C., Tang, Y., Ricard, J., et al.: Paxillin binds schwannomin and regulates its density-dependent localization and effect on cell morphology. Nat. Genet. 31:354, 2002. Ferri, C. C., and Bisby, M. A.: Improved survival of injured sciatic nerve Schwann cells in mice lacking the p75 receptor. Neurosci. Lett. 272:191, 1999. Fischbach, G. D., and Rosen, K. M.: ARIA: a neuromuscular junction neuregulin. Annu. Rev. Neurosci. 20:429, 1997. Frade, J. M., and Barde, Y. A.: Genetic evidence for cell death mediated by nerve growth factor and the neurotrophin receptor p75 in the developing mouse retina and spinal cord. Development 126:683, 1999.

368

Neurobiology of the Peripheral Nervous System

92. Franklin, R. J., and Barnett, S. C.: Olfactory ensheathing cells and CNS regeneration: the sweet smell of success? Neuron 28:15, 2000. 93. French, S. W., Hoyer, K. K., Shen, R. R., and Teitell, M. A.: Transdifferentiation and nuclear reprogramming in hematopoietic development and neoplasia. Immunol. Rev. 187:22, 2002. 94. Fromm, L., and Burden, S. J.: Neuregulin-1-stimulated phosphorylation of GABP in skeletal muscle cells. Biochemistry 40:5306, 2001. 95. Furukawa, T., Mukherjee, S., Bao, Z. Z., et al.: Rax, Hes1, and Notch1 promote the formation of Muller glia by postnatal retinal progenitor cells. Neuron 26:383, 2000. 96. Gabella, G.: Ultrastructure of the nerve plexuses of the mammalian intestine: the enteric glial cells. Neuroscience 6:425, 1981. 97. Gaiano, N., Nye, J. S., and Fishell, G.: Radial glial identity is promoted by Notch1 signaling in the murine forebrain. Neuron 26:395, 2000. 98. Garbern, J. Y., Cambi, F., Tang, X. M., et al.: Proteolipid protein is necessary in peripheral as well as central myelin. Neuron 19:205, 1997. 99. Garratt, A. N., Britsch, S., and Birchmeier, C.: Neuregulin, a factor with many functions in the life of a Schwann cell. Bioessays 22:987, 2000. 100. Garratt, A. N., Voiculescu, O., Topilko, P., et al.: A dual role of erbB2 in myelination and in expansion of the Schwann cell precursor pool. J. Cell Biol. 148:1035, 2000. 101. Gassman, M., and Lemke, G.: Neuregulins and neuregulin receptors in neural development. Curr. Opin. Neurobiol. 7:87, 1997. 102. Ge, W., Martinowich, K., Wu, X., et al.: Notch signaling promotes astrogliogenesis via direct CSL-mediated glial gene activation. J. Neurosci. Res. 69:848, 2002. 103. Gershon, M. D.: V. Genes, lineages, and tissue interactions in the development of the enteric nervous system. Am. J. Physiol. 275:G869, 1998. 104. Ghazvini, M., Mandemakers, W., Jaegle, M., et al.: A cell type-specific allele of the POU gene Oct-6 reveals Schwann cell autonomous function in nerve development and regeneration. EMBO J. 21:4612, 2002. 105. Ghislain, J., Desmarquet-Trin-Dinh, C., Jaegle, M., et al.: Characterisation of cis-acting sequences reveals a biphasic, axon-dependent regulation of Krox20 during Schwann cell development. Development 129:155, 2002. 106. Gillespie, C. S., Sherman, D. L., Blair, G. E., and Brophy, P. J.: Periaxin, a novel protein of myelinating Schwann cells with a possible role in axonal ensheathment. Neuron 12:497, 1994. 107. Gillespie, C. S., Sherman, D. L., Fleetwood-Walker, S. M., et al.: Peripheral demyelination and neuropathic pain behavior in periaxin-deficient mice. Neuron 26:523, 2000. 108. Grinspan, J. B., Marchionni, M. A., Reeves, M., et al.: Axonal interactions regulate Schwann cell apoptosis in developing peripheral nerve: neuregulin receptors and the role of neuregulins. J. Neurosci. 16:6107, 1996. 109. Guenard, V., Gwynn, L. A., and Wood, P. M.: Transforming growth factor-beta blocks myelination but not ensheathment of axons by Schwann cells in vitro. J. Neurosci. 15:419, 1995.

110. Guenard, V., Rosenbaum, T., Gwynn, L. A., et al.: Effect of transforming growth factor-beta 1 and -beta 2 on Schwann cell proliferation on neurites. Glia 13:309, 1995. 111. Guennoun, R., Benmessahel, Y., Delespierre, B., et al.: Progesterone stimulates Krox-20 gene expression in Schwann cells. Brain Res. Mol. Brain Res. 90:75, 2001. 112. Haack, H., and Hynes, R. O.: Integrin receptors are required for cell survival and proliferation during development of the peripheral glial lineage. Dev. Biol. 233:38, 2001. 113. Hadjiargyrou, M., Kaprielian, Z., Kato, N., and Patterson, P. H.: Association of the tetraspan protein CD9 with integrins on the surface of S-16 Schwann cells. J. Neurochem. 67:2505, 1996. 114. Hagedorn, L., Paratore, C., Brugnoli, G., et al.: The Ets domain transcription factor Erm distinguishes rat satellite glia from Schwann cells and is regulated in satellite cells by neuregulin signaling. Dev. Biol. 219:44, 2000. 115. Hagedorn, L., Suter, U., and Sommer, L.: P0 and PMP22 mark a multipotent neural crest-derived cell type that displays community effects in response to TGF-beta family factors. Development 126:3781, 1999. 116. Hall, D. E., Reichardt, L. F., Crowley, E., et al.: The ␣1␤1 and ␣6␤1 integrin heterodimers mediate cell attachment to distinct sites on laminin. J. Cell Biol. 110:2175, 1990. 117. Haney, C. A., Sahenk, Z., Li, C., et al.: Heterophilic binding of L1 on unmyelinated sensory axons mediates Schwann cell adhesion and is required for axonal survival. J. Cell Biol. 146:1173, 1999. 118. Harari, D., Tzahar, E., Romano, J., et al.: Neuregulin-4: a novel growth factor that acts through the ErbB-4 receptor tyrosine kinase. Oncogene 18:2681, 1999. 119. Harris, B.: Zinc-finger transcription factors in the Schwann cell lineage. PhD thesis, University of London, 2001. 120. He, X., Gerrero, R., Simmons, D. M., et al.: Tst-1, a member of the POU domain gene family, binds the promoter of the gene encoding the cell surface adhesion molecule P0. Mol. Cell. Biol. 11:1739, 1991. 121. Henion, P. D., and Weston, J. A.: Timing and pattern of cell fate restrictions in the neural crest lineage. Development 124:4351, 1997. 122. Hippenmeyer, S., Shneider, N. A., Birchmeier, C., et al.: A role for neuregulin1 signaling in muscle spindle differentiation. Neuron 36:1035, 2002. 123. Hirata, H., Hibasami, H., Yoshida, T., et al.: Nerve growth factor signaling of p75 induces differentiation and ceramide-mediated apoptosis in Schwann cells cultured from degenerating nerves. Glia 36:245, 2001. 124. Hogervorst, F., Admiraal, L. G., Niessen, C., et al.: Biochemical characterization and tissue distribution of the A and B variants of the integrin alpha 6 subunit. J. Cell Biol. 121:179, 1993. 125. Hoke, A., Ho, T., Crawford, T. O., et al.: Glial cell line-derived neurotrophic factor alters axon Schwann cell units and promotes myelination in unmyelinated nerve fibers. J. Neurosci. 23:561, 2003. 126. Honig, M. G., Camilli, S. J., and Xue, Q. S.: Effects of L1 blockade on sensory axon outgrowth and pathfinding in the chick hindlimb. Dev. Biol. 243:137, 2002. 127. Howe, D. G., and McCarthy, K. D.: Retroviral inhibition of cAMP-dependent protein kinase inhibits myelination but

Molecular Signaling in Schwann Cell Development

128.

129.

130.

131.

132.

133.

134.

135.

136.

137. 138.

139.

140.

141.

142.

143.

144.

not Schwann cell mitosis stimulated by interaction with neurons. J. Neurosci. 20:3513, 2000. Huan, Y., Rutkowski, J. L., and Tennekoon, G. I.: The role of CCAAT/enhancer binding protein in Schwann cell differentiation (No. 2105). Presented at the 34th Annual Meeting of the American Society for Cell Biology, San Francisco, 1994. Itoh, K., Ozaki, M., Stevens, B., and Fields, R. D.: Activitydependent regulation of N-cadherin in DRG neurons: differential regulation of N-cadherin, NCAM, and L1 by distinct patterns of action potentials. J. Neurobiol. 33:735, 1997. Itoh, K., Stevens, B., Schachner, M., and Fields, R. D.: Regulated expression of the neural cell adhesion molecule L1 by specific patterns of neural impulses. Science 270:1369, 1995. Jaakkola, S., Savunen, O., Halme, T., et al.: Basement membranes during development of human nerve: Schwann cells and perineurial cells display marked changes in their expression profiles for laminin subunits and beta 1 and beta 4 integrins. J. Neurocytol. 22:215, 1993. Jaegle, M., Ghazvini, M., Mandemakers, W., et al.: The POU proteins Brn-2 and Oct-6 share important functions in Schwann cell development. Genes Dev. 17:1380, 2003. Jaegle, M., Mandemakers, W., Broos, L., et al.: The POU factor Oct-6 and Schwann cell differentiation. Science 273:507, 1996. Jessen, K. R., Brennan, A., Morgan, L., et al.: The Schwann cell precursor and its fate: a study of cell death and differentiation during gliogenesis in rat embryonic nerves. Neuron 12:509, 1994. Jessen, K. R., and Mirsky, R.: Astrocyte-like glia in the peripheral nervous system: an immunohistochemical study of enteric glia. J. Neurosci. 3:2206, 1983. Jessen, K. R., and Mirsky, R.: Schwann cells and their precursors emerge as major regulators of nerve development. Trends Neurosci. 22:402, 1999. Jessen, K. R., and Mirsky, R.: Signals that determine Schwann cell identity. J. Anat. 200:367, 2002. Jessen, K. R., and Mirsky, R.: Schwann cell development. In Lazzarini, R. A. (ed.): Myelin Biology and Disorders, Vol. 1. San Diego, CA, Academic Press, p. 329, 2003. Jessen, K. R., and Mirsky, R.: The Schwann cell lineage. In Kettenmann, H., and Ransom, B. R. (eds.): Neuroglia, 2nd ed. Oxford University Press, 2004. Kaempchen, K., Mielke, K., Utermark, T., et al.: Upregulation of the Rac1/JNK signaling pathway in primary human schwannoma cells. Hum. Mol. Genet. 12:1211, 2003. Kamholz, J., Sessa, M., Scherer, S., et al.: Structure and expression of proteolipid protein in the peripheral nervous system. J. Neurosci. Res. 31:231, 1992. Kaprielian, Z., Cho, K. O., Hadjiargyrou, M., and Patterson, P. H.: CD9, a major platelet cell surface glycoprotein, is a ROCA antigen and is expressed in the nervous system. J. Neurosci. 15:562, 1995. Kelsh, R. N., Dutton, K., Medlin, J., and Eisen, J. S.: Expression of zebrafish fkd6 in neural crest-derived glia. Mech. Dev. 93:161, 2000. Kendall, S., Goldhawk, D., Kubu, C., et al.: Expression analysis of a novel p75(NTR) signaling protein, which regulates cell cycle progression and apoptosis. Mech. Dev. 117:187, 2002.

369

145. Khursigara, G., Bertin, J., Yano, H., et al.: A prosurvival function for the p75 receptor death domain mediated via the caspase recruitment domain receptor-interacting protein 2. J. Neurosci. 21:5854, 2001. 146. Khursigara, G., Orlinick, J. R., and Chao, M. V.: Association of the p75 neurotrophin receptor with TRAF6. J. Biol. Chem. 274:2597, 1999. 147. Kim, H. A., DeClue, J. E., and Ratner, N.: cAMP-dependent protein kinase A is required for Schwann cell growth: interactions between the cAMP and neuregulin/tyrosine kinase pathways. J. Neurosci. Res. 49:236, 1997. 148. Kim, H. A., Pomeroy, S. L., Whoriskey, W., et al.: A developmentally regulated switch directs regenerative growth of Schwann cells through cyclin D1. Neuron 26:405, 2000. 149. Kim, H. A., Ratner, N., Roberts, T. M., and Stiles, C. D.: Schwann cell proliferative responses to cAMP and Nf1 are mediated by cyclin D1. J. Neurosci. 21:1110, 2001. 150. Kioussi, C., Gross, M. K., and Gruss, P.: Pax3: a paired domain gene as a regulator in PNS myelination. Neuron 15:553, 1995. 151. Koenig, H. L., Schumacher, M., Ferzaz, B., et al.: Progesterone synthesis and myelin formation by Schwann cells. Science 268:1500, 1995. 152. Kondo, T., and Raff, M.: The Id4 HLH protein and the timing of oligodendrocyte differentiation. EMBO J. 19:1998, 2000. 153. Kondo, T., and Raff, M.: Oligodendrocyte precursor cells reprogrammed to become multipotential CNS stem cells. Science 289:1754, 2000. 154. Krasnoselsky, A., Massay, M. J., DeFrances, M. C., et al.: Hepatocyte growth factor is a mitogen for Schwann cells and is present in neurofibromas. J. Neurosci. 14:7284, 1994. 155. Kubu, C. J., Orimoto, K., Morrison, S. J., et al.: Developmental changes in Notch1 and numb expression mediated by local cell-cell interactions underlie progressively increasing delta sensitivity in neural crest stem cells. Dev. Biol. 244:199, 2002. 156. Kuhlbrodt, K., Herbarth, B., Sock, E., et al.: Cooperative function of POU proteins and SOX proteins in glial cells. J. Biol. Chem. 273:16050, 1998. 157. Kuhlbrodt, K., Herbarth, B., Sock, E., et al.: Sox10, a novel transcriptional modulator in glial cells. J. Neurosci. 18:237, 1998. 158. Kuhn, R., Monuki, E. S., and Lemke, G.: The gene encoding the transcription factor SCIP has features of an expressed retroposon. Mol. Cell. Biol. 11:4642, 1991. 159. Kurtz, A., Zimmer, A., Schnutgen, F., et al.: The expression pattern of a novel gene encoding brain-fatty acid binding protein correlates with neuronal and glial cell development. Development 120:2637, 1994. 160. Kury, P., Greiner-Petter, R., Cornely, C., et al.: Mammalian achaete scute homolog 2 is expressed in the adult sciatic nerve and regulates the expression of Krox24, Mob-1, CXCR4, and p57kip2 in Schwann cells. J. Neurosci. 22:7586, 2002. 161. Labatut-Cazabat, I., Vekris, A., and Petry, K. G.: A protein with the characters of a zinc-finger is implicated in the differentiation of Schwann cells. Neuroreport 10:3037, 1999. 162. Lallemand, D., Curto, M., Saotome, I., et al.: NF2 deficiency promotes tumorigenesis and metastasis by destabilizing adherens junctions. Genes Dev. 17:1090, 2003.

370

Neurobiology of the Peripheral Nervous System

163. Landmesser, L., Dahm, L., Schultz, K., and Rutishauser, U.: Distinct roles for adhesion molecules during innervation of embryonic chick muscle. Dev. Biol. 130:645, 1988. 164. Landon, D. N, and Wiseman, O. J.: A pacinian corpuscle in the human bladder lamina propria. J. Neurocytol. 30:457, 2001. 165. Le, X. F., Varela, C. R., and Bast, Jr. R. C.: Heregulin-induced apoptosis. Apoptosis 7:483, 2002. 166. Le Douarin, N. M., and Kalcheim, C.: The Neural Crest. Cambridge, UK, Cambridge University Press, 1999. 167. Lee, J. E., Hollenberg, S. M., Snider, L., et al.: Conversion of Xenopus ectoderm into neurons by NeuroD, a basic helix-loop-helix protein. Science 268:836, 1995. 168. Lee, M.-J., Brennan, A., Blanchard, A., et al.: P0 is constitutively expressed in the rat neural crest and embryonic nerves and is negatively and positively regulated by axons to generate non-myelin-forming and myelin-forming Schwann cells, respectively. Mol. Cell. Neurosci. 8:336, 1997. 169. Lee, M.-J., Calle, E., Brennan, A., et al.: In early development of the rat mRNA for the major myelin protein P0 is expressed in nonsensory areas of the embryonic inner ear, notochord, enteric nervous system, and olfactory ensheathing cells. Dev. Dyn. 222:40, 2001. 170. Lee, M. M., Sato-Bigbee, C., and De Vries, G. H.: Schwann cells stimulated by axolemma-enriched fractions express cyclic AMP responsive element binding protein. J. Neurosci. Res. 46:204, 1996. 171. Leimeroth, R., Lobsiger, C., Lüssi, A., et al.: Membranebound neuregulin1 type III actively promotes Schwann cell differentiation of multipotent progenitor cells. Dev. Biol. 246:245, 2002. 172. Lemke, G.: Neuregulins in development. Mol. Cell. Neurosci. 7:247, 1996. 173. Lemke, G.: Glial control of neuronal development. Annu. Rev. Neurosci. 24:87, 2001. 174. Lemke, G., and Chao, M.: Axons regulate Schwann cell expression of the major myelin and NGF receptor genes. Development 102:499, 1988. 175. Lemke, G., Kuhn, R., Monuki, E. S., and Weinmaster, G.: Expression and activity of the transcription factor SCIP during glial differentiation and myelination. Ann. N. Y. Acad. Sci. 633:189, 1991. 176. Lentz, S. I., Miner, J. H., Sanes, J. R, and Snider, W. D.: Distribution of the ten known laminin chains in the pathways and targets of developing sensory axons. J. Comp. Neurol. 378:547, 1997. 177. Levi, A. D., Bunge, R. P., Lofgren, J. A., et al.: The influence of heregulins on human Schwann cell proliferation. J. Neurosci. 15:1329, 1995. 178. Li, H., Wigley, C., and Hall, S. M.: Chronically denervated rat Schwann cells respond to GGF in vitro. Glia 24:290, 1998. 179. Li, Y., Tennekoon, G. I., Birnbaum, M., et al.: Neuregulin signaling through a PI3K/Akt/Bad pathway in Schwann cell survival. Mol. Cell. Neurosci. 17:761, 2001. 180. Lin, D. M., and Goodman, C. S.: Ectopic and increased expression of Fasciclin II alters motoneuron growth cone guidance. Neuron 13:507, 1994. 181. Livesey, F. J., O’Brien, J. A., Li, M., et al.: A Schwann cell mitogen accompanying regeneration of motor neurons. Nature 390:614, 1997.

182. Lloyd, A. C., Obermuller, F., Staddon, S., et al.: Cooperating oncogenes converge to regulate cyclin/cdk complexes. Genes Dev. 11:663, 1997. 183. Lo, L., Tiveron, M. C., and Anderson, D. J.: MASH1 activates expression of the paired homeodomain transcription factor Phox2a, and couples pan-neuronal and subtype-specific components of autonomic neuronal identity. Development 125:609, 1998. 184. Lobsiger, C. S., Taylor, V., and Suter, U.: The early life of a Schwann cell. Biol. Chem. 383:245, 2002. 185. Loeb, J. A., Khurana, T. S., Robbins, J. T., et al.: Expression patterns of transmembrane and released forms of neuregulin during spinal cord and neuromuscular synapse development. Development 126:781, 1999. 186. Lubischer, J. L., and Thompson, W. J.: Neonatal partial denervation results in nodal but not terminal sprouting and a decrease in efficacy of remaining neuromuscular junctions in rat soleus muscle. J. Neurosci. 19:8931, 1999. 187. Lyons, S. A., Morell, P., and McCarthy, K. D.: Schwann cells exhibit P2Y purinergic receptors that regulate intracellular calcium and are up-regulated by cyclic AMP analogues. J. Neurochem. 63:552, 1994. 188. Lyons, S. A., Morell, P., and McCarthy, K. D.: Schwann cell ATP-mediated calcium increases in vitro and in situ are dependent on contact with neurons. Glia 13:27, 1995. 189. Magnaghi, V., Cavarretta, I., Galbiati, M., et al.: Neuroactive steroids and peripheral myelin proteins. Brain Res. Brain Res. Rev. 37:360, 2001. 190. Mandemakers, W., Zwart, R., Jaegle, M., et al.: A distal Schwann cell-specific enhancer mediates axonal regulation of the Oct-6 transcription factor during peripheral nerve development and regeneration. EMBO J. 19:2992, 2000. 191. Marchionni, M. A., Goodearl, A. D. J., Chen, M. S., et al.: Glial growth factors are alternatively spliced erbB2 ligands expressed in the nervous system. Nature 362:312, 1993. 192. Martini, R., Bollensen, E., and Schachner, M.: Immunocytological localization of the major peripheral nervous system glycoprotein P0 and the L2/HNK-1 and L3 carbohydrate structures in developing and adult mouse sciatic nerve. Dev. Biol. 129:330, 1988. 193. Masaki, T., Matsumura, K., Hirata, A., et al.: Expression of dystroglycan and the laminin-alpha 2 chain in the rat peripheral nerve during development. Exp. Neurol. 174:109, 2002. 194. Matheny, C., DiStephano, P. S., and Milbrandt, J.: Differential expression of NGF receptor and early response genes in neural crest-derived cells. Mol. Brain Res. 13:75, 1992. 195. Mathon, N. F., Malcolm, D. S., Harrisingh, M. C., et al.: Lack of replicative senescence in normal rodent glia. Science 291:872, 2001. 196. Matsumura, K., Chiba, A., Yamada, H., et al.: A role of dystroglycan in Schwannoma cell adhesion to laminin. J. Biol. Chem. 272:13904, 1997b. 197. Matsumura, K., Yamada, H., Saito, F., et al.: Peripheral nerve involvement in merosin-deficient congenital muscular dystrophy and dy mouse. Neuromuscul. Disord. 7:7, 1997a. 198. Maurel, P., and Salzer, J. L.: Axonal regulation of Schwann cell proliferation and survival and the initial events of myelination require PI 3-kinase activity. J. Neurosci. 20:4635, 2000.

Molecular Signaling in Schwann Cell Development 199. Mayer, C., Wachtler, J., Kamleiter, M., and Grafe, P.: Intracellular calcium transients mediated by P2 receptors in the paranodal Schwann cell region of myelinated rat spinal root axons. Neurosci. Lett. 224:49, 1997. 200. Meier, C., Parmantier, E., Brennan, A., et al.: Developing Schwann cells acquire the ability to survive without axons by establishing an autocrine circuit involving IGF, NT-3 and PDGF-BB. J. Neurosci. 19:3847, 1999. 201. Meintanis, S., Thomaidou, D., Jessen, K. R., et al.: The neuron-glia signal beta-neuregulin promotes Schwann cell motility via the MAPK pathway. Glia 34:39, 2001. 202. Memberg, S. P., and Hall, A. K.: Dividing neuron precursors express neuron-specific tubulin. J. Neurobiol. 27:26, 1995. 203. Mews, M., and Meyer, M.: Modulation of Schwann cell phenotype by TGF-beta 1: inhibition of P0 mRNA expression and downregulation of the low affinity NGF receptor. Glia 8:208, 1993. 204. Meyer, D., and Birchmeier, C.: Multiple essential functions of neuregulin in development. Nature 378:386, 1995. 205. Meyer, D., Yamaii, T., Garratt, A., et al.: Isoform-specific expression and function of neuregulin. Development 124:3575, 1997. 206. Milner, R., Wilby, M., Nishimura, S., et al.: Division of labor of Schwann cell integrins during migration on peripheral nerve extracellular matrix ligands. Dev. Biol. 185:215, 1997. 207. Mirsky, R., Dubois, C., Morgan, L., and Jessen, K. R.: 04 and A007-sulfatide antibodies bind to embryonic Schwann cells prior to the appearance of galactocerebroside; regulation of the antigen by axon-Schwann cell signals and cyclic AMP. Development 109:105, 1990. 208. Mirsky, R., and Jessen, K. R.: Schwann cell development, differentiation and myelination. Curr. Opin. Neurobiol. 6:89, 1996. 209. Mirsky, R., and Jessen, K. R.: The neurobiology of Schwann cells. Brain Pathol. 9:293, 1999. 210. Mirsky, R., and Jessen, K. R.: Embryonic and early postnatal development of Schwann cells. In Jessen, K. R., and Richardson, W. D. (eds.): Glial Cell Development: Basic Principles and Clinical Relevance, 2nd ed. Oxford, UK, Oxford University Press, p. 1, 2001. 211. Monuki, E. S., Kuhn, R., and Lemke, G.: Repression of the myelin P0 gene by the POU transcription factor SCIP. Mech. Dev. 42:15, 1993. 212. Monuki, E. S., Kuhn, R., Weinmaster, G., et al.: Expression and activity of the POU transcription factor SCIP. Science 249:1300, 1990. 213. Monuki, E. S., Weinmaster, G., Kuhn, R., and Lemke, G.: SCIP: a glial POU domain gene regulated by cyclic AMP. Neuron 3:783, 1989. 214. Morgan, L., Jessen, K. R., and Mirsky, R.: The effects of cAMP on differentiation of cultured Schwann cells: progression from an early phenotype (04⫹ ) to a myelin phenotype (P0⫹ , GFAP⫺, N-CAM⫺, NGF-receptor⫺) depends on growth inhibition. J. Cell Biol. 112:457, 1991. 215. Morgan, L., Jessen, K. R., and Mirsky, R.: Negative regulation of the P0 gene in Schwann cells: suppression of P0 mRNA and protein induction in cultured Schwann cells by FGF2 and TGF␤1, TGF␤2 and TGF␤3. Development 120:1399, 1994. 216. Morris, J. K., Lin, W., Hauser, C., et al.: Rescue of the cardiac defect in ErbB2 mutant mice reveals essential roles of ErbB2

217.

218.

219.

220.

221.

222.

223.

224.

225.

226.

227.

228.

229.

230.

231.

232.

371

in peripheral nervous system development. Neuron 23:273, 1999. Morrison, H., Sherman, L. S., Legg, J., et al.: The NF2 tumor suppressor gene product, merlin, mediates contact inhibition of growth through interactions with CD44. Genes Dev. 15:968, 2001. Morrison, S. J., Perez, S. E., Qiao, Z., et al.: Transient Notch activation initiates an irreversible switch from neurogenesis to gliogenesis by neural crest stem cells. Cell 101:499, 2000. Morrison, S. J., White, P. M., Zock, C., and Anderson, D. J.: Prospective identification, isolation by flow cytometry, and in vivo self-renewal of multipotent mammalian neural crest stem cells. Cell 96:737, 1999. Morrissey, T. K., Levi, A. D., Nuijens, A., et al.: Axon-induced mitogenesis of human Schwann cells involves heregulin and p185erbB2. Proc. Natl. Acad. Sci. U.S.A. 92:1431, 1995. Mukai, J., Hachiya, T., Shoji-Hoshino, S., et al.: NADE, a p75NTR-associated cell death executor, is involved in signal transduction mediated by the common neurotrophin receptor p75NTR. J. Biol. Chem. 275:17566, 2000. Mukai, J., Shoji, S., Kimura, M. T., et al.: Structure-function analysis of NADE: identification of regions that mediate nerve growth factor-induced apoptosis. J. Biol. Chem. 277:13973, 2002. Murphy, P., Topilko, P., Schneider-Manoury, S., et al.: The regulation of Krox-20 expression reveals important steps in the control of peripheral glial cell development. Development 122:2847, 1996. Nagano, S., Takeda, M., Ma, L., and Soliven, B.: Cytokineinduced cell death in immortalized Schwann cells: roles of nitric oxide and cyclic AMP. J. Neurochem. 77:1486, 2001. Nagarajan, R., Le, N., Mahoney, H., et al.: Deciphering peripheral nerve myelination by using Schwann cell expression profiling. Proc. Natl. Acad. Sci. U.S.A. 99:8998, 2002. Nagarajan, R., Svaren, J., Le, N., et al.: EGR2 mutations in inherited neuropathies dominant-negatively inhibit myelin gene expression. Neuron 30:355, 2001. Nakagawa, M., Miyagoe-Suzuki, Y., Ikezoe, K., et al.: Schwann cell myelination occurred without basal lamina formation in laminin alpha2 chain-null mutant (dy3K/dy3K) mice. Glia 35:101, 2001. Nataf, V., and Le Douarin, N. M.: Induction of melanogenesis by tetradecanoylphorbol-13 acetate and endothelin 3 in embryonic avian peripheral nerve cultures. Pigment Cell Res. 13:172, 2000. Nave, K.-A.: Myelin-specific genes and their mutations in the mouse. In Jessen, K. R., and Richardson, W. D. (eds.): Glial Cell Development: Basic Principles and Clinical Relevance, 2nd ed. Oxford, UK, Oxford University Press, p. 177, 2001. Neve, R. M., Holbro, T., and Hynes, N. E.: Distinct roles for phosphoinositide 3-kinase, mitogen-activated protein kinase and p38 MAPK in mediating cell cycle progression of breast cancer cells. Oncogene 21:4567, 2002. Ng, V., Zanazzi, G., Timpl, R., et al.: Role of the cell wall phenolic glycolipid-1 in the peripheral nerve predilection of Mycobacterium leprae. Cell 103:511, 2000. Nichols, D. H., and Weston, J. A.: Melanogenesis in cultures of peripheral nervous tissue. I. The origin and prospective fate of cells giving rise to melanocytes. Dev. Biol. 60:217, 1977.

372

Neurobiology of the Peripheral Nervous System

233. Nickols, J. C., Valentine, W., Kanwal, S., and Carter, B. D.: Activation of the transcription factor NF-kappaB in Schwann cells is required for peripheral myelin formation. Nat. Neurosci. 6:161, 2003. 234. Niemann, C., Brinkmann, V., and Birchmeier, W.: Hepatocyte growth factor and neuregulin in mammary gland cell morphogenesis. Adv. Exp. Med. Biol. 480:9, 2000. 235. Niessen, C. M., Cremona, O., Daams, H., et al.: Expression of the integrin ␣6␤4 in peripheral nerves: localization in Schwann and perineural cells and different variants of the ␤4 subunit. J. Cell Sci. 107:543, 1994. 236. Nikam, S. S., Tennekoon, G. I., Christy, B. A., et al.: The zinc finger transcription factor Zif268/Egr-1 is essential for Schwann cell expression of the p75 NGF receptor. Mol. Cell. Neurosci. 6:337, 1995. 237. Notterpek, L., Snipes, G. J., and Shooter, E. M.: Temporal expression pattern of peripheral myelin protein 22 during in vivo and in vitro myelination. Glia 25:358, 1999. 238. Okazaki, S., Tanase, S., Choudhury, B. K., et al.: A novel nuclear protein with zinc fingers down-regulated during early mammalian cell differentiation. J. Biol. Chem. 269:6900, 1994. 239. Orr-Urtreger, A., Trakhtenbrot, L., Ben-Levy, R., et al.: Neural expression and chromosomal mapping of Neu differentiation factor to 8p12-p21. Proc. Natl. Acad. Sci. U.S.A. 90:1867, 1993. 240. Pannese, E.: The satellite cells of the sensory ganglia. Adv. Anat. Embryol. Cell Biol. 65:1, 1981. 241. Pannese, E.: Neurocytology: Fine Structure of Neurons, Nerve Processes, and Neuroglial Cells. Stuttgart, Thieme, 1994. 242. Paratore, C., Brugnoli, G., Lee, H. Y., et al.: The role of the Ets domain transcription factor Erm in modulating differentiation of neural crest stem cells. Dev. Biol. 250:168, 2002. 243. Paratore, C., Goerich, D. E., Suter, U., et al.: Survival and glial fate acquisition of neural crest cells are regulated by an interplay between the transcription factor Sox10 and extrinsic combinatorial signaling. Development 128:3949, 2001. 244. Pareyson, D., Taroni, F., Botti, S., et al.: Cranial nerve involvement in CMT disease type 1 due to early growth response 2 gene mutation. Neurology 54:1696, 2000. 245. Parkinson, D. B., Bhaskaran, A., Droggiti, A., et al.: Krox-20 inhibits Jun-NH2-terminal kinase/c-jun to control Schwann cell proliferation and death. J. Cell Biol. 164:385, 2004. 246. Parkinson, D. B., Dickinson, S., Bhaskaran, A., et al.: Krox 20 activates a set of complex changes in Schwann cells that characterize myelination. Glia Suppl. 1:S69, 2002. 247. Parkinson, D. B., Dickinson, S., Bhaskaran, A., et al.: Regulation of the myelin gene periaxin provides evidence for Krox-20 independent myelin-related signalling in Schwann cells. Mol. Cell. Neurosci. 23:13, 2003. 248. Parkinson, D. B., Dong, Z., Bunting, H., et al.: Transforming growth factor ␤ (TGF␤) mediates Schwann cell death in vitro and in vivo: examination of c-Jun activation, interactions with survival signals, and the relationship of TGF␤ mediated death to Schwann cell differentiation. J. Neurosci. 21:8572, 2001. 249. Parkinson, D. B., Langner, K., Namini, S. S., et al.: ␤-Neuregulin and autocrine mediated survival of Schwann

250.

251.

252.

253.

254.

255.

256.

257.

258.

259. 260.

261.

262.

263.

264.

265.

266.

267.

cells requires activity of Ets family transcription factors. Mol. Cell. Neurosci. 20:154, 2002. Parmantier, E., Braun, C., Thomas, J. L., et al.: PMP-22 expression in the central nervous system of the embryonic mouse defines potential transverse segments and longitudinal columns. J. Comp. Neurol. 378:159, 1997. Parmantier, E., Cabon, F., Braun, C., et al.: Peripheral myelin protein-22 is expressed in rat and mouse brain and spinal cord motoneurons. Eur. J. Neurosci. 7:1080, 1995. Parmantier, E., Lynn, B., Lawson, D., et al.: Schwann cell-derived Desert Hedgehog controls the development of peripheral nerve sheaths. Neuron 23:713, 1999. Peirano, R. I., Goerich, D. E., Riethmacher, D., and Wegner, M.: Protein zero gene expression is regulated by the glial transcription factor Sox10. Mol. Cell. Biol. 20:3198, 2000. Plantinga, L. C., Verhaagen, J., Edwards, P. M., et al.: The expression of B-50/GAP-43 in Schwann cells is upregulated in degenerating peripheral nerve stumps following nerve injury. Brain Res. 602:69, 1993. Previtali, S., Nodari, A., Dina, G., et al.: Cell-type specific knockouts in the study of glia-basement membrane interactions. Glia Suppl. 1:S7, 2002. Previtali, S. C., Feltri, M. L., Archelos, J. J., et al.: Role of integrins in the peripheral nervous system. Prog. Neurobiol. 64:35, 2001. Pusch, C., Hustert, E., Pfeifer, D., et al.: The SOX10/Sox10 gene from human and mouse: sequence, expression, and transactivation by the encoded HMG domain transcription factor. Hum. Genet. 103:115, 1998. Rahmatullah, M., Schroering, A., Rothblum, K., et al.: Synergistic regulation of Schwann cell proliferation by heregulin and forskolin. Mol. Cell. Biol. 18:6245, 1998. Raisman, G.: Olfactory ensheathing cells—another miracle cure for spinal cord injury? Nat. Rev. Neurosci. 2:369, 2001. Rambukkana, A., Salzer, J. L., Yurchenco, P. D., and Tuomanen, E. I.: Neural targeting of Mycobacterium leprae mediated by the G domain of the laminin-alpha2 chain. Cell 88:811, 1997. Rambukkana, A., Yamada, H., Zanazzi, G., et al.: Role of alpha-dystroglycan as a Schwann cell receptor for Mycobacterium leprae. Science 282:2076, 1998. Rambukkana, A., Zanazzi, G., Tapinos, N., and Salzer, J. L.: Contact-dependent demyelination by Mycobacterium leprae in the absence of immune cells. Science 296:927, 2002. Ramon-Cueto, A., and Santos-Benito, F. F.: Cell therapy to repair injured spinal cords: olfactory ensheathing glia transplantation. Restor. Neurol. Neurosci. 19:149, 2001. Ridley, A. J., Davis, J. B., Stroobant, P., and Land, H.: Transforming growth factors-beta 1 and beta 2 are mitogens for rat Schwann cells. J. Cell Biol. 109:3419, 1989. Ridley, A. J., Paterson, H. F., Noble, M., and Land, H.: Rasmediated cell cycle arrest is altered by nuclear oncogenes to induce Schwann cell transformation. EMBO J. 7:1635, 1988. Riethmacher, D., Sonnenberg-Riethmacher, E., Brinkmann, V., et al.: Severe neuropathies in mice with targeted mutations in the ErbB3 receptor. Nature 389:725, 1997. Rizvi, T. A., Huang, Y., Sidani, A., et al.: A novel cytokine pathway suppresses glial cell melanogenesis after injury to adult nerve. J. Neurosci. 22:9831, 2002.

Molecular Signaling in Schwann Cell Development 268. Robitaille, R.: Modulation of synaptic efficacy and synaptic depression by glial cells at the frog neuromuscular junction. Neuron 21:847, 1998. 269. Rosenbaum, C., Karyala, S., Marchionni, M. A., et al.: Schwann cells express NDF and SMDF/n-ARIA mRNAs, secrete neuregulin, and show constitutive activation of erbB3 receptors: evidence for a neuregulin autocrine loop. Exp. Neurol. 148:604, 1997. 270. Rouleau, G. A., Merel, P., Lutchman, M., et al.: Alteration in a new gene encoding a putative membrane-organizing protein causes neuro-fibromatosis type 2. Nature 363:515, 1993. 271. Russell, J. W., Cheng, H. L., and Golovoy, D.: Insulin-like growth factor-I promotes myelination of peripheral sensory axons. J. Neuropathol. Exp. Neurol. 59:575, 2000. 272. Sabourin, L. A., and Rudnicki, M. A.: The molecular regulation of myogenesis. Clin. Genet. 57:16, 2000. 273. Sadoulet-Puccio, H. M., and Kunkel, L. M.: Dystrophin and its isoforms. Brain Pathol. 6:25, 1996. 274. Sainio, M., Zhao, F., Heiska, L., et al.: Neurofibromatosis 2 tumor suppressor protein colocalizes with ezrin and CD44 and associates with actin-containing cytoskeleton. J. Cell Sci. 110:2249, 1997. 275. Salehi, A. H., Roux, P. P., Kubu, C. J., et al.: NRAGE, a novel MAGE protein, interacts with the p75 neurotrophin receptor and facilitates nerve growth factor-dependent apoptosis. Neuron 27:279, 2000. 276. Saunders, R. D., and DeVries, G. H.: Schwann cell proliferation is accompanied by enhanced inositol phospholipid metabolism. J. Neurochem. 50:876, 1988. 277. Scherer, S. S.: The biology and pathobiology of Schwann cells. Curr. Opin. Neurol. 10:386, 1997. 278. Scherer, S. S., and Gutmann, D. H.: Expression of the neurofibromatosis 2 tumor suppressor gene product, merlin, in Schwann cells. J. Neurosci. Res. 46:595, 1996. 279. Scherer, S. S., and Salzer, J. L.: Axon-Schwann cell interactions during peripheral nerve degeneration and regeneration. In Jessen, K. R., and Richardson, W. D. (eds.): Glial Cell Development: Basic Principles and Clinical Relevance, 2nd ed. Oxford, UK, Oxford University Press, p. 299, 2001. 280. Scherer, S. S., Wang, D.-Y., Kuhn, R., et al.: Axons regulate Schwann cell expression of the POU transcription factor SCIP. J. Neurosci. 14:1930, 1994. 281. Scherer, S. S., Xu, Y. T., Bannerman, P. G., et al.: Periaxin expression in myelinating Schwann cells: modulation by axon-glial interactions and polarized localization during development. Development 121:4265, 1995. 282. Scherer, S. S., Xu, Y. T., Roling, D., et al.: Expression of growth-associated protein-43 kD in Schwann cells is regulated by axon-Schwann cell interactions and cAMP. J. Neurosci. Res. 38:575, 1994. 283. Schneider, C., Wicht, H., Enderich, J., et al.: Bone morphogenetic proteins are required in vivo for the generation of sympathetic neurons. Neuron 24:861, 1999. 284. Schulze, K. M., Hanemann, C. O., Muller, H. W., and Hanenberg, H.: Transduction of wild-type merlin into human schwannoma cells decreases schwannoma cell growth and induces apoptosis. Hum. Mol. Genet. 11:69, 2002.

373

285. Schumacher, M., Guennoun, R., Mercier, G., et al.: Progesterone synthesis and myelin formation in peripheral nerves. Brain Res. Brain Res. Rev. 37:343, 2001. 286. Scoles, D. R., Huynh, D. P., Morcos, P. A., et al.: Neurofibromatosis 2 tumour suppressor schwannomin interacts with ␤11-spectrin. Nat. Gen. 18:354, 1998. 287. Seilheimer, B., Persohn, E., and Schachner, M.: Antibodies to the L1 adhesion molecule inhibit Schwann cell ensheathment of neurons in vitro. J. Cell Biol. 109:3095, 1989. 288. Seilheimer, B., and Schachner, M.: Studies of adhesion molecules mediating interactions between cells of peripheral nervous system indicate a major role for L1 in mediating sensory neuron growth on Schwann cells in culture. J. Cell Biol. 107:341, 1988. 289. Sensenbrenner, M., Lucas, M., and Deloulme, J. C.: Expression of two neuronal markers, growth-associated protein 43 and neuron-specific enolase, in rat glial cells. J. Mol. Med. 75:653, 1997. 290. Shah, N. M., Groves, A. K., and Anderson, D. J.: Alternative neural crest cell fates are instructively promoted by TGF␤ superfamily members. Cell 85:331, 1996. 291. Shah, N. M., Marchionni, M. A., Isaacs, I., et al.: Glial growth factor restricts mammalian neural crest stem cells to a glial fate. Cell 77:349, 1994. 292. Shaw, R. J., Paez, J. G., Curto, M., et al.: The Nf2 tumor suppressor, merlin, functions in Rac-dependent signaling. Dev. Cell 1:63, 2001. 293. Sherman, D. L., Fabrizi, C., Gillespie, C. S., and Brophy, P. J.: Specific disruption of a Schwann cell dystrophin-related protein complex in a demyelinating neuropathy. Neuron 30:677, 2001. 294. Sherman, L., Stocker, K. M., Morrison, R., and Ciment, G.: Basic fibroblast growth factor (bFGF) acts intracellularly to cause the transdifferentiation of avian neural crest-derived Schwann cell precursors into melanocytes. Development 118:1313, 1993. 295. Sherman, L. S., and Gutmann, D. H.: Merlin: hanging tumor suppression on the Rac. Trends Cell Biol. 11:442, 2001. 296. Sherman, L. S., Rizvi, T. A., Karyala, S., and Ratner, N.: CD44 enhances neuregulin signaling by Schwann cells. J. Cell Biol. 150:1071, 2000. 297. Shibuya, Y., Mizoguchi, A., Takeichi, M., et al.: Localization of N-cadherin in the normal and regenerating nerve fibers of the chicken peripheral nervous system. Neuroscience 67:253, 1995. 298. Sim, F., Zhao, C., Li, W., et al.: Expression of the POUdomain transcription factors SCIP/Oct-6 and Brn-2 is associated with Schwann cell but not oligodendrocyte remyelination of the CNS. Mol. Cell. Neurosci. 20:669, 2002. 299. Skoff, A. M., Lisak, R. P., Bealmear, B., and Benjamins, J. A.: TNF-alpha and TGF-beta act synergistically to kill Schwann cells. J. Neurosci. Res. 53:747, 1998. 300. Sobue, G., and Pleasure, D.: Schwann cell galactocerebroside induced by derivatives of adenosine 3',5'-monophosphate. Science 224:72, 1984. 301. Soilu-Hanninen, M., Ekert, P., Bucci, T., et al.: Nerve growth factor signaling through p75 induces apoptosis in Schwann cells via a Bcl-2-independent pathway. J. Neurosci. 19:4828, 1999.

374

Neurobiology of the Peripheral Nervous System

302. Sonnenberg-Riethmacher, E., Miehe, M., Stolt, C. C., et al.: Development and degeneration of dorsal root ganglia in the absence of the HMG-domain transcription factor Sox10. Mech. Dev. 109:253, 2001. 303. Southard-Smith, E. M., Kos, L., and Pavan, W. J.: Sox 10 mutation disrupts neural crest development in Dom Hirschsprung mouse model. Nat. Genet. 18:60, 1998. 304. Spassky, N., Goujet-Zalc, C., Parmantier, E., et al.: Multiple restricted origin of oligodendrocytes. J. Neurosci. 18:8331, 1998. 305. Spencer, P. S., and Schaumburg, H. H.: An ultrastructural study of the inner core of the pacinian corpuscle. J. Neurocytol. 2:217, 1973. 306. Stevens, B., and Fields, R. D.: Response of Schwann cells to action potentials in development. Science 287:2267, 2000. 307. Stevens, B., Tanner, S., and Fields, R. D.: Control of myelination by specific patterns of neural impulses. J. Neurosci. 18:9303, 1998. 308. Stewart, H. J., Curtis, R., Jessen, K. R., and Mirsky, R.: TGF-betas and cAMP regulate GAP-43 expression in Schwann cells and reveal the association of this protein with the trans-Golgi network. Eur. J. Neurosci. 7:1761, 1995. 309. Stewart, H. J. S.: Expression of c-Jun, Jun B, Jun D and cAMP response element binding protein by Schwann cells and their precursors in vivo and in vitro. Eur. J. Neurosci. 7:1366, 1995. 310. Stewart, H. J. S., Bradke, F., Tabernero, A., et al.: Regulation of rat Schwann cell P0 expression and DNA synthesis by insulin-like growth factors in vitro. Eur. J. Neurosci. 8:553, 1996. 311. Stewart, H. J. S., Brennan, A., Rahman, M., et al.: Developmental regulation and overexpression of the transcription factor AP-2, a potential regulator of the timing of Schwann cell generation. Eur. J. Neurosci. 14:363, 2001. 312. Stewart, H. J. S., Mirsky, R., and Jessen, K. R.: The Schwann cell lineage: embryonic and early postnatal development. In Jessen, K. R., and Richardson, W. D. (eds.): Glial Cell Development: Basic Principles and Clinical Relevance. Oxford, UK, Bios Scientific Publishers Ltd, p. 1, 1996. 313. Stewart, H. J. S., Morgan, L., Jessen, K. R., and Mirsky, R.: Changes in DNA synthesis rate in the Schwann cell lineage in vivo are correlated with the precursor-Schwann cell transition and myelination. Eur. J. Neurosci. 5:1136, 1993. 314. Stewart, H. J. S., Zoidl, G., Rossner, M., et al.: Helix-loophelix proteins in Schwann cells: a study of regulation and subcellular localization of Ids, REB and E12/47 during embryonic and postnatal development. J. Neurosci. Res. 50:684, 1997. 315. Stiles, C. D., Pledger, W. J., Tucker, R. W., et al.: Regulation of the Balb/c-3T3 cell cycle-effects of growth factors. J. Supramol. Struct. 13:489, 1980. 316. Stocker, K. M., Sherman, L., Rees, S., and Ciment, G.: Basic FGF and TGF-␤1 influence commitment to melanogenesis in neural crest-derived cells of avian embryos. Development 111:635, 1991. 317. Stolt, C. C., Rehberg, S., Ader, M., et al.: Terminal differentiation of myelin-forming oligodendrocytes depends on the transcription factor Sox10. Genes Dev. 16:165, 2002.

318. Sudhalter, J., Whitehouse, L., Rusche, J. R., et al.: Schwann cell heparan sulfate proteoglycans play a critical role in glial growth factor/neuregulin signaling. Glia 17:28, 1996. 319. Sulaiman, O. A., and Gordon, T.: Transforming growth factorbeta and forskolin attenuate the adverse effects of long-term Schwann cell denervation on peripheral nerve regeneration in vivo. Glia 37:206, 2002. 320. Sun, C. X., Haipek, C., Scoles, D. R., et al.: Functional analysis of the relationship between the neurofibromatosis 2 tumor suppressor and its binding partner, hepatocyte growth factor-regulated tyrosine kinase substrate. Hum. Mol. Genet. 11:3167, 2002. 321. Syroid, D. E., Maycox, P. J., Soilu-Hanninen, M., et al.: Induction of postnatal Schwann cell death by the low-affinity neurotrophin receptor in vitro and after axotomy. J. Neurosci. 20:5741, 2000. 322. Syroid, D. E., Maycox, P. R., Burrola, P. G., et al.: Cell death in the Schwann cell lineage and its regulation by neuregulin. Proc. Natl. Acad. Sci. U.S.A. 93:9229, 1996. 323. Syroid, D. E., Zorick, T. S., Arbet-Engels, C., et al.: A role for insulin-like growth factor-I in the regulation of Schwann cell survival. J. Neurosci. 19:2059, 1999. 324. Tabernero, A., Stewart, H. J. S., Jessen, K. R., and Mirsky, R.: The neuron-glia signal ␤ neuregulin induces sustained CREB phosphorylation on Ser-133 in cultured rat Schwann cells. Mol. Cell. Neurosci. 10:309, 1998. 325. Takashima, H., Boerkoel, C. F., De Jonghe, P., et al.: Periaxin mutations cause a broad spectrum of demyelinating neuropathies. Ann. Neurol. 51:709, 2002. 326. Tanigaki, K., Nogaki, F., Takahashi, J., et al.: Notch1 and Notch3 instructively restrict bFGF-responsive multipotent neural progenitor cells to an astroglial fate. Neuron 29:45, 2001. 327. Taylor, A. R., Geden, S. E., and Fernandez-Valle, C.: Formation of a beta1 integrin signaling complex in Schwann cells is independent of rho. Glia 41:94, 2003. 328. Thatikunta, P., Qin, W., Christy, B. A., et al.: Reciprocal Id expression and myelin gene regulation in Schwann cells. Mol. Cell. Neurosci. 14:519, 1999. 329. Thompson, M. A., and Ziff, E. B.: Structure of the gene encoding peripherin, an NGF-regulated neuronal-specific type III intermediate filament protein. Neuron 2:1043, 1989. 330. Tikoo, R., Zanazzi, G., Shiffman, D., et al.: Cell cycle control of Schwann cell proliferation: role of cyclin-dependent kinase-2. J. Neurosci. 20:4627, 2000. 331. Timmerman, V., De Jonghe, P., Ceuterick, C., et al.: Novel missense mutation in the early growth response 2 gene associated with Dejerine-Sottas syndrome phenotype. Neurology 52:1827, 1999. 332. Timsit, S. G., Bally-Cuif, L., Colman, D. R., and Zalc, B.: DM-20 mRNA is expressed during the embryonic development of the nervous system of the mouse. J. Neurochem. 58:1172, 1992. 333. Tiveron, M.-C., Hirsch, M.-R., and Brunet, J.-F.: The expression pattern of the transcription factor Phox2 delineates synaptic pathways of the autonomic nervous system. J. Neurosci. 16:7649, 1996. 334. Tontonoz, P., Hu, E., and Spiegelman, B. M.: Stimulation of adipogenesis in fibroblasts by PPAR gamma 2, a lipidactivated transcription factor. Cell 79:1147, 1994.

Molecular Signaling in Schwann Cell Development 335. Topilko, P., Levi, G., Merlo, G., et al.: Differential regulation of the zinc finger genes Krox-20 and Krox-24 (Egr-1) suggests antagonistic roles in Schwann cells. J. Neurosci. Res. 50:702, 1997. 336. Topilko, P., and Meijer, D.: Transcription factors that control Schwann cell development and myelination. In Jessen, K. R., and Richardson, W. D. (eds.): Glial Cell Development: Basic Principles and Clinical Relevance, 2nd ed. Oxford, UK, Oxford University Press, p. 223, 2001. 337. Topilko, P., Schneider-Manoury, S., Levi, G., et al.: Krox-20 controls myelination in the peripheral nervous system. Nature 371:796, 1994. 338. Topilko, P., Schneider-Maunoury, S., Levi, G., et al.: Multiple pituitary and ovarian defects in Krox-24 (NGFI-A, Egr-1)-targeted mice. Mol. Endocrinol. 12:107, 1998. 339. Tosh, D., and Slack, J. M.: How cells change their phenotype. Nat. Rev. Mol. Cell Biol. 3:187, 2002. 340. Trachtenberg, J. T., and Thompson, W. J.: Schwann cell apoptosis at developing neuromuscular junctions is regulated by glial growth factor. Nature 379:174, 1996. 341. Trofatter, J. A., MacCollin, M. M., Rutter, J. L., et al.: A novel moesin-, ezrin-, radixin-like gene is a candidate for the neurofibromatosis 2 tumor suppressor. Cell 72:791, 1993. 342. Tsiper, M. V., and Yurchenco, P. D.: Laminin assembles into separate basement membrane and fibrillar matrices in Schwann cells. J. Cell Sci. 115:1005, 2002. 343. Uziyel, Y., Hall, S., and Cohen, J.: Influence of laminin-2 on Schwann cell-axon interactions. Glia 32:109, 2000. 344. van der Neut, R., Krimpenfort, P., Calafat, J., et al.: Epithelial detachment due to absence of hemidesmosomes in integrin beta 4 null mice. Nat. Genet. 13:366, 1996. 345. Vartanian, T., Goodearl, A., Viehover, A., and Fischbach, G.: Axonal neuregulin signals cells of the oligodendrocyte lineage through activation of HER4 and Schwann cells through HER2 and HER3. J. Cell Biol. 137:211, 1997. 346. Wakamatsu, Y., Maynard, T. M., and Weston, J. A.: Fate determination of neural crest cells by NOTCH-mediated lateral inhibition and asymmetrical cell division during gangliogenesis. Development 127:2811, 2000. 347. Wang, J. Y., Miller, S. J., and Falls, D. L.: The N-terminal region of neuregulin isoforms determines the accumulation of cell surface and released neuregulin ectodomain. J. Biol. Chem. 276:2841, 2001. 348. Wang, S., Sdrulla, A., Johnson, J. E., et al.: A role for the helix-loop-helix protein Id2 in the control of oligodendrocyte development. Neuron 29:603, 2001. 349. Wanner, I. B., and Wood, P. M.: N-cadherin mediates axon-aligned process growth and cell-cell interaction in rat Schwann cells. J. Neurosci. 22:4066, 2002. 350. Warner, L. E., Garcia, C. A., and Lupski, J. R.: Hereditary peripheral neuropathies: clinical forms, genetics, and molecular mechanisms. Annu. Rev. Med. 50:263, 1999. 351. Warner, L. E., Mancias, P., Butler, I. J., et al.: Mutations in the early growth response 2 (EGR2) gene are associated with hereditary myelinopathies. Nat. Genet. 18:382, 1998. 352. Wegner, M.: From head to toes: the multiple facets of Sox proteins. Nucleic Acids Res. 27:1409, 1999.

375

353. Wegner, M.: Transcriptional control in myelinating glia: the basic recipe. Glia 29:118, 2000. 354. Weiner, J. A., and Chun, J.: Schwann cell survival mediated by the signaling phospholipid lysophosphatidic acid. Proc. Natl. Acad. Sci. U.S.A. 96:5233, 1999. 355. Weiner, J. A., Fukushima, N., Contos, J. J., et al.: Regulation of Schwann cell morphology and adhesion by receptormediated lysophosphatidic acid signaling. J. Neurosci. 21:7069, 2001. 356. Weinmaster, G., and Lemke, G.: Cell-specific cyclic AMP-mediated induction of the PDGF receptor. EMBO J. 9:915, 1990. 357. Weinstein, D. E., Burrola, P. G., and Lemke, G.: Premature Schwann cell differentiation and hypermyelination in mice expressing a targeted antagonist of the POU transcription factor SCIP. Mol. Cell. Neurosci. 6:212, 1995. 358. White, P. M., Morrison, S. J., Orimoto, K., et al.: Neural crest stem cells undergo cell-intrinsic developmental changes in sensitivity to instructive differentiation signals. Neuron 29:57, 2001. 359. Williams, A. C., and Brophy, P. J.: The function of the periaxin gene during nerve repair in a model of CMT4F. J. Anat. 200:323, 2002. 360. Winseck, A. K., Calderó, J., Ciutat, D., et al.: In vivo analysis of Schwann cell programmed cell death in the embryonic chick: regulation by axons and glial growth factor. J. Neurosci. 22:4509, 2002. 361. Woldeyesus, M. T., Britsch, S., Riethmacher, D., et al.: Peripheral nervous system defects in erbB2 mutants following genetic rescue of heart development. Genes Dev. 13:2538, 1999. 362. Wolpowitz, D., Mason, T. B., Dietrich, P., et al.: Cysteinerich domain isoforms of the neuregulin-1 gene are required for maintenance of peripheral synapses. Neuron 25:79, 2000. 363. Wood, P. M., Schachner, M., and Bunge, R. P.: Inhibition of Schwann cell myelination in vitro by antibody to the L1 adhesion molecule. J. Neurosci. 10:3635, 1990. 364. Woodhoo, A., Dean, C. H., Droggiti, A., et al.: The trunk neural crest and its early glial derivatives: a study of survival responses, developmental schedules and autocrine mechanisms. Mol. Cell. Neurosci. 25:40, 2004. 365. Woolf, C. J., Reynolds, M. L., Chong, M. S., et al.: Denervation of the motor endplate results in the rapid expression by terminal Schwann cells of the growth-associated protein GAP-43. J. Neurosci. 12:3999, 1992. 366. Wu, R., Jurek, M., Sundarababu, S., and Weinstein, D. E.: The POU gene Brn-5 is induced by neuregulin and is restricted to myelinating Schwann cells. Mol. Cell. Neurosci. 17:683, 2001. 367. Yamashita, T., Tucker, K. L., and Barde, Y. A.: Neurotrophin binding to the p75 receptor modulates Rho activity and axonal outgrowth. Neuron 24:585, 1999. 368. Yoshihara, T., Kanda, F., Yamamoto, M., et al.: A novel missense mutation in the early growth response 2 gene associated with late-onset Charcot-Marie-Tooth disease type 1. J. Neurol. Sci. 184:149, 2001. 369. Yoshimura, T., Goda, S., Kobayashi, T., and Goto, I.: Involvement of protein kinase C in the proliferation of cultured Schwann cells. Brain Res. 617:55, 1993.

376

Neurobiology of the Peripheral Nervous System

370. Zanazzi, G., Einheber, S., Westreich, R., et al.: Glial growth factor/neuregulin inhibits Schwann cell myelination and induces demyelination. J. Cell Biol. 152:1289, 2001. 371. Zelena, J.: Nerves and Mechanoreceptors. London, Chapman and Hall, 1994. 372. Zhang, D., Sliwkowski, M. X., Mark, M., et al.: Neuregulin-3 (NRG3): a novel neural tissue-enriched protein that binds and activates ErbB4. Proc. Natl. Acad. Sci. U.S.A. 94:9562, 1997. 373. Zhang, J. Y., Luo, X. G., Xian, C. J., et al.: Endogenous BDNF is required for myelination and regeneration of injured sciatic nerve in rodents. Eur. J. Neurosci. 12:4171, 2000.

374. Zilian, O., Saner, C., Hagedorn, L., et al.: Multiple roles of mouse Numb in tuning developmental cell fates. Curr. Biol. 11:494, 2001. 375. Zorick, T. S., Syroid, D. E., Arroyo, E., et al.: The transcription factors SCIP and Krox-20 mark distinct stages and cell fates in Schwann cell differentiation. Mol. Cell. Neurosci. 8:129, 1996. 376. Zorick, T. S., Syroid, D. E., Brown, A., et al.: Krox-20 controls SCIP expression, cell cycle exit and susceptibility to apoptosis in developing myelinating Schwann cells. Development 126:1397, 1999.

17 Neurotrophic Factors in the Peripheral Nervous System ANTHONY J. WINDEBANK AND ELIZABETH S. MCDONALD

Neurotrophin Families Nerve Growth Factor Brain-Derived Neurotrophic Factor Neurotrophin-3/Neurotrophin-4

Ciliary Neurotrophic Factor Glial Cell Line–Derived Neurotrophic Factor Insulin-like Growth Factor

Growth factors may be defined as soluble extracellular macromolecules that influence the proliferation, growth, or differentiation of target cells by a cell surface receptor– mediated mechanism. “Soluble” is included in the definition because the known growth factors are assumed to work in vivo by signaling from a target tissue (e.g., muscle) to a neuronal subpopulation (e.g., the anterior horn cell). This process involves local diffusion from the target to the neuron. However, it is probable that extracellular matrix proteins such as laminin, fibronectin, and collagen play important trophic roles for neurons, especially during peripheral nerve regeneration. Similarly, direct cell membrane–to–cell membrane contact by way of surface-bound neuronal cell adhesion molecules may influence neuronal growth and differentiation. For operational convenience, small molecules, such as neurotransmitters, are not included in the definition of growth factors because little is known about their roles in modulating growth and differentiation. It is possible that these small molecules play an important part in modulating or determining cell survival or specific aspects of cellular differentiation such as maintenance of synapse structure. Practically all of the neuronal growth factors studied to date are polypeptides that have biologically active monomeric subunits with molecular weights in the 10,000- to 30,000-Da range. Nonpolypeptide macromolecules have not been identified as growth factors, although glycosaminoglycans such as heparin may be important in modulating the action of growth factors in vivo.

Neurotrophin Signaling Clinical Applications of Neurotrophic Factors

Growth factors are presumably involved in many different stages of the development and maintenance of the final shape and connectivity of neurons. A simplified conceptual framework is illustrated in Figure 17–1. This type of scheme has been developed from the work of Hamburger and Keefe34 and others, who demonstrated that ablation of a target, such as a limb bud, or addition of extra target tissue during embryonic development resulted in a decrease or increase, respectively, in the number or size of neurons innervating that target. These observations led directly to the discovery of nerve growth factor (NGF). It is assumed that timing of interaction, especially during embryonic development, is critical. Thus a population of neurons may express specific growth factor receptors on their surface at one stage in their development and not at others. It is also assumed that these molecules work over short distances within tissues and that this local interaction may provide for some of the specificity of action. Understanding timing of receptivity and tissue localization becomes increasingly important as human recombinant growth factors become available as potential therapeutic agents. A final general issue to be considered is the naming of growth factors. Names in present conventional use reflect either the tissue from which the factor was originally derived (e.g., ciliary neurotrophic factor [CNTF], brain-derived neurotrophic factor [BDNF]) or the target cell upon which the factor was found to act in vitro (e.g., fibroblast growth factor). These original discoveries may not reflect the true functions of the factors in vivo. However, because none 377

378

Neurobiology of the Peripheral Nervous System

NGF NGF

NGF NGF Cell membrane

Y

Grb2 Y490 Shc SOS Gab1 Shp-2 Y

785

Ras Raf

PI3k PLCγ IP3 + DAG

GSK3

PKCδ

P* P*

MEK1/2

PDKs

Akt1/2

14-3-3 P*

MAPK1/2

p53

BAD ser136

IKKα

Forkhead P*

BAD Bcl-2

neurite outgrowth

Rsk NFκB

IκB P* NFκB CREB IκB Nucleus

Bcl-xL FasL Bax Bcl-2 IAPs (+) (+) (+) (+) Regulation of survival

FIGURE 17–1 Tyrosine receptor kinase A (TrkA) signal transduction pathways. Nerve growth factor binding induces autophosphorylation of the TrkA receptor. This recruits proteins to the TrkA phosphotyrosine residues that activate a variety of signaling pathways that aid neuron survival and differentiation. These include the phosphatidylinositol 3 kinase (PI3K)–Akt, extracellular signal-regulated kinase (ERK)–Ras, and phospholipase C-␥ 1 (PLC-␥ 1) pathways. Proteins serving an adaptor function are colored green, the small G protein is blue, kinases are purple, and transcription factors are red. (Data from Huang and Reichardt38 and Kaplan and Miller.41) See Color Plate

of these “true” actions is completely understood, it seems reasonable to retain for now the conventional names.

NEUROTROPHIN FAMILIES There are three families of characterized growth factors that promote differentiation and survival in the nervous system. The first is the classic neurotrophins: NGF, BDNF, and neurotrophins 3 through 7 (NT-3 through NT-7). The other two main families are glial cell line–derived neurotrophic factor (GDNF) and CNTF.13 Another family of growth factors, insulin-like growth factors (IGFs), have been traditionally studied as cytokine-like growth factors for non-neuronal cells but also play a role in neuronal development.92 The three-dimensional structures of the classic neurotrophins (NGF, BDNF, NT-3, and NT-4) have been determined through x-ray crystallography. They have

a common core structure of three disulfide bonds that join two pairs of antiparallel two-strand beta sheets (a “cysteine knot” structure). The GDNF family members also have the cysteine knot fold. The CNTF structure is different in that it has no cysteine knot but consists of four helices with several crossover loops.13

Nerve Growth Factor NGF, the first neurotrophin to be characterized, was discovered in the 1950s through the joint work of Viktor Hamburger and Rita Levi-Montalcini.20 NGF was originally studied because of a mouse sarcoma’s unexpected ability to support sympathetic and sensory spinal neurons in vitro. While trying to purify the supportive factor, it was discovered that snake venom and mouse submandibular salivary glands had an even more potent neurotrophic effect.51 The breakthrough that led to the isolation and characterization of the

Neurotrophic Factors in the Peripheral Nervous System

NGF protein was the finding that it was present in large quantities in the tubular portions of adult male mouse submandibular glands.12,51,78 The active form of NGF is a dimer of 13-kDa polypeptide chains that each have three intrachain disulfide bridges.79 The targets of NGF in the peripheral nervous system (PNS) are sympathetic and small-diameter, unmyelinated (C fiber) temperature and pain dorsal root ganglia (DRG) sensory neurons. These neurons express the NGF receptors tyrosine receptor kinase A (TrkA) and p75 low-affinity neurotrophin receptor (p75NTR).27,92 Cells that produce NGF include many targets of neurons during development, such as keratinocytes, melanocytes, vascular and smooth muscle cells, testis and ovarian cells, and cells of the pituitary, thyroid, parathyroid, and exocrine salivary glands.79 In the adult, NGF also plays a role in injury response. Cytokines released by macrophages after peripheral nerve injury cause Schwann cells and fibroblasts to synthesize NGF. Mast cells also synthesize and release NGF.38,52 NGF is part of a neurotrophin family that acts in an autocrine or paracrine fashion on cell surface receptors. The growth factors are present in limited quantities during development, when they function to promote neuronal survival and differentiation.39,51 In addition, neurotrophins have been shown to regulate cell proliferation, axonal and dendritic growth and remodeling, assembly of the cytoskeleton, membrane trafficking and fusion, and synapse formation and function.38 The four neurotrophins characterized in mammals (NGF, BDNF, NT-3, and NT-4)38 promote vertebrate nervous system development.48 Neurotrophin-deficient mutant mice die postnatally, with the exception of those deficient in NT-4.27 Unlike insulin or transforming growth factor-␤, these classic neurotrophins do not have homologues in Caenorhabditis elegans or Drosophila melanogaster.48 Two more neurotrophins, NT-6 and NT-7, have been isolated in fish and are not thought to have mammalian homologues.38 The neurotrophins are made as 30- to 35-kDa precursor proteins that are intracellularly cleaved to 118 to 120 amino acids. The precursor proteins have glycosylation and processing enzyme sites and a signal peptide. The mature proteins are 12- to 14-kDa C-terminal products. They are active as 26-kDa homodimers.16 NGF interacts with two receptors: p140 tyrosine receptor kinase (TrkA or p140TRK) and p75NTR. The TrkA receptor is responsible for the neurotrophic and neuroprotective actions of NGF.30,32 Cells that express the TrkA receptor include sympathetic neurons, DRG neurons, and cholinergic interneurons.45 Tyrosine receptor kinase (Trk) was first discovered as an oncogene in colon carcinoma. The extracellular TrkA kinase domain was fused to tropomyosin, which allows Trk constitutive tyrosine kinase activity. The proto-oncogene, TrkA, was expressed in the nervous system. Later researchers determined that neurotrophins bound to and activated this receptor, which was important in developmental neuronal death and neuronal differentiation.62,67

379

The Trk receptors have been well studied and have five domains: A cysteine-rich cluster constitutes domains one and three, three leucine-rich repeats form domain two, and the last two domains are immunoglobulin-like. The evidence so far supports the involvement of domain five in ligand binding.87 TrkA signaling primarily affects survival and differentiation of small-diameter pain and temperature sensory neurons and sympathetic neurons.62 Mice engineered to be deficient in TrkA do not make thermoceptive or nociceptive neurons.67 The mature form of the neurotrophin NGF interacts with the TrkA receptor with high affinity (KD ⫽ 2.3 ⫻ 10⫺11 M) and the p75NTR receptor with lower affinity (KD ⫽ 1.7 ⫻ 10⫺9 M).80 The immature or precursor protein, which is cleaved intracellularly to the active NGF homodimer, binds to the p75NTR fivefold stronger than mature NGF (10⫺10 M). The ratio of immature to mature NGF influences p75NTR activation. Thus proteolytic cleavage of neurotrophins may be a primary regulator of neuronal survival. The affinity of other immature neurotrophins for p75NTR is still under investigation.16,48 High-affinity binding to endogenous TrkA only occurs with concomitant expression of p75NTR.5,64 The binding of low-concentration NGF to TrkA is regulated by both TrkA and p75NTR levels.47 All of the neurotrophins bind p75NTR. In addition, BDNF and NT-4 bind specifically to tyrosine receptor kinase B (TrkB) and NT-3 specifically to tyrosine receptor kinase C (TrkC).39 NT-3 and adenosine, normally a G protein–coupled receptor agonist, can also activate TrkA and TrkB.46,67 TrkB and TrkC signaling primarily affects proprioceptive and light touch sensory neurons as well as motor neurons.62 p75NTR can prevent or promote cell death.6 Cells that express p75NTR include sympathetic, small-diameter sensory as well as alpha motor neurons and Schwann cells during development and injury.67 About 10 years before transport and endocytosis were studied in non-neuronal cells, the mechanism of NGF internalization and transport was initially defined. NGF is taken into the cell at the axon terminal by a receptordependent mechanism and transported along axons to the cell body in membrane vesicles by microtubule- and energy-dependent mechanisms. Eventually, the transport vesicles fuse with late endosomes at the cell bodies and then to lysosomes, where NGF is degraded.38,64 The neuronal cell body can be up to 1 m from the axon, and the speed of this transport mimics vesicular transport, with estimates ranging from 2 to 20 mm/h in peripheral neurons.57 The transport vesicles use dynein motor proteins to transport retrogradely along axonal microtubules.64 When axons in Campenot chambers are stimulated with neurotrophins, the Trk receptor quickly moves to the cell-body compartment. Some of this Trk is bound to neurotrophin, but there is also evidence that retrogradely transported Trk receptor signals without bound ligand.57 Recently, clathrin-coated vesicles retrogradely transporting NGF

380

Neurobiology of the Peripheral Nervous System

and TrkA were also found to contain activated proteins of the Ras–myelin-associated protein kinase (MAPK) pathway. These endosomes were able to activate a downstream target in the MAPK pathway.37 Autophosphorylated Trk movement to cell bodies is essential for NGF-mediated signal transduction of cyclic AMP–regulated enhancer binding protein (CREB) to neuronal nuclei.70

Brain-Derived Neurotrophic Factor The second neurotrophin, discovered in 1982, was purified from porcine brain and called brain-derived neurotrophic factor. Neurons responsive to BDNF include DRGs, spinal motor neurons, and nodose and trigeminal ganglion sensory neurons. BDNF helps the survival of motor neurons after nerve axotomy and can also prevent developmental motor neuron cell death. After nerve transection, BDNF messenger RNA (mRNA) increases in muscle and TrkB mRNA increases in motor neurons.92 BDNF and NT-3 regulate myelination in the PNS.15 BDNF can rescue about a third of postnatal DRG neurons in culture. Likewise, the BDNF null mutant mouse loses about a third of its DRG neurons after birth.27 BDNF may also have a role in memory and learning, because long-term potentiation is impaired in the mutant mouse.89 Neurotrophins are produced in target cells for presynaptic neurons. They are then retrogradely transported to the neuronal cell body. There is recent evidence that BDNF can act as an anterograde as well as a retrograde trophic factor. It is released at brain neuron synapse terminals and can also act on postsynaptic neurons.63 BDNF has been used intrathecally in Phase I/II65 and subcutaneously in Phase I through III clinical trials for amyotrophic lateral sclerosis (ALS)82 and in Phase II trials for diabetic polyneuropathy86,92 with no benefit demonstrated.

Neurotrophin-3/Neurotrophin-4 NT-3 is mainly expressed in muscle spindles, Merkel cells, and Golgi tendon organs. There is evidence that it aids differentiation of sympathetic cholinergic neurons and that its effects are antagonized by NGF.10 NT-3 mutant mice lose about 60% of their DRG neurons during embryogenesis between embryonic day (E) 10.5 and E13, before NGF or BDNF dependence has begun. Mice engineered to be deficient in the NT-3 receptor TrkC do not make proprioceptive neurons.67 When TrkC is experimentally removed, there is evidence that NT-3 can signal through the TrkA and TrkB receptors. NT-4 (also known as NT-5) regulates low-threshold hair afferent mechanoreceptors.27

Ciliary Neurotrophic Factor CNTF was first purified in 1979 from chick nerve and eye extracts that promoted survival of ciliary ganglionic neurons. It also supports dopaminergic, retinal rod, sympathetic, and

motor neurons. It may play a role in body weight homeostasis and PNS injury response.13,50,64 CNTF is a nonsecreted ␣-helical cytokine that first binds to ciliary neurotrophic factor receptor ␣ (CNTFR␣), followed by two other receptor subunits: gp130 and leukemia inhibitory factor receptor ␤ (LIFR␤). CNTFR␣ is tethered to the cell membrane with a glycosyl phosphatidylinositol anchor. It confers binding specificity. Gp130 is a nonspecific, high-affinity, signaltransducing subunit. The LIFR␤ and gp130 intracellular domains can bind STAT or JAK.13,50,64 A secreted ligand for CNTFR␣ was discovered in 2000: cardiotrophin-like cytokine (CLC), also identified as novel neurotrophin-1. CLC binds to the CNTF receptor after forming a complex with another secreted protein, cytokine-like factor-1 (CLF-1). CLF-1 was originally cloned as a cytokine receptor but functions in this case as a ligand.50

Glial Cell Line–Derived Neurotrophic Factor GDNF was first purified in 1993 from a glial cell line culture that promoted neuronal survival. It can support embryonic midbrain dopaminergic, sympathetic, and spinal motor neurons. It is required for development of the enteric nervous system, ureters, and kidneys.13,26,59,68,73,76 The GDNF family includes three other trophic factors: persephin (PSP), artemin (ART), and neurturin (NTN). They share a common tyrosine kinase receptor, c-Ret. In addition, they each have a receptor that allows ligand specificity. These receptors are named GDNF family receptor ␣ (GFR␣)-1 through -4 and are tethered to the cell membrane with a glycosyl phosphatidylinositol anchor. GDNF␣-1 preferentially binds GDNF, GDNF␣-2 preferentially binds NTN, GDNF␣-3 preferentially binds ART, and GDNF␣-4 preferentially binds PSP. Some GDNF family members can bind more than one “specific” receptor.64 This is similar to the crossover binding seen between classic neurotrophic factors and Trk receptors. The active GDNF family receptor is a heterohexameric complex made up of two Ret receptors and two GDNF ␣ receptors. It then signals through the phosphatidylinositol 3 kinase (PI3K) and MAPK pathways.13

Insulin-like Growth Factor The structure of IGF is about 50% homologous to insulin. The tyrosine kinase IGF-I receptor is also homologous to the insulin receptor and is expressed throughout the nervous system. The IGFs promote survival and neuronal outgrowth in developing sympathetic and sensory neurons. IGF-I promotes motor neuron survival of developing neurons after axotomy. IGF-I injection can also induce motor neuron axonal growth in adult mammals, and antiserum to IGF-I and IGF-II inhibits regeneration of motor axons after spinal root injury. IGF-I has been used in two Phase III clinical trials for ALS. The first showed a modest benefit and the second showed no benefit. A meta-analysis suggested a modest

Neurotrophic Factors in the Peripheral Nervous System

but unproven beneficial effect for IGF-I in the treatment of ALS.58 It has also been used in Phase II trials for diabetic polyneuropathy, chemotherapy-induced neuropathy, and postpolio syndrome.92 IGF-I is a potential therapeutic modality to promote neuronal regeneration.

NEUROTROPHIN SIGNALING The extracellular TrkA-d5 immunoglobulin-like domain proximal to the cell membrane is both necessary and sufficient for the binding of NGF to TrkA.88 After binding, the tyrosine kinase in the receptor cytoplasmic domain is phosphorylated and activated. The Trk receptors contain 10 conserved cytoplasmic domain tyrosines. Three of these (Y670, Y674, and Y675) form an autoregulatory loop in the receptor kinase domain that controls kinase activity. When the other tyrosine residues are phosphorylated, adapter proteins with Src homology 2 or phosphotyrosine-binding motifs can bind and activate different intracellular signaling cascades. The intracellular signaling cascades that are activated include PI3K-Akt, extracellular signal-regulated kinase (ERK)-Ras, and phospholipase C-␥ 1 (PLC-␥ 1)38 (see Fig. 17–1). PLC-␥ 1 is recruited to the membrane and phosphorylated by the TrkA kinase after phosphorylation of TrkA Y785. PLC-␥ 1 hydrolyses phosphatidylinositides to form inositol triphosphate and diacylglycerol (DAG). Inositol triphosphate stimulates release of Ca2⫹ into the cytoplasm from the endoplasmic reticulum. The cytoplasmic calcium then activates Ca2⫹-regulated isoforms of protein kinase C (PKC) and Ca2⫹-calmodulin–regulated protein kinases and phosphatases. DAG can activate DAG-regulated PKC isoforms.38,83 One DAG-dependent, Ca2⫹-independent isoform of PKC is PKC␦. PKC␦ activity is required in PC12 cells for neurite outgrowth and activation of ERK in response to NGF.19 Phosphorylation of the TrkA receptor on Y490 activates the Ras-MAPK pathway. The adaptor protein Shc is recruited to the TrkA receptor, and this allows subsequent binding of the adaptor protein Grb-2 and Ras exchange factor SOS. SOS, when recruited to the cell membrane, can activate Ras by allowing exchange of GDP for GTP. This activates the c-Raf–ERK and PI3K pathways, which are the main regulators of Ras- and neurotrophin-mediated survival.41,54 There are many important signaling targets of the serine-threonine kinases c-Raf and ERK, including ribosomal S6 kinase. Activation of these kinases can lead to the phosphorylation of CREB, a transcription factor that promotes neuron survival. Inhibition of CREB phosphorylation can trigger apoptosis.9 Upon activation, this transcription factor binds to a cyclic AMP response element inside the c-fos promoter, allowing transcription of genes needed for initiation and maintenance of differentiation.7,79

381

Ras activation of PI3K is mediated through the adaptor protein Grb2-associated binder 1 (Gab-1). The phosphatidylinositides formed by the PI3K activate the kinases PDK1 and PDK2, which in turn phosphorylate and activate the protein kinase Akt (protein kinase B ␣).41 Akt is a serine-threonine kinase that promotes cell survival. It acts through downstream effectors such as Bad, caspase-9, forkhead, nuclear factor-␬B (NF-␬B), and Pak.11,14,21,23,40,69,81 The phosphorylation of Bad causes association with 14-3-3 proteins that inhibit its pro-apoptotic function.21 Akt also phosphorylates the inhibitor of I␬B kinase (IKK␣) at threonine 23. NF-␬B is usually sequestered in the cytoplasm bound to the I␬B inhibitor protein. The phosphorylation of IKK␣ by both Akt and NF-␬B–inducing kinase (NIK) allows subsequent degradation of the I␬B complex (IKK␣ and IKK␤). NF-␬B can then translocate to the nucleus, where it acts as a survival-promoting transcription factor.24,66,71,91 Another Akt substrate, forkhead, can control the expression of genes involved in apoptosis, such as Fas ligand.11 The IAP caspase inhibitor family is also potentially activated by Akt.41 There is in vivo evidence that Akt is necessary for the neuroprotective effects of IGF-I.43 Mutations that activate Akt have been found in certain tumors.28 Activated Akt has also been shown to inhibit both apoptosis and changes in Bax expression induced by nitric oxide in primary hippocampal neurons.53 It also has the ability to prevent cytochrome c release from the mitochondria42 and delay the onset of p53-mediated apoptosis.72 p75NTR binds to all of the mature neurotrophins with a similar affinity. Receptor activation can both prevent and promote cell death6 as well as regulate axon growth and mediate neurotrophin retrograde transport.67 Neurotrophin activation of NF-␬B through p75NTR can promote neuronal survival.33 NGF-mediated activation of p75NTR allows association of a tumor necrosis factor receptor–associated factor (TRAF) adaptor protein.44 TRAF2, TRAF5, and TRAF6 activate NF-␬B.4,44 NIK is part of the signaling complex assembled with TRAF proteins after they are recruited to a surface receptor. NF-␬B is activated when NIK phosphorylates IKK. IKK then phosphorylates I␬B, which is degraded by proteases after ubiquination. This releases NF-␬B to translocate to the nucleus, where it can bind and activate promoters of responsive genes4 (Fig. 17–2). Activation of the p75NTR has been associated with apoptosis of sensory and motor neurons, sympathetic neurons, hippocampal neurons, and oligodendrocytes.56 p75NTR can cause apoptosis via activation of the Jun N-terminal kinase (JNK)–mediated signaling pathway, with subsequent activation of p53.2 Activation of the JNK pathway leads to Fas ligand upregulation.49 p53 can bind and transactivate the fas gene at a responsive element located within the first intron.61 There is evidence that neuronal death from neurotrophin withdrawal depends on the presence of p75NTR,6 p53,22,84 and Fas receptor.55 Thus death from NGF withdrawal probably

382

Neurobiology of the Peripheral Nervous System

p75NTR Cell membrane TRAF6

Traf 2/4

Ras TrkA RhoA

SC-1

NRIF NIK PI3K

Ceramide

NRAGE

P*

P* IκB P* p53 IκB

IKKα

Bax

NFκB

Cell death

Jun kinase pathway Nucleus NFκB

Apoptosis Cell survival

FIGURE 17–2 p75NTR signal transduction pathways. Neurotrophin binding causes the association of adaptor proteins that mediate prosurvival and prodeath pathways. Activation of the pro-apoptotic JNK pathway can be inhibited by concurrent TrkA activation. Activation of NF-␬B through the association of TRAF proteins leads to cell survival. Proteins serving an adaptor function are colored green, small G proteins are blue, kinases are purple, and transcription factors are red. (Data from Huang and Reichardt38 and Kaplan and Miller.41) See Color Plate

occurs through both a JNK-p53-Bax and a JNK-p53-Fas pathway. In the presence of NGF and TrkA signaling, the pro-apoptotic p75NTR pathways are suppressed while the prosurvival pathways are activated.41 If NGF is present in cells without TrkA (such as oligodendrocytes), then p75NTR activation will signal for apoptosis. If BDNF is present in cells without TrkB (such as sympathetic neurons), then p75NTR activation will signal for apoptosis. Thus Trk receptor expression regulates the pro- or antiapoptotic potential of p75NTR. Trk may inhibit p75NTR-mediated JNK activation via Ras and the PI3K-Akt pathway.56 p75NTR activation by NGF, in the absence of a TrkA receptor, induces sphingomyelin hydrolysis, resulting in the formation of the sphingolipid metabolite ceramide.25 Ceramide binds to Raf-1 and promotes its association with GTP-Ras without subsequent Raf-1 kinase activation. This inhibits the survival-promoting MAPK/ERK cascade.60 Ceramide also inhibits PI3K signaling.93,94 There are other proteins that interact with the p75NTR receptor, such as neurotrophin receptor interacting factor, neurotrophin

receptor-interacting MAGE (melanoma-associated antigen) homologue, and Schwann cell-1. These proteins may be involved in ceramide production or JNK activation.38 Activation of p75NTR may be partially responsible for injury-induced apoptosis, because p75NTR⫺/⫺ animals show less Schwann cell apoptosis after sciatic nerve axotomy. p75NTR is also induced in central nervous system neurons dying from excitotoxicity.56 NGF protects neurons from death resulting from various insults, including axotomy, ischemia, oxidative stress, and glutamate receptor–mediated excitotoxicity. NGF can also protect PC12 cells and DRG neurons from anoxia and glucose deprivation.8,35,36,79 The addition of NGF rescues PC12 cells from death after treatment with 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP).77 Pretreatment with NGF can prevent nitric oxide cytotoxicity in PC12 cells85 as well as excitotoxic injury in vivo.1 NGF also protects sensory neurons from death after treatment with chemotherapeutic drugs such as cisplatin. Cisplatin causes apoptosis of DRG neurons.29 High-dose NGF protects against cisplatin-mediated apoptotic DRG

383

Neurotrophic Factors in the Peripheral Nervous System

death.31 NGF may protect neurons from cisplatin toxicity through NGF-mediated activation of Akt kinase and inhibition of Bax translocation. Cisplatin causes Bax translocation and subsequent death in DRG neurons. High-dose NGF prevents this translocation.55 DRG neurons that lack the gene for Bax have very little death when exposed to lethal doses of cisplatin. Activated Akt inhibits apoptosis and changes in Bax expression induced by nitric oxide in primary hippocampal neurons.53 Apoptosis of primary hippocampal neurons as a result of nitric oxide or hypoxia is dependent on the pro-apoptotic effects of p53. Exogenous Akt kinase suppresses p53-mediated transcriptional activation of Bax, caspase-3 activity, and cell death.90 Akt also has the ability to prevent cytochrome c release from the mitochondria42 and delay the onset of temperature-dependent p53-mediated apoptosis.72 There is in vitro and in vivo evidence that Akt is necessary for the neuroprotective effects of IGF-I.18,43 Akt may be necessary for the neuroprotective effects of NGF. The rescue of PC12 cells from death resulting from treatment with MPTP is prevented by co-treatment with a dominant negative form of Akt.77 The TrkA receptor inhibits p75NTR-mediated activation of the pro-apoptotic JNK pathway.41 Cisplatin can also activate the JNK pathway. Inhibition of this pathway decreases cisplatin-mediated cellular death.74,75 Thus TrkA activation could also be inhibiting cisplatin-mediated JNK activation and subsequent cell death.

CLINICAL APPLICATIONS OF NEUROTROPHIC FACTORS Neurotrophic factors have been successfully used to treat a variety of animal disease states, including motor neuron disease, toxic neuropathy, diabetic neuropathy, spinal cord and nerve trauma, stroke, and models of Alzheimer’s and Parkinson’s diseases. Unfortunately, when used in human trials, there have been few successes. NGF infusion has been tested for possible treatment of human immunodeficiency virus–associated neuropathy, diabetic neuropathy, and Alzheimer’s disease. Attempts to treat ALS have included clinical trials with BDNF, CNTF, and IGF-I.3 NGF, BDNF, NT-3, and IGF-I have all been tested in Phase II or III human clinical trials for various neurologic diseases. Adverse reactions have prevented GDNF from moving beyond Phase I clinical trials.92 Thus far, no Phase III trial of growth factor treatment has led to a new standard in disease therapy. Many of these trials have used systemically administered growth factors, which may not have access to neurons protected by a blood-nerve or blood-brain barrier. In addition, intermittent-bolus dosing has been used instead of more physiologic steady exposure. Systemic side effects have limited dosage below the amounts used in the successful animal trials even when effective delivery (such as to

sensory neurons exposed to systemic circulation) was possible.3 Thus it may be early to conclude that growth factors will never be used therapeutically. Targeted drug delivery and more physiologic dosing, which would also mitigate systemic reaction, may allow successful use of growth factors in the treatment of neurologic disease. Recently, an advance has been made with successful NT-3 gene transfer using herpes simplex virus. Endogenous herpes simplex virus targets peripheral sensory neurons, where it establishes a latent infection. Expression of NT-3 from a replication-incompetent viral vector protected DRG neurons from pyridoxine-mediated neuropathy.17 Viral vectors have also been used to successfully treat animal models of Parkinson’s disease, Huntington’s disease, and motor neuron disease.3 Thus viral vectors are one possibility to allow administration of neurotrophins to treat neurologic disease in both the peripheral and central nervous systems.

REFERENCES 1. Aloe, L.: Intracerebral pretreatment with nerve growth factor prevents irreversible brain lesions in neonatal rats injected with ibotenic acid. Biotechnology 5:1085, 1987. 2. Aloyz, R. S., Bamji, S. X., Pozniak, C. D., et al.: p53 is essential for developmental neuron death as regulated by the TrkA and p75 neurotrophin receptors. J. Cell Biol. 143:1691, 1998. 3. Apfel, S. C.: Is the therapeutic application of neurotrophic factors dead? Ann. Neurol. 51:8, 2002. 4. Arch, R. H., Gedrich, R. W., and Thompson, C. B.: Tumor necrosis factor receptor-associated factors (TRAFs)—a family of adapter proteins that regulates life and death. Genes Dev. 12:2821, 1998. 5. Barker, P. A., and Shooter, E. M.: Disruption of NGF binding to the low affinity neurotrophin receptor p75LNTR reduces NGF binding to TrkA on PC12 cells. Neuron. 13:203, 1994. 6. Barrett, G. L., and Bartlett, P. F.: The p75 nerve growth factor receptor mediates survival or death depending on the stage of sensory neuron development. Proc. Natl. Acad. Sci. U. S. A. 91:6501, 1994. 7. Berkowitz, L. A., Riabowol, K. T., and Gilman, M. Z.: Multiple sequence elements of a single functional class are required for cyclic AMP responsiveness of the mouse c-fos promoter. Mol. Cell. Biol. 9:4272, 1989. 8. Boniece, I. R., and Wagner, J. A.: Growth factors protect PC12 cells against ischemia by a mechanism that is independent of PKA, PKC, and protein synthesis. J. Neurosci. 13:4220, 1993. 9. Bonni, A., Brunet, A., West, A. E., et al.: Cell survival promoted by the Ras-MAPK signaling pathway by transcriptiondependent and -independent mechanisms. Science 286:1358, 1999. 10. Brodski, C., Schnurch, H., and Dechant, G.: Neurotrophin-3 promotes the cholinergic differentiation of sympathetic neurons. Proc. Natl. Acad. Sci. U. S. A. 97:9683, 2000. 11. Brunet, A., Bonni, A., Zigmond, M. J., et al.: Akt promotes cell survival by phosphorylating and inhibiting a forkhead transcription factor. Cell 96:857, 1999.

384

Neurobiology of the Peripheral Nervous System

12. Burdman, J. A., and Goldstein, M. N.: Synthesis and storage of a nerve growth protein in mouse submandibular glands. J. Exp. Zool. 160:183, 1965. 13. Butte, M. J.: Neurotrophic factor structures reveal clues to evolution, binding, specificity, and receptor activation. Cell. Mol. Life Sci. 58:1003, 2001. 14. Cardone, M. H., Roy, N., Stennicke, H. R., et al.: Regulation of cell death protease caspase-9 by phosphorylation. Science 282:1318, 1998. 15. Chan, J. R., Cosgaya, J. M., Wu, Y. J., and Shooter, E. M.: Neurotrophins are key mediators of the myelination program in the peripheral nervous system. Proc. Natl. Acad. Sci. U. S. A. 98:14661, 2001. 16. Chao, M. V., and Bothwell, M.: To cleave or not to cleave. Neuron. 33:9–12, 2001. 17. Chattopadhyay, M., Wolfe, D., Huang, S., et al.: In vivo gene therapy for pyridoxine-induced neuropathy by herpes simplex virus-mediated gene transfer of neurotrophin-3. Ann. Neurol. 51:19, 2002. 18. Cheng, H. L., Steinway, M., Delaney, C. L., et al.: IGF-I promotes Schwann cell motility and survival via activation of Akt. Mol. Cell. Endocrinol. 170:211, 2000. 19. Corbit, K. C., Foster, D. A., and Rosner, M. R.: Protein kinase C␦ mediates neurogenic but not mitogenic activation of mitogen-activated protein kinase in neuronal cells. Mol. Cell. Biol. 19:4209, 1999. 20. Cowan, W. M.: Viktor Hamburger and Rita Levi-Montalcini: the path to the discovery of nerve growth factor. Annu. Rev. Neurosci. 24:551, 2001. 21. Datta, S. R., Dudek, H., Tao, X., et al.: Akt phosphorylation of BAD couples survival signals to the cell-intrinsic death machinery. Cell 91:231, 1997. 22. Deckwerth, T. L., Elliott, J. L., Knudson, C. M., et al.: BAX is required for neuronal death after trophic factor deprivation and during development. Neuron 17:401, 1996. 23. del Peso, L., Gonzalez-Garcia, M., Page, C., et al.: Interleukin-3-induced phosphorylation of BAD through the protein kinase Akt. Science 278:687, 1997. 24. Demarchi, F., Verardo, R., Varnum B., et al.: Gas6 antiapoptotic signaling requires NF-kappa B activation. J. Biol. Chem. 276:31738, 2001. 25. Dobrowsky, R. T., Jenkins, G. M., and Hannun, Y. A.: Neurotrophins induce sphingomyelin hydrolysis: modulation by co-expression of p75NTR with Trk receptors. J. Biol. Chem. 270:22135, 1995. 26. Enomoto, H., Araki, T., Jackman, A., et al.: GFR ␣1-deficient mice have deficits in the enteric nervous system and kidneys. Neuron 21:317, 1998. 27. Ernfors, P.: Local and target-derived actions of neurotrophins during peripheral nervous system development. Cell. Mol. Life Sci. 58:1036, 2001. 28. Evan, G. I., and Vousden, K. H.: Proliferation, cell cycle and apoptosis in cancer. Nature 411:342, 2001. 29. Fischer, S. J., McDonald, E. S., Gross, L., and Windebank, A. J.: Alterations in cell cycle regulation underlie cisplatin induced apoptosis of dorsal root ganglion neurons in vivo. Neurobiol. Dis. 8:1027, 2001. 30. Fischer, S. J., Podratz, J. L., and Windebank, A. J.: Nerve growth factor rescue of cisplatin neurotoxicity is mediated

31.

32.

33.

34.

35.

36.

37.

38.

39.

40. 41.

42.

43.

44.

45. 46.

through the high affinity receptor: studies in PC12 cells and p75 null mouse dorsal root ganglia. Neurosci. Lett. 308:1, 2001. Gill, J. S., and Windebank, A. J.: Cisplatin-induced apoptosis in rat dorsal root ganglion neurons is associated with attempted entry into the cell cycle. J. Clin. Invest. 101:2842, 1998. Greene, S. H., Rydel, R. E., Connolly, J. L., and Greene, L. A.: PC12 cell mutants that possess low- but not highaffinity nerve growth factor receptors neither respond to nor internalize nerve growth factor. J. Cell Biol. 102:830, 1986. Hamanoue, M., Middleton, G., Wyatt, S., et al.: p75-mediated NF-kappaB activation enhances the survival response of developing sensory neurons to nerve growth factor. Mol. Cell. Neurosci. 14:28, 1999. Hamburger, V., and Keefe, E. L.: The effects of peripheral factors on the proliferation and differentiation in the spinal cord of chick embryos. J. Exp. Zool. 96:223, 1944. Honma, H., Gross, L., and Windebank, A. J.: Hypoxia-induced apoptosis of dorsal root ganglion neurons is associated with DNA damage recognition and cell cycle disruption. Neurosci. Lett. 354:95, 2004. Honma, H., Podratz, J. L., and Windebank, A. J.: Acute glucose deprivation leads to apoptosis in a cell model of acute diabetic neuropathy. J. Peripher. Nerv. Syst. 8:65, 2003. Howe, C. L., Valletta, J. S., Rusnak, A. S., and Mobley, W. C.: NGF signaling from clathrin-coated vesicles: evidence that signaling endosomes serve as a platform for the Ras-MAPK pathway. Neuron 32:801, 2001. Huang, E. J., and Reichardt, L. F.: Neurotrophins: roles in neuronal development and function. Annu. Rev. Neurosci. 24:677, 2001. Johnson, J. E.: Neurotrophic factors. In Zigmund, M. J., Bloom, R. E., Landis, S. C., et al. (eds.): Fundamental Neuroscience. San Diego, Academic Press, p. 611, 1999. Kane, L. P., Shapiro, V. S., Stokoe, D., and Weiss, A.: Induction of NF-␬B by the Akt/PKB kinase. Curr. Biol. 9:601, 1999. Kaplan, D. R., and Miller, F. D.: Neurotrophin signal transduction in the nervous system. Curr. Opin. Neurobiol. 10:381, 2000. Kennedy, S. G., Kandel, E. S., Cross, T. K., and Hay, N.: Akt/protein kinase B inhibits cell death by preventing the release of cytochrome c from mitochondria. Mol. Cell. Biol. 19:5800, 1999. Kermer, P., Klöcker, N., Labes, M., and Bähr, M.: Insulin-like growth factor-I protects axotomized rat retinal ganglion cells from secondary death via PI3-K-dependent Akt phosphorylation and inhibition of caspase-3 in vivo. J. Neurosci. 20:722, 2000. Khursigara, G., Orlinick, J. R., and Chao, M. V.: Association of the p75 neurotrophin receptor with TRAF6. J. Biol. Chem. 274:2597, 1999. Korsching, S.: The neurotrophic factor concept: a reexamination. J. Neurosci. 13:2739, 1993. Le-Niculescu, H., Bonfoco, E., Kasuya, Y., et al.: Withdrawal of survival factors results in activation of the JNK pathway in neuronal cells leading to Fas ligand induction and cell death. Mol. Cell. Biol. 19:751, 1999.

Neurotrophic Factors in the Peripheral Nervous System 47. Lee, F. S., and Chao, M. V.: Activation of Trk neurotrophin receptors in the absence of neurotrophins. Proc. Natl. Acad. Sci. U. S. A. 98:3555, 2001. 48. Lee, F. S., Kim, A. H., Khursigara, G., and Chao, M. V.: The uniqueness of being a neurotrophin receptor. Curr. Opin. Neurobiol. 11:281, 2001. 49. Lee, R., Kermani, P., Teng, K. K., and Hempstead, B. L.: Regulation of cell survival by secreted proneurotrophins. Science 294:1945, 2001. 50. Lesser, S. S., and Lo, D. C.: CNTF II, I presume? Nat. Neurosci. 3:851, 2000. 51. Levi-Montalcini, R.: The nerve growth factor thirty-five years later. In Vitro Cell. Dev. Biol. 23:227, 1987. 52. Levi-Montalcini, R., Skaper, S. D., Dal Toso, R., et al.: Nerve growth factor: from neurotrophin to neurokine. Trends Neurosci. 19:514, 1996. 53. Matsuzaki, H., Tamatani, M., Mitsuda, N., et al.: Activation of Akt kinase inhibits apoptosis and changes in Bcl-2 and Bax expression induced by nitric oxide in primary hippocampal neurons. J. Neurochem. 73:2037, 1999. 54. McCormick, F.: Activators and effectors of ras p21 proteins. Curr. Opin. Genet. Dev. 4:71, 1994. 55. McDonald, E. S., and Windebank, A. J.: Cisplatin-induced apoptosis of DRG neurons involves Bax redistribution and cytochrome c release but not fas receptor signaling. Neurobiol. Dis. 9:220, 2002. 56. Miller, F. D., and Kaplan, D. R.: Neurotrophin signalling pathways regulating neuronal apoptosis. Cell. Mol. Life Sci. 58:1045, 2001. 57. Miller, F. D., and Kaplan, D. R.: On Trk for retrograde signaling. Neuron 32:767, 2001. 58. Mitchell, J. D., Wokke, J. H., and Borasio, G. D.: Recombinant human insulin-like growth factor I (rhIGF-I) for amyotrophic lateral sclerosis/motor neuron disease. Cochrane Database Syst. Rev. 3:CD002064, 2002. 59. Moore, M. W., Klein, R. D., Farinas, I., et al.: Renal and neuronal abnormalities in mice lacking GDNF. Nature 382:76, 1996. 60. Muller, G., Storz, P., Bourteele, S., et al.: Regulation of Raf-1 kinase by TNF via its second messenger ceramide and cross-talk with mitogenic signalling. EMBO J. 17:732, 1998. 61. Muller, M., Wilder, S., Bannasch, D., et al.: p53 activates the CD95 (APO-1/Fas) gene in response to DNA damage by anticancer drugs. J. Exp. Med. 188:2033, 1998. 62. Nakagawara, A.: Trk receptor tyrosine kinases: a bridge between cancer and neural development. Cancer Lett. 169:107, 2001. 63. Nawa, H., and Takei, N.: BDNF as an anterophin: a novel neurotrophic relationship between brain neurons. Trends Neurosci. 24:683, 2001. 64. Neet, K. E., and Campenot, R. B.: Receptor binding, internalization, and retrograde transport of neurotrophic factors. Cell. Mol. Life Sci. 58:1021, 2001. 65. Ochs, G., Penn, R. D., York, M., et al.: A Phase I/II trial of recombinant methionyl human brain derived neurotrophic factor administered by intrathecal infusion to patients with amyotrophic lateral sclerosis. Amyotroph. Lateral Scler. Other Motor Neuron Disord. 1:201, 2000.

385

66. Ozes, O. N., Mayo, L. D., Gustin, J. A., et al.: NF-␬B activation by tumour necrosis factor requires the Akt serine-threonine kinase. Nature 401:82, 1999. 67. Patapoutian, A., and Reichardt, L. F.: Trk receptors: mediators of neurotrophin action. Curr. Opin. Neurobiol. 11:272, 2001. 68. Pichel, J. G., Shen, L., Sheng, H. Z., et al.: Defects in enteric innervation and kidney development in mice lacking GDNF. Nature 382:73, 1996. 69. Rena, G., Guo, S., Cichy, S. C., et al.: Phosphorylation of the transcription factor forkhead family member FKHR by protein kinase B. J. Biol. Chem. 274:17179, 1999. 70. Riccio, A., Pierchala, B. A., Ciarallo, C. L., and Ginty, D. D.: An NGF-TrkA-mediated retrograde signal to transcription factor CREB in sympathetic neurons. Science 277:1097, 1997. 71. Romashkova, J. A., and Makarov, S. S.: NF-␬B is a target of AKT in anti-apoptotic PDGF signalling. Nature 401:86, 1999. 72. Sabatini, P., and McCormick, F.: Phosphoinositide 3-OH kinase (PI3K) and PKB/Akt delay the onset of p53-mediated, transcriptionally dependent apoptosis. J. Biol. Chem. 274:24263, 1999. 73. Sanchez, M. P., Silos-Santiago, I., Frisen, J., et al.: Renal agenesis and the absence of enteric neurons in mice lacking GDNF. Nature 382:70, 1996. 74. Sanchez-Perez, I., Martinez-Gomariz, M., Williams, D., et al.: CL100/MKP-1 modulates JNK activation and apoptosis in response to cisplatin. Oncogene 19:5142, 2000. 75. Sanchez-Perez, I., Murguia, J. R., and Perona, R.: Cisplatin induces a persistent activation of JNK that is related to cell death. Oncogene 16:533, 1998. 76. Schuchardt, A., D’Agati, V., Larsson-Blomberg, L., et al.: Defects in the kidney and enteric nervous system of mice lacking the tyrosine kinase receptor Ret. Nature 367:380, 1994. 77. Shimoke, K., and Chiba, H.: Nerve growth factor prevents 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-induced cell death via the Akt pathway by suppressing caspase-3-like activity using PC12 cells: relevance to therapeutical application for Parkinson’s disease. J. Neurosci. Res. 63:402, 2001. 78. Shooter, E. M.: Early days of the nerve growth factor proteins. Annu. Rev. Neurosci. 24:601, 2001. 79. Sofroniew, M. V., Howe, C. L., and Mobley, W. C.: Nerve growth factor signaling, neuroprotection, and neural repair. Annu. Rev. Neurosci. 24:1217, 2001. 80. Sutter, A., Riopelle, R. J., Harris-Warrick, R. M., and Shooter, E. M.: Nerve growth factor receptors: characterization of two distinct classes of binding sites on chick embryo sensory ganglia cells. J. Biol. Chem. 254:5972, 1979. 81. Tang, Y., Zhou, H., Chen, A., et al.: The Akt proto-oncogene links ras to pak and cell survival signals. J. Biol. Chem. 275:9106, 2000. 82. The BDNF Study Group (Phase III): A controlled trial of recombinant methionyl human BDNF in ALS. Neurology 52:1427, 1999. 83. Vetter, M. L., Martin-Zanca, D., Parada, L. F., et al.: Nerve growth factor rapidly stimulates tyrosine phosphorylation of phospholipase C-g1 by a kinase activity associated with the product of the trk protooncogene. Proc. Natl. Acad. Sci. U. S. A. 88:5650, 1991.

386

Neurobiology of the Peripheral Nervous System

84. Vogel, K. S., and Parada, L. F.: Sympathetic neuron survival and proliferation are prolonged by loss of p53 and neurofibromin. Mol. Cell. Neurosci. 11:19, 1998. 85. Wada, K., Okada, N., Yamamura, T., and Koizumi, S.: Nerve growth factor induces resistance of PC12 cells to nitric oxide cytotoxicity. Neurochem. Int. 29:461, 1996. 86. Wellmer, A., Misra, V. P., Sharief, M. K., et al.: A doubleblind placebo-controlled clinical trial of recombinant human brain-derived neurotrophic factor (rhBDNF) in diabetic polyneuropathy. J. Peripher. Nerv. Syst. 6:204, 2001. 87. Wiesmann, C., and de Vos, A. M.: Nerve growth factor: structure and function. Cell. Mol. Life Sci. 58:748, 2001. 88. Wiesmann, C., Ultsch, M. H., Bass, S. H., and de Vos, A. M.: Crystal structure of nerve growth factor in complex with the ligand-binding domain of the TrkA receptor. Nature 401:184, 1999. 89. Yamada, K., Mizuno, M., and Nabeshima, T.: Role for brainderived neurotrophic factor in learning and memory. Life Sci. 70:735, 2002.

90. Yamaguchi, A., Tamatani, M., Matsuzaki, H., et al.: Akt activation protects hippocampal neurons from apoptosis by inhibiting transcriptional activity of p53. J. Biol. Chem. 276:5256, 2001. 91. Yang, C. H., Murti, A., Pfeffer, S. R., et al.: Interferon ␣/␤ promotes cell survival by activating nuclear factor ␬B through phosphatidylinositol 3-kinase and Akt. J. Biol. Chem. 276:13756, 2001. 92. Yuen, E. C.: The role of neurotrophic factors in disorders of peripheral nerves and motor neurons. Phys. Med. Rehabil. Clin. N. Am. 12:293, 2001. 93. Zhou, H., Summers, S. A., Birnbaum, M. J., and Pittman, R. N.: Inhibition of Akt kinase by cell-permeable ceramide and its implications for ceramide-induced apoptosis. J. Biol. Chem. 273:16568, 1998. 94. Zundel, W., Swiersz, L. M., and Giaccia, A.: Caveolin 1-mediated regulation of receptor tyrosine kinase-associated phosphatidylinositol 3-kinase activity by ceramide. Mol. Cell. Biol. 20:1507, 2000.

18 Axonal Transport: Properties, Mechanisms, and Role in Nerve Disease STEPHEN BRIMIJOIN

Properties and Mechanisms Overview Fast Anterograde Transport Fast Retrograde Transport Slow Transport Transport Motors Theoretical Implications of Transport Abnormalities Loading Stage Defects Anterograde Transport Defects Turnaround Defects

Defects of Retrograde Axonal Transport Potential Causes of Transport Defects Evidence for Transport Abnormalities in Human Neuropathy Axonal Transport in Experimental Neuropathies Overview of Toxicant-Induced Neuropathies Experimental Diabetes Acrylamide

PROPERTIES AND MECHANISMS Overview Neurons, because of their extended geometry and inability to synthesize proteins in axons, depend on special pathways for transport of macromolecules to their distal extremities. Waller’s original 1850 study of nerve transection233 identified the neuronal cell body as a “trophic center.” Fifty years later, Scott209 proposed that the trophic relationship between center and periphery involved actual nutrient flow the interruption of which was responsible for the degeneration Waller described. Though that seems self-evident today, Ranvier’s discussion of the mechanism of wallerian degeneration183 referred only to “vital energy” supplied by cell bodies, and Ramon y Cajal explicitly rejected the idea that soluble substances or enzymes move along axons.182 Axonal transport is now seen as a system of intracellular motility enabling nerve cells to deliver essential proteins and membrane components to peripheral sites, and to receive from these locations chemical signals and materials for disposal. Axonal transport is required to preserve cell viability, to elaborate and support axonal and dendritic arborization, to sustain impulse conduction, and to maintain the release

␤,␤⬘-Iminodipropionitrile Hexacarbons Zinc Pyridinethione p-Bromophenylacteylurea Genetically Based Animal Neuropathies Murine Dystrophy Related Dystrophies Other Genetically Based, Transgenic, and Cell Culture Models Conclusion

of transmitters at nerve endings. In addition, axonal transport provides a two-way line of chemical communication between neurons, on the one hand, and their synaptic targets and supporting cells, on the other. Anterograde transport delivers neurotrophic components to the nerve terminals, where their release affects muscles and other cells, and retrograde transport supplies the cell bodies with neurotrophic substances captured from the synaptic environment. The first part of this chapter outlines the basic phenomena of axonal transport, described in more detail in classic monographs by Grafstein and Forman87 and Sidney Ochs,168 and introduces recent advances in our understanding of the mechanisms involved. Axonal transport is a ubiquitous, ATP-driven process with multiple “components,” each moving at a distinctive macroscopic velocity. 239,240 The best defined components are (1) slow component a (SCa, averaging about 1 mm/day); (2) slow component b (SCb, averaging 2 to 10 mm/day); (3) fast anterograde transport (averaging 200 to 400 mm/day), and (4) fast retrograde transport (averaging 150 to 300 mm/day). These transport components each reflect the behavior of a particular subset of cargoes, such as the neurofilament proteins (slow moving) and synaptic 387

388

Neurobiology of the Peripheral Nervous System

vesicles (fast moving). It seems increasingly likely that all components are propelled by a common mechanism as Ochs suggested 30 years ago in his “unitary hypothesis,”167 which recent work largely confirms.49,211 From this perspective the differences in macroscopic transport rates might stem from different relative likelihoods of motion and stasis among cargoes whose instantaneous peak velocities are similar. Nonetheless, it is clear that transport involves at least two different families of transport motors, the kinesins and the dyneins, along with multiple adapter molecules that determine overall behavior. Before turning to motors and molecular mechanisms, however, we will consider the basic phenomena displayed by the different transport components, along with the identities of their characteristic cargoes.

Fast Anterograde Transport One useful way to study fast anterograde transport is to examine the outflow of proteins that have been radiolabeled by incorporation of a radioactive amino acid precursor. Typical data obtained with radiotracer methods are shown in Figure 18–1. A few hours after introducing 3H-leucine into a spinal ganglion, much radioactivity remains at the injection site but significant amounts are spread distally along the sciatic nerve. This actively transported label appears as a plateau rising distally to a crest, then descending steeply to background levels. The foot of the wave front progresses steadily along the nerve at a rate that precise

experiments determine to be near 400 mm/day at mammalian body temperature.163 Maximal rates are similar in myelinated and unmyelinated fibers, in motor and sensory nerves, in animals of sizes from mouse to goat, and at ages from perinatal to older adult.164,170 The rate varies with local temperature and exhibits a temperature coefficient (Q10) of 2.2.172 This high value means that transport slows markedly when nerves are cooled. Fast transport ceases altogether at a critical temperature near 11° C, probably because the cytoskeleton is disturbed, but it readily recovers after rewarming.37,45 Fast transport depends on local oxidative metabolism and is blocked within 30 minutes by incubation in a nitrogen atmosphere.140,165,169 Transport through any given region of nerve typically fails whenever oxidative metabolism is impaired to the point at which the level of highenergy phosphate (ATP ⫹ phosphocreatine) falls to about half of normal. This happens regardless of whether the impairment is due to pure anoxia, to ischemia, or to metabolic poisons such as dinitrophenol.192 Transport failure after metabolic insult is typically accompanied by a failure of impulse conduction. Both of these functional deficits are reversible if physiologic conditions are restored within 90 minutes. After anoxia or ischemia for up to a few hours, however, transport recovery can require days.140 Still longer periods of metabolic impairment lead to irreversible collapse and wallerian degeneration of the axon. Along with a supply of energy, microtubules are equally critical for fast axonal transport. Transport ceases when

FIGURE 18–1 Distribution of radioactivity in the dorsal root ganglia and sciatic nerves of five cats taken between 2 and 10 hours after injecting 3H-leucine into L7 ganglia (G). The activity present in consecutive 5-mm segments of roots, ganglia, and nerves is shown in counts per minute (CPM). Top left scale for 10-hour nerve, bottom left scale for 2-hour nerve (partial scales for 4-, 6-, and 8-hour nerves at the right). Arrows mark the wave fronts, defined as the points where the steeply sloping forward curves intersect the background. (Data from Ochs, S.: Fast transport of materials in mammalian nerve fibers. Science 176:252, 1972.)

Axonal Transport: Properties, Mechanisms, and Role in Nerve Disease

microtubules disaggregate, either because of treatment with antimitotic drugs such as colchicine, maytansine, or vinblastine,59,84,130 or because of exposure to temperatures below 11° C.45 The first studies using video-enhanced contrast–differential interference contrast microscopy (VECDIC) revealed fast-moving axonal particles following well-defined tracks that proved, in electron micrographs, to represent microtubules.3,27 Later it was possible to reconstruct similar motility in highly simplified cell-free systems consisting only of microtubules, membranous organelles, and a source of crude or purified kinesin. As a result of such studies, it became clear that microtubules represent the scaffolding or stationary elements along which molecular motors propel specific organelles in specific directions. Microtubules are composed of ␣- and ␤-tubulin dimers oriented end to end to form protofilaments, 13 of which combine to make up the tubule wall.70 Formed microtubules are dynamic structures from 1 to 100 ␮m in length, in which tubulin subunits are constantly being added to the “plus” ends, oriented away from the cell body, and being removed at the opposite or “minus” ends.51,107 Fast-transport cargoes were identified by years of painstaking observations with two-dimensional gel electrophoresis and autoradiography, video microscopy, and nerve ligation, among other methods (see Grafstein and Forman87). Early on, biochemical and electron microscopic data indicated that proteins, peptides, and neurotransmitters undergoing fast transport were associated with a membranous compartment that included synaptic vesicles, vesicle precursors, and a smooth reticular network of axonal membranes.60,68,69,149 The predominant entities whose motion can now be visualized directly by VEC-DIC microscopy are small membrane-bound organelles whose apparent dimensions are consistent with those of synaptic vesicles.2 When it comes to specific substances, a “stopflow” approach involving localized cooling and rewarming has been effective (Fig. 18–2). This method can directly demonstrate the fast transport of materials associated with neurotransmission, such as acetylcholinesterase (AChE),48 norepinephrine and its biosynthetic enzymes,35,36,46 and substance P.44 It is now widely agreed that all substances undergoing fast anterograde transport are associated with membrane-limited organelles, either as integral components or as passengers engaged in reversible associations.

Fast Retrograde Transport Retrograde transport closely resembles fast anterograde axonal transport. It returns many anterogradely transported proteins to the cell soma after their delivery to the periphery. Ligation of a nerve that is actively transporting radiolabeled protein results in an immediate accumulation of radioactivity on the “upstream” side and a delayed accumulation on the “downstream” side.15 The latter is driven by retrograde transport. Analysis of band patterns by gel elec-

389

FIGURE 18–2 Downflow of accumulated dopamine-␤-hydroxylase activity in vitro after rewarming of rabbit sciatic nerves locally cooled to 4° C for 90 minutes. Numbered arrows indicate duration of rewarming to 37° C, in hours. Mean values from four to eight nerves. The foot of the wave front progresses at the rate of 12.5 ⫾ 0.7 mm/h, or 300 mm/day. (Data from Brimijoin, S.: Stop-flow: a new technique for measuring axonal transport, and its application to the transport of dopamine-␤-hydroxylase. J. Neurobiol. 6:379, 1975.)

trophoresis and autoradiography shows that the proteins accumulating on both sides are similar.18,72,136,166 Many proteins must therefore reverse direction at some point, either in midpassage or after reaching the axon terminals. As with fast anterograde transport, most retrogradely transported proteins are probably associated with membrane-limited organelles. Thus, when transport is blocked by a local constriction or cooling, tubelike membranous structures accumulate above the blockade and large, multilamellar structures accumulate below.218,229 The latter are bigger, which probably explains why most of the motion resolved by VEC-DIC microscopy is directed toward the cell bodies. In any event, the speed of retrograde transport averages about 70% that of anterograde transport, whether determined optically or by stop-flow studies of marker proteins such as AChE and dopamine-␤-hydroxylase (DBH).48,83,108 Retrograde transport also carries exogenous components. Either along with endocytosis of nerve terminal membrane or following entry through special channels, materials in the immediate environment of the nerve terminals are taken up

390

Neurobiology of the Peripheral Nervous System

and transported retrogradely. Retrograde transport allows nerve growth factor (NGF) and other neurotrophins to play a major role in promoting the early development of sensory, sympathetic, and motor neurons.207 On the negative side, toxicity and neurologic disease can result when foreign proteins such as tetanus toxin74,181,206,208 and viruses such as herpes simplex and poliovirus are taken up by the nerve terminals and transported to the cell body.132,133 In experimental settings, the retrograde transport of dyes and markers is useful for the tracing of neuroanatomic pathways. Like anterograde transport, retrograde transport depends on local temperature, oxygen supply, and microtubular integrity. Because the motor for retrograde transport is not a kinesin, however, this component can be selectively inhibited, for example, by the agent 9-erythro2-(hydroxy-3-nonyl)adenine.82 Retrograde transport may play an important role in the pathologic reaction of chromatolysis (see Chapter 31), which follows an axonal interruption owing to trauma such as nerve crush, tear, or transection. Chromatolysis begins in the parent cell bodies after a lag time consistent with delivery of a signal by retrograde transport from the site of injury.131,225 The nature of the chromatolytic signal is still unclear. Possibilities are (1) a loss of some transported substance needed to sustain the normal state (e.g., a neurotrophin); (2) arrival of a “foreign” substance from the site of injury (e.g., a captured cytokine); or (3) excessive return of normally transported materials. Indications that retrograde transport does play a role in initiating chromatolysis come from experiments with the antimicrotubule drug colchicine. Local application of colchicine to the proximal stump of a transected nerve prevents at least one of the metabolic changes that precede full-blown chromatolysis, namely an increased uptake of 2-deoxyglucose by cell bodies.131 It is reasonable to conclude that the signal for this early reaction requires a transport process that depends on microtubular integrity.

Slow Transport Slow transport is classically considered to comprise two components, SCa and SCb. In the mid-1970s, Hoffman and Lasek extensively investigated SCa, the slower component moving at approximately 1 mm/day. Their studies demonstrated that SCa consists primarily of the three proteins making up the neurofilament triplet, plus the ␣- and ␤-tubulin subunits of microtubules.110 Impressed by the coherence of the radiolabeled wave associated with this phase of transport, these investigators proposed that SCa represented a slow and synchronous motion of fully formed cytoskeletal elements. This “structural hypothesis” generated years of controversy. Other scientists argued instead that the materials actually moving in SCa are soluble protein subunits of microtubules and neurofilaments, which must be locally incorporated into structural polymers after transport to distal sites.

Controversy continued even after photobleaching techniques were able to demonstrate the translocation of injected cytoskeletal proteins with fluorescent tags. Now newer methods with the power to reveal the behavior of individual microtubules and neurofilaments in live axons have gone a long way toward resolving the issue. The outcome, which seems obvious in hindsight, is at variance with both the “structural hypothesis” and the “local incorporation” proposal. Namely, it appears that both neurofilaments and microtubules move as formed polymeric structures rather than soluble subunits. However, such motion is anything but slow and synchronous; rather, it is rare, irregular, and surprisingly rapid when it does occur (for recent reviews, see Brown49 and Baas9). The most powerful and direct experiments in support of the new picture utilized neurofilaments tagged with green fluorescent protein, in sympathetic nerve axons with a sparse and discontinuous cytoskeletal array that permitted observation of individual polymers.187,234 Such work has revealed that neurofilaments move at peak speeds up to 3 ␮m/sec, well within the range for rapid axonal transport, although average speeds are 10- to 100-fold lower. However, periods of movement are interrupted by long pauses, so that the probability of motion at any one time is low (as little as 1%). This behavior is easily explained in terms of simple transitions between moving and stationary phases, as earlier envisioned by Ochs.167 The elucidation of microtubule transport has posed greater technical difficulties because of the dynamic nature of these organelles, which are constantly lengthening and shortening as they add and lose subunits. Nonetheless, rigorous studies with techniques capable of resolving individual fluorescencetagged polymers have now demonstrated that microtubules do move as formed elements, like neurofilaments.64,235 The current focus of research in this area is on identifying the molecular motors responsible for transport of cytoskeletal elements. In the case of microtubules, a likely candidate is cytoplasmic dynein, which is optimally designed to propel the anterograde movement of plus-ended structures.9 Less is known about SCb, a component of slow transport that includes a far more diverse collection of materials, many of which are soluble cytoplasmic proteins. In fact this component includes clathrin, actin, actin-binding proteins, and the enzymes of glycolysis. The unsolved question is whether such cargoes are propelled by direct interaction with specific molecular motors (and if so, which ones), or they move by piggybacking on faster moving carrier molecules.

Transport Motors As noted earlier, two major types of transport motor have been identified, the kinesins (a large and diverse family) and the dyneins (a smaller and less diverse family). The properties of the dyneins and kinesins are important

Axonal Transport: Properties, Mechanisms, and Role in Nerve Disease

because these proteins are potential targets of toxins and genetic defects that might lead to nerve disease. Indeed, certain inherited neuropathies, such as Charcot-MarieTooth disease type 2A, have already been shown to result from mutations in specific kinesins.246 For treatment of molecular motors in greater depth, the reader is referred to more specialized reviews by Goldstein and associates.4,86 An abbreviated outline follows. Kinesins share a similar architecture with a conserved motor domain located at the amino terminus of a heavy chain, which may be present as a homodimer or heterodimer. This basic unit is typically attached to a light chain (kinesin I) or to a regulator/adapter subunit. These constituents define the specificity of a kinesin for its cargo, either by associating directly with integral cargo proteins, or perhaps more often by recruiting other linker proteins that mediate the interaction (Fig. 18–3). Several of the partner proteins that are relevant for axonal transport have recently been identified. One is a transmembrane protein known as “Sunday driver” (Syd).23 The other is amyloid precursor protein, also a transmembrane protein.121 Both of these molecules interact directly with the light chain subunit of kinesin I through its tetratricopeptide repeat domain. Mammals express up to 100 different kinesin molecules, each with their own specificities. Among these motors, the monomeric kinesin “Unc104”176 appears to be of major importance for the fast axonal transport of synaptic vesicle precursors, even though its low “processivity” means that multiple motors per organelle are required.179 The diversity of kinesins is undoubtedly multiplied by a wide variety of adapter proteins, including many not yet

391

discovered. It therefore becomes possible to envision highly specific and individualized propulsion systems underlying the rapid anterograde transport of each of the various membrane elements found moving in axons. In the limit, these systems may be able to deliver particular proteins or functional complexes to precise destinations. A motor cannot move a vehicle unless it has something to push (or pull) against. All kinesins derive their motor functions from directional interactions with microtubules, the underlying scaffold for rapid axonal transport. Most kinesins interact specifically with the plus end of microtubules and thereby direct motion away from the cell body, generating anterograde transport. A few kinesins, however, are “minus-end directed.” The axonally expressed protein KIFC2 appears to be of this class and may therefore participate along with dynein in powering rapid retrograde transport of certain cargoes. Cytoplasmic dyneins are the main drivers of retrograde transport. The lower diversity among dyneins may be related to a lesser need for address-specific delivery, because many materials undergoing retrograde transport are destined only for lysosomal degradation. Be that as it may, dyneins are notable for their large size and complex subunit composition, which suggest multiple possibilities for regulatory control that have only begun to be investigated. Dyneins, like kinesins, are proteins that are themselves subject to rapid anterograde transport. An active delivery system is required to maintain their presence and function in distal regions of the axon. Therefore, a disorder that impairs anterograde transport will eventually impair retrograde transport as well.

FIGURE 18–3 Motors and accessory factors in a model for the mechanism of organelle movement. Motor enzymes (kinesin or cytoplasmic dynein) and accessory factors are both required for transport of the organelle. The complex formed by the motor, the accessory factor, and a motor receptor on the translocated organelle is termed the organelle translocation complex. It is likely that different subfamilies of kinesins and dyneins, and different accessory factors, are utilized to supportaddress specific transport of diverse organelles and protein complexes in the axon. (From Sheetz, M. R., Stener, E. R., and Schroer, T. A.: The mechanism and regulation of fast axonal transport. Trends Neurosci. 12:474, 1989, with permission.)

392

Neurobiology of the Peripheral Nervous System

THEORETICAL IMPLICATIONS OF TRANSPORT ABNORMALITIES Understanding of the physiologic role of axonal transport has generated persistent interest in the possibility that defects of motility underlie some pathologies of peripheral nerve. The role of interrupted transport in wallerian degeneration is well accepted. In addition, although outright failure of transport is not likely to be the root cause of any human neuropathy, subtle disturbances of axonal transport may contribute to a variety of disorders. A look at the steps in the life cycle of a transported protein makes clear why this is so. Each component of transport involves several stages. In rapid transport the following stages can be identified: (1) packaging of newly synthesized proteins into organelles and loading into the proximal axon; (2) distally directed, kinesin-driven motion (with occasional pauses and transient reversals); (3) arrival at the destination and temporary stasis; (4) turnaround, a directional reversal sometimes preceded by insertion into axolemmal membrane and later retrieval; (5) rapid dynein-driven retrograde transport (also with occasional pauses); and (6) lysosomal digestion in transit or after arrival in the cell body. Not every protein follows the entire sequence because metabolism or release from the nerve cell may intervene. Nevertheless, there are multiple points at which the transport system could be attacked by toxicants or disease, with varying outcomes.

Loading Stage Defects Many years ago Hammerschlag100 showed that rapid axonal transport is initiated by a loading phase that displays a distinctive pharmacology. As compared with motion along the axon, the initiation of axonal transport depends strongly on calcium and is sensitive to antimitotic drugs and monovalent ionophores.101–104 One can thus entertain the concept of selective diseases of transport loading. The pathophysiology of such a disorder would include a reduced axonal flow of synaptic vesicles, other formed organelles, and associated proteins, neuropeptides, and neurotransmitters. Many transported substances show distally declining concentration gradients, suggesting preferred distribution to proximal regions.143 A defect in the loading of such substances might especially hinder renewal and repair of surface membranes at the ends of long axons. Paradoxically, therefore, a loading defect in the cell body is a potential cause for a “dying-back” axonopathy that appears first in distal regions of nerve.54 However, another deleterious consequence could be that materials destined for transport would accumulate in the cell body, causing local swelling and structural dislocation. Defects in the Golgi apparatus need not affect proteins with alternative routes to the axon. Such proteins include many slowly transported materials, for example, soluble

enzymes and the components of microtubules, microfilaments, and neurofilaments. How these proteins reach the axon is still obscure, but there is evidence that their delivery can be selectively impaired by neurotoxins and neurologic disease. For example, as discussed later in more detail, the entry of neurofilament proteins into axons is specifically affected by ␤,␤⬘-iminodipropionitrile (IDPN), which causes ballooning of cell bodies and giant, filamentpacked swellings in the proximal axon.94

Anterograde Transport Defects One easily imagined sort of malfunction is retardation, or slowing. Though pathophysiologic studies have tended to look for that kind of malfunction, present evidence suggests that slowing of rapid anterograde transport is atypical. However, the slower components of transport appear more vulnerable to retardation, as will become apparent. Normal axons have surplus capacity for rapidly transported materials and can compensate for reduced velocity by increasing the concentration of material in transit.37,39,85 For this reason a small change in transport velocity is not in itself physiologically relevant. The meaningful variable is the distal flux of material (amount moved per unit time). To reduce this flux may require extreme slowing or a simultaneous loss of carrying capacity, for example, because microtubules are disorganized or kinesin is deficient. Another way to reduce delivery of materials to distant sites without greatly changing average velocity would be to lower the probability of onward transport from one region to the next. Causes could include premature reversal, accelerated degradation of cargo, or increased incorporation of transported material into fixed structures. Suppose that the probability of continued transport over a 1-cm distance were reduced only 3%, from 1.0 to 0.97. The cumulative probability of transport over a 1-m distance (typical for long sensory nerves in humans) would then drop to (0.97)100 ⫽ 0.05. Experimentally, such a transport defect would manifest itself as a steepened profile of transported radiolabel, with normal displacement of the wave front. Such effects have been seen with cytoskeletal proteins in the mutant diabetic mouse.244 A defect that reduced the anterograde transport of macromolecules by limiting supply would have important consequences. Failure to replenish neuropeptides, which must be synthesized in the cell body, would quickly alter the character of neurotransmission, and depletion of synaptic vesicles would cause persistent failure. Loss of membrane precursors and associated ion channels would compromise impulse conduction and the renewal of the axolemma. Reduced delivery of neurofilament proteins, which form skeletal elements that determine cross-sectional area,137 would produce axonal dwindling. Many such abnormalities are characteristic of advanced neuropathies.

Axonal Transport: Properties, Mechanisms, and Role in Nerve Disease

Turnaround Defects Bray and colleagues32 and Bisby15,16 showed decades ago that about half the protein rapidly transported to distal axons ultimately returns to nerve cell bodies by retrograde transport. This protein circulation involves directional reversal at the terminals, or wherever fast transport is interrupted (there is apparently no “turnaround” of slow components). An impaired turnaround would lead to swellings packed with the membranous organelles that normally undergo rapid anterograde transport. Terminal and preterminal axons would be likely sites for such lesions. There is ample evidence that turnaround defects contribute to distal axonopathy in one or more welldescribed disorders.

Defects of Retrograde Axonal Transport Rapid retrograde transport appears to reflect the movement of large particles, including multivesicular bodies and prelysosomal structures returning to the cell bodies for disposal, and pinocytotic vesicles carrying trophic signals from the periphery. The similar mechanisms for rapid retrograde and anterograde transport generate a shared sensitivity to antitubule drugs, the same temperature and ionic dependence, and identical energy requirements (for review of early work, see Kristensson132). However, the maximal velocity of retrograde transport is only about two-thirds that of anterograde transport, and different motors are involved, as indicated earlier. Furthermore, retrograde transport is differentially sensitive to certain chemical agents.81 These observations open the possibility of a selective block of retrograde transport by neurotoxins or disease. A selective block of retrograde transport would probably resemble impaired turnaround, except that the supply of extraneuronal trophic factors to nerve cell bodies would also be reduced. Moderate deficits of trophic factors would alter the patterns of gene expression and protein synthesis, and severe deficits could cause atrophy and even cell death.

Potential Causes of Transport Defects A disease could act in many ways to impair axonal transport. Potential mechanisms include local pressure, ischemia, inhibition of glycolysis or oxidative phosphorylation, disturbances in the amounts or organization of cytoskeletal proteins, interference with kinesin or dynein ATPases, and alterations of the ionic balance in axons. Each of these mechanisms deserves brief consideration. Local Pressure Local pressure is difficult to elevate without also preventing access to oxygen and nutrients, but sophisticated experiments suggest that high pressure by itself impedes axonal transport.98,99,188 Transport fails totally under pres-

393

sures of 200 mm Hg. If pressure is released within 2 hours, transport resumes within 1 to 3 days, but sustained pressure causes irreversible damage.188 In any event, pressures sufficient to cause outright collapse of the nerve, and mechanical obstruction of axons, are classic causes of compression neuropathies. A case of special interest in this connection is the so-called double-crush syndrome proposed by Upton and McComas.231 According to this concept, an initial compression lesion in the proximal course of a nerve sensitizes the patient to a subsequent compression at a more distal site, facilitating, for example, the development of an entrapment neuropathy. Animal experimentation provides some support for this mechanism,210 but its applicability to real-life conditions such as cervical radiculopathy remains doubtful.156 Ischemia Ischemia reduces the availability of oxygen and glucose together. Both are important in generating the ATP needed for rapid axonal transport.162,171,192 Actually, transport is surprisingly resistant to moderate hypoxia,159 though it is readily interrupted by complete anoxia even when highly localized.162 Nevertheless, ischemic impairment of transport has been demonstrated experimentally129 and is a potential contributor to the neuropathology of glaucoma,5 compression neuropathies (see Chapter 57), necrotizing angiopathic neuropathy,71 thromboangiitis obliterans, and advanced diabetes (see Chapter 85). Impaired Glycolysis or Oxidative Phosphorylation Any deficiency of energy metabolism, either from inherited enzyme defects or from exogenous toxins, tends to deplete cellular stores of ATP. Because fast transport depends on locally available ATP,125,169,171 it certainly would be impaired; other transport components would probably be affected as well. Especially interesting is the possibility that a toxic neuropathy might result from a localized ATP deficit with localized failure of transport. Cytoskeletal Disturbances Antimitotic drugs, including colchicine and the vinca alkaloids vinblastine and vincristine, induce the disaggregation of microtubules. Consistent with evidence for common mechanisms, these agents appear to block all components of transport.11,12,131 Such effects provided the initial evidence that microtubules are part of the machinery for transport. Now we recognize that microtubules serve to direct the propulsive forces provided by extrinsic motors and to determine transport capacity. The well-known peripheral neuropathy induced by vincristine is the limiting side effect in the use of vinca alkaloids to treat leukemia.237 It is highly likely that vincristine neuropathy results from a severe loss of transport capacity after microtubular disaggregation.26,136 Another agent that causes

Neurobiology of the Peripheral Nervous System

cytoskeletal disturbance is IDPN. As mentioned earlier, IDPN neuropathy involves marked retardation of the slow transport of neurofilament proteins, accompanied by giant, filament-rich swellings in the proximal axon. Disturbed Ionic Balance Fast transport depends, at least indirectly, upon appropriate axonal cations. The importance of Ca2⫹ for the loading stage of fast transport has already been mentioned. Studies with permeabilized cells and reconstituted systems of kinesin, microtubules, and isolated organelles show that Mg2⫹-ATP is sufficient to support transport per se.28,80 However, earlier indications that Ca2⫹ also affects transport173 are supported by some recent evidence for a regulatory role.33,151 Accordingly, long-term changes in serum electrolytes, particularly divalent cations, may be pathologically significant. Toxic Effects on Transport Motors Kinesins and dyneins are known to exhibit different sensitivities to inhibition by drugs.73,81,178 Therefore, differential toxicity for anterograde and retrograde transport motors becomes a reasonable topic of investigation. At another level, certain neurotoxic agents may eventually be found to impair individual motors, leading to disrupted transport of the corresponding cargoes. As yet, however, we know of no instances in which differential inhibition of particular transport motors contributes to the pathogenesis of neuropathy.

EVIDENCE FOR TRANSPORT ABNORMALITIES IN HUMAN NEUROPATHY If abnormalities of transport are probable causes of peripheral neuropathy, transport should be affected in actual diseases of human nerve. Although radiolabeling in vivo is not applicable to studies in humans, isolated axons continue transport in vitro. This transport can be studied by biochemical and optical methods. For biochemical experiments in human nerve, biopsy samples ligated at both ends are incubated in oxygenated physiologic salt solution. Later the distribution of rapidly transported enzyme or neurotransmitter is examined in consecutive short segments of the sample. So long as the total amounts of protein or transmitter are constant, an accumulation in the distal segment can be attributed to redistribution by anterograde axonal transport. The rate of accumulation is the transport flux, that is, the amount of material delivered per unit time. From the length of nerve that must contribute its content each hour to produce the accumulation, one calculates an overall average velocity of transport for a given marker. Interpretation of this measure is limited by uncertainties about the size of

the moving fraction.38 In favorable cases one can estimate the moving fraction by observing the clearance of marker from the middle of the sample or by identifying a break in the course of accumulation. Even when this is not possible, the average velocity of transport is useful because it is independent of the number of axons in the sample. A reduced average velocity of transport implies an actual slowing or an increased proportion of stationary material. Either would point to an abnormality of transport as opposed to a simple loss of fibers. Our initial studies of enzyme redistribution in human sural nerve utilized two enzyme markers, DBH and AChE, that catalyze the biosynthesis of norepinephrine and the hydrolysis of acetylcholine, respectively (Fig. 18–4). Observations of DBH and AChE redistribution indicated consistent reductions in transport in certain hereditary peripheral neuropathies. Transport abnormalities were especially pronounced when the sensory system was involved, alone or together with the motor system.40–42

Hours 5.0 4.0 3.5 3.0 2.5 DBH activity

394

2.0 1.5 1.0

Distance

0 Proximal 0

Distal 15

30

mm along nerve Ligature FIGURE 18–4 Redistribution of dopamine-␤-hydroxylase (DBH) activity in biopsy specimens of normal human sural nerve incubated in vitro. Each profile represents enzyme activity (measure of DBH protein) in consecutive 3-mm segments along an individual nerve ligated for the indicated number of hours. Accumulation proximal to the ligature occurs at a rate corresponding to transport at an average velocity of 2 mm/h. (From original data reported by Brimijoin, S., Capek, P., and Dyck P. J.: Axonal transport of dopamine-␤-hydroxylase by human sural nerves in vitro. Science 180:1295, 1973.)

Axonal Transport: Properties, Mechanisms, and Role in Nerve Disease

Dramatic deficits were seen in several cases of CharcotMarie-Tooth disease. That finding is consistent with the latest data. For example, it is now known that some varieties of this disorder reflect mutations in the light subunit of neurofilaments (NF-L),61 which disrupt neurofilament assembly and axonal transport.50 Of equal interest, another variety of the same disorder has been shown to involve mutations in the kinesin motor protein KIF1b␤.246 Another early study, by Behrens and co-workers,13 reported complete failure of AChE accumulation in a case of myotonic dystrophy, although enzyme content was more than 3 standard deviations above the mean from a series of normal nerves. In diabetic neuropathy, statistically significant reductions of enzyme content and of average transport velocity have been observed both with DBH and with AChE. We recognize that some of these findings might reflect end-stage axonal damage. However, the abnormalities of enzyme redistribution in neuropathic biopsies are also consistent with fundamental disturbances of transport. Optical methods are well suited to analysis of transport kinetics in neuropathy (Fig. 18–5). Nearly 20 years ago, DIC microscopy was applied to axonal transport in large myelinated sural nerve axons from cases of peroneal muscular

FIGURE 18–5 Video-enhanced contrast–differential interference contrast microscopy image of a single organelle undergoing rapid axonal transport in vitro. Digitized images of an axon were captured at 3-second intervals and electronically compiled while maintaining precise spatial registration. The transported organelle was digitally highlighted. Note that successive organelle displacements follow a consistent track but are not all equally long. Average velocity of this organelle was approximately 0.5 ␮m/s. In observations on nearly 300 particles from 19 axons, mean anterograde velocity was 2.2 ⫾ 0.6 ␮m/s and mean retrograde velocity was 1.4 ⫾ 0.4 ␮m/s. (D. Brat and S. Brimijoin, unpublished data, 1992.)

395

atrophy, Friedreich’s ataxia, and diabetic polyneuropathy.123 Visible particles were transported in some axons of all samples studied, and transport velocity varied considerably from case to case. Conclusions were limited because control biopsy specimens were not available. Another older optical study161 examined motor nerves in intercostal muscles of patients with amyotrophic lateral sclerosis (ALS). Estimated transport velocity was slowed by 30% in comparison with normal nerves from patients undergoing radical mastectomy. With the arrival of modern VEC-DIC microscopy and image processing, Breuer and colleagues were able to evaluate retrograde and anterograde transport in the motor branch of the median nerve in patients with ALS.33,34 Controls were various somatic nerves from patients undergoing limb amputation for removal of tumors, but at least one median nerve from a brain-dead donor was included. The design limitation imposed by the use of different types of nerve in patients and controls was mitigated by the absence of any known variation in the basic kinetics of transport in normal somatic nerves. Interestingly, the only significant kinetic abnormality in ALS nerves was a 57% increase in speed of anterograde transport. In contrast, the same analysis showed a 70% decrease in traffic density, or total flux expressed as the number of retrogradely moving organelles per second per square meter of axoplasm. This result is interesting because anterograde traffic density was only slightly reduced. The imbalance between anterograde and retrograde traffic suggests a defect in transport turnaround, and it may reflect an important biologic abnormality in ALS. Neurofilament dynamics are frankly disturbed in a number of neuropathic conditions. For example, axonal aggregations of neurofilamentous material are commonly observed in Alzheimer’s disease, in diabetic neuropathy, and in ALS.53,109,200,201 Such aggregations may very well reflect disruptions in the axonal transport of the underlying proteins or the neurofilaments themselves. Arguments for a primary transport disturbance are favored by observations in transgenic mouse models of ALS, which are discussed at the end of this chapter. Putting together the pathologic picture of neurofilament aggregation in disease states and the data from enzyme and optical studies, one concludes that many human peripheral nerve diseases probably do involve abnormalities of axonal transport.154 To determine whether the abnormalities cause the associated nerve damage or merely reflect structural damage from other causes, animal models of neuropathy must be investigated. The current status of research in this area is dealt with next. Meanwhile, it is worth emphasizing that even a secondary failure of transport could determine the course of neuropathy by inducing axonal atrophy, deficits in axolemmal renewal, impaired nerve conduction, failure of synaptic transmission, degeneration of nerve terminals, loss of trophic interactions, and reactive changes in the perikaryon.

396

Neurobiology of the Peripheral Nervous System

AXONAL TRANSPORT IN EXPERIMENTAL NEUROPATHIES Overview of Toxicant-Induced Neuropathies In 1969, Pleasure and associates reported marked reduction of slow axonal transport of radiolabeled protein in the central processes of sensory neurons of cats with acrylamide neuropathy.180 Since that original study, both fast and slow anterograde transport have been measured by various methods in many experimental neuropathies (Table 18–1). Minor reductions in fast-transport velocity have been encountered in the neuropathies induced by zinc pyridinethione (ZPT),147 acrylamide,26 and vincristine,26 although some studies have reported normal velocities in the same conditions.217 More substantial slowing of fast axonal transport, up to 36%, has been seen in hexacarbon neuropathy.152 Overall, however, one is struck by the extent to which fast anterograde transport seems to persist at full speed despite substantial axonal pathology. Few if any of the toxicant-induced neuropathies involve gross defects of fast anterograde transport. Some more pronounced effects on slow anterograde transport in several different toxic neuropathies are detailed subsequently.

Experimental Diabetes There has been disagreement about the status of rapid anterograde transport in diabetes (see Table 18–1). One early study reported slight reductions in the flux of AChE, a fast-transported enzyme.203 Others could neither confirm this observation nor detect changes in the velocity of fast-transported radiolabeled protein.10,17,116,238 A fair conclusion is that velocity and flux of fast anterograde transport are minimally affected in models of diabetes, whether induced by streptozotocin or alloxan. Experimental diabetes does seem to impair slow transport. This impairment is manifested as a reduction in the accumulation of enzymes such as choline acetyltransferase and as a retardation of the wave of radiolabeled protein in SCa.119,152,203,227,244 Certain of these defects are reversed or prevented by insulin, by myoinositol, or by aldose reductase inhibitors.147,226,228 Such beneficial effects suggest that the transport abnormalities arise specifically from the defect in carbohydrate metabolism. Conversely, inhibition of aldose reductase seems not to improve axonal transport of SCa, and there is conflicting information about the effect of insulin.135,227 Thus present information leaves room for doubt as to which transport effects represent streptozotocin neurotoxicity and which are truly metabolic. The retardation and reduced flux of neurofilament protein are nonetheless significant and might explain the modest but definite axonal dwindling in experimental diabetes.113,114 Turnaround and retrograde transport also appear to be impaired substantially in experimental diabetes. Jakobsen

and Sidenius found a significant reduction in the turnaround and retrograde transport of label associated with rapidly transported glycoconjugates in the sensory nerves of rats with streptozotocin diabetes.118 This reduction appeared within 1 day after onset of the diabetic state216 and persisted for at least 8 weeks.118 Significantly, turnaround was restored when blood sugar was normalized by insulin treatment.116 Hence the defect was probably not a primary neurotoxic effect of streptozotocin but a consequence of deranged carbohydrate metabolism. Unfortunately, fullblown neuropathy does not usually develop in animals with chemically induced diabetes, so the relevance of these observations to the neuropathology of human diabetes can still be questioned. We would argue nonetheless that diabetic neuropathy does involve deficits in transport reversal and retrograde transport. Additional work on this problem has emphasized the potential importance of deficits in the retrograde transport of neurotrophic factors from target organs. In 1981 our research team found that retrograde axonal transport of exogenous NGF was reduced in sciatic nerves of diabetic rats.116 This observation has been repeatedly confirmed.139,204 Such effects have been shown to persist in normoglycemic media in vitro.202 Hence they must reflect an abnormality of the nerve cell, not of its external environment. In vivo, deficits of NGF transport per se are compounded by a deficient expression of NGF in target tissues.78 Similar findings have been reported with regard to the related neurotrophin-3 (NT-3).79 Those observations take on added importance from the fact that NT-3 is a critical growth factor for many of the neurons that are most susceptible to diabetic neuropathy.75,77,126 It appears that deficits in neurotrophin transport are related in part to deficient uptake, owing to reduced expression of the relevant receptor proteins, among them tyrosine kinase A (TrkA) and p75.62,63,144 Other factors may also play a role, such as phosphatidylinositol 3-kinase/Akt signaling. This pathway, which is crucial for the internalization and retrograde transport of NGF in certain types of nerve,105,185 shows depressed activity in vagal neurons of diabetic rats.52

Acrylamide Until recently, opinion was divided as to whether acute administration of acrylamide was directly toxic to axonal transport and to what extent such toxicity could account for acrylamide neuropathy. A good deal of the evidence was negative. For example, in a careful study of proteins radiolabeled by 3H-leucine injection into dorsal root ganglia of the rat, Sidenius and Jakobsen found no significant change in the velocity of any transport component.217 Moreover, our own research group and that of Padilla found normal fluxes of transported vesicles in nerves and neurites observed by video microscopy, even when acrylamide was present in concentrations up to 1 mM for up to

Marked slowing (dorsal root only) No change — Large drop in vesicle flux 40% drop in vesicle flux at 1 h — — 25% slowing — — 40% less ChAT accumulation — 20% slowing — Normal velocity — — Normal velocity No change

— — SCa totally blocked — No change in ChAT accumulation — — NF transport speeds then slows No change 30% slowing No change — — —

Rat sciatic (sensory) Rat sciatic (sensory) Rat sciatic (motor) Rabbit sciatic Rat sciatic (cholinergic) Rat sciatic (sensory) Chicken sciatic Chicken sciatic Cat spinal roots Cat sciatic (sensory) Cat sciatic (sensory) Cat vagus Rat sciatic (motor) Rat sciatic (motor, sensory)

Slow Transport Effects

Cat spinal roots Cat sciatic Chicken oculomotor Cultured rat DRG neurons Mouse sciatic Rat sciatic (sensory) Rat sciatic (motor) Rat sciatic (sensory) Rat sciatic (motor) Neuroblastoma culture Rat sciatic (cholinergic) Rat sciatic (sensory) Rat sciatic (sensory) Rat sciatic (mixed) Rat sciatic (sensory) Rat ileal mesenteric Rat vagus (sensory) Rabbit vagus (sensory) Chicken sciatic Rat sural

Nerve

Reduced retrograde transport — — 10% slowing 25% slowing Irregular slowing Reduced AChE flux, antero- & retrograde Minor slowing, major block of turnaround

20% acceleration

Normal velocity, irregular block of retrograde transport

Focal blockade in swellings 10%–35% slowing; partial multifocal blockade No change

Normal AChE & DBH accumulation, impaired turnaround Normal velocity, reduced retrograde transport of NGF Reduced retrograde transport of NGF Reduced retrograde transport of NGF and NT-3 Normal velocity — 20% decreased velocity

Delayed onset, normal velocity No slowing, no deposition at sites of focal demyelination Delayed onset, normal velocity, impaired turnaround Accelerated onset, normal velocity, impaired turnaround Decreased anterograde and retrograde vesicle flux 20% reduction of AChE accumulation Delayed onset, normal velocity, impaired turnaround

— 10% slowing Focal retention in distal axon

Fast Transport Effects

AChE ⫽ acetylcholinesterase; BPAU ⫽ p-bromophenylacetylurea; ChAT ⫽ choline acetyltransferase; DBH ⫽ dopamine-␤-hydroxylase; DFP ⫽ diisopropylfluorophosphate; DRG ⫽ dorsal root ganglia; IDPN ⫽ ␤,␤⬘-imidodipropionitrile; NF ⫽ neurofilament; NGF ⫽ nerve growth factor; NT-3 ⫽ neurotrophin-3; SCa ⫽ slow component a; STZ ⫽ streptozotocin; TOCP ⫽ triorthocresylphosphate; — ⫽ not reported.

Acrylamide Acrylamide24 Acrylamide56,220 Acrylamide106 Acrylamide222 Adriamycin214 Antigalactocerebroside7 BPAU115 BPAU157,158 Cremophor EL31 Diabetes (STZ)203 Diabetes (STZ)215 Diabetes (STZ)120 Diabetes (STZ)116 Diabetes (STZ)17 Diabetes (STZ)204 Diabetes (STZ)138 Diabetes (alloxan)10 Diphtheria toxin122 Ethylene oxide160 Hexacarbons 2,5-Hexanedione96 Methyl n-butyl ketone152 IDPN93 Metals Aluminum141 Lead174 Methylmercury232 Organophosphates DFP, Mipafox, others155 DFP97 TOCP180 TOCP26 Vincristine26 Vincristine90 Vitamin E deficiency219 Zinc pyridinethione131

180

Agent (Reference)

Table 18–1. Experimental Neuropathy and Axonal Transport

398

Neurobiology of the Peripheral Nervous System

48 hours.30,177 By contrast, in an experimental system similar to that of Sidenius and Jacobsen, Sickles observed a substantial depression in the amount of radiolabeled protein undergoing fast anterograde transport.212 Such divergent results prompted a debate over factors that could spuriously affect measurements of transport. Over the past 10 years, however, a body of data has accumulated to show that acrylamide does impair axonal transport, and to provide an explanation for previous divergent findings. Much of the new evidence on acrylamide neuropathy comes from research by Sickles and colleagues. As summarized in a recent review,213 this work has established several important points. First, the effects of single exposures to acrylamide are limited in duration and easily overlooked when measurements are made more than a few hours after dosing, as happened in many earlier studies. Second, the amount of transported cargo is affected more severely than the maximal velocity. Third, acrylamide is capable of interacting directly with neural kinesins to impair transport capacity. Fourth, in different model systems based on cultured neurons and neural explants, acute exposure to acrylamide does reduce the traffic of rapidly transported, microscopically detected particles. Among the few older findings without obvious explanation to date are the observations of Brat and Brimijoin,30 which were obtained with neuron-like neuroblastoma cells. We now suppose that this discrepancy arises from a special feature, as yet unrecognized, of this tumor-derived cell line. When it comes to the severe clinical neuropathy that develops after repeated exposure to acrylamide, debate has focused not on the existence of transport abnormality but on its relationship to the neuropathology. Is it a cause or an epiphenomenon? It has long been recognized that at least one aspect of transport is severely disturbed in acrylamide neuropathy, as in experimental diabetes: turnaround and fast retrograde transport. Sahenk and Mendell197 obtained initial evidence for a turnaround defect, which was confirmed by Jakobsen and Sidenius.120 Droz and co-workers used elegant autoradiography in the ciliary ganglion of the chicken to clarify the structural basis of the defect. Application of this technique after incorporation of radioactive amino acid in cell bodies made it possible to show that acrylamide neuropathy was associated with abnormal retention of labeled material in the preterminal portions of axons.220 Thus the turnaround defect did not involve loss of proteins by metabolism or by release from the nerve terminal. Furthermore, electron microscope autoradiography demonstrated that the retained label coincided with accumulations of smooth endoplasmic reticulum.56 Because structurally similar membranes are chief among the elements undergoing rapid axonal transport, it is logical to infer that impaired turnaround directly causes the distal accumulation of membranous material in acrylamide neuropathy. Some investigators, however, have argued that the transport abnormalities in acrylamide neuropathy are nonetheless

epiphenomena.142 One reason for this contrary view is that large doses of acrylamide can cause behavioral signs of severe neuropathy, including foot splay and muscle weakness, without generalized axonopathy. Another is that the development of axonopathy is strongly influenced by the rate at which a given total dose is delivered. The pathology is greatest when administration is prolonged by repeated administration of small quantities, and least when a given dose is administered as a single large bolus.142 Such findings might suggest that axons are not the primary sites of acrylamide toxicity. This argument is not fully persuasive because a primary defect in transport might well manifest first in the nerve ending or the cell soma rather than the axon itself. Of course it is entirely possible that acrylamide neuropathy involves toxic actions on additional processes and proteins. Meanwhile, it remains reasonable to consider that a direct attack on the molecular motors for anterograde transport probably does represent an early and important step in the cascade of events leading to acrylamide neuropathy.

␤,␤⬘-Iminodipropionitrile IDPN neuropathy is still a prime example of a nerve disease in which impaired axonal transport is causative. The selective effect of IDPN on transport of neurofilament proteins in SCa93,128 provides a direct explanation for the giant filamentous swellings in the proximal axon and for the atrophy of the distal axon. The earliest effect of IDPN is a segregation of neurofilaments and microtubules. Local applications of IDPN cause neurofilaments to cluster in the subaxolemmal region, while an inner core of normally spaced microtubules continues to support fast axonal transport.91,95 These findings imply that IDPN does not attack transport per se but rather alters the properties of neurofilaments or their subunits so that they are no longer readily transported. Segregation of filaments in treated axons initially suggested an increased interaction among these structures (e.g., cross-linking) that does not directly involve the transport machinery. The distinction may be semantic because, no matter how it is initiated, transport blockade has a clear role in the expression of IDPN neurotoxicity. However, detailed biochemical studies of such phenomena are shedding new light on the general mechanism of filamentous neuropathies. In the case of IDPN, the underlying chemistry is still not clear, but there is reason to suspect that cyanoacetaldehyde or another reactive metabolic intermediate may covalently combine with lysine residues on the neurofilament subunits.199 The altered charge distribution on the protein surface would be a plausible basis for increased aggregation. Another plausible mechanism for altered intermolecular associations of neurofilaments in IDPN neuropathy was proposed by Eyer and associates.76 These workers found no covalent cross-linking but noticed increased autophosphorylation and in vitro gelation capacity of neurofilaments isolated from IDPN-treated rats. Similar effects

Axonal Transport: Properties, Mechanisms, and Role in Nerve Disease

occurred when IDPN was added directly to isolated neurofilaments but not when neurofilament-associated proteins were first removed by salt treatment. It was hypothesized that IDPN promotes interaction between neurofilament polymers through a preferential effect on associated proteins that form a minor but important part of the cytoskeletal network.

Hexacarbons A striking disturbance of fast anterograde axonal transport has been observed in the toxic neuropathy induced by methyl n-butyl ketone (MnBK) and its metabolite, 2,5hexanedione (2,5-HD). This disturbance bears a rough quantitative relation to the severity of neuropathy.152 Structural studies reveal that the neurotoxic hexacarbons cause multiple filament-packed axonal swellings, especially in the distal parts of long motor nerves.45 Mendell and coworkers, noting that transport was progressively impaired as increasingly long stretches of axon were examined, proposed that MnBK induced partial, multifocal blockade.152 Direct evidence for this concept was provided by Griffin and co-workers, who showed that rapidly transported proteins tend to accumulate in and proximal to the axonal swellings in 2,5-HD neuropathy.96 At present it can be argued that the cumulative effects of partial multifocal transport blockade, by impairing the delivery of fast-transported proteins to distal axons and terminals, partly account for dying-back degeneration. The pathogenic event, however, may be the appearance of filamentous swellings that represent deposits of the slowly transported neurofilament proteins. Two major explanations have been offered to account for focal failure of slow transport in 2,5-HD neuropathy. Sabri and Spencer have argued that local energy deficits are responsible.189–191,193,221 Their in vitro experiments show that MnBK and 2,5-HD are weak inhibitors of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and to a lesser extent, of phosphofructokinase (PFK).191 Both enzymes are essential for glycolysis and hence for generation of axonal ATP. A curious feature is the slow onset of the inhibition of GAPDH and PFK by neurotoxic hydrocarbons.193 This delay might explain why neuropathy develops only after weeks of daily dosing. Slow cumulative inhibition, with the target enzymes becoming critically compromised only as they reach the distal axon, also provides an explanation for the characteristic locus of pathology in hexacarbon neuropathy. However, new evidence suggests that hexacarbon axonopathy has more to do with the chemistry of neurofilaments than with axonal energy metabolism. It is now apparent that 2,5-HD reacts directly with free amino groups on proteins, especially neurofilaments, leading to pyrrole formation and, perhaps, extensive crosslinking.88,89 Such a mechanism could account for the

399

selective impairment of the transport of neurofilament components. It may also explain the location of the filamentous lesions. Thus analogues of the parent ␥-diketone can be shown to produce distal, proximal, or medial lesions, depending on their neurotoxic potency.6,92 Likewise, varying the dosage of toxicants that ordinarily induce proximal lesions (e.g., 3,4-dimethyl-2,5-hexanedione and IDPN) can lead to dramatic shifts in the location of axonal lesions.6,92 Direct evidence is still needed, but these results allow us to speculate that the locus of lesions in both cases is determined by the speed with which neurofilaments are chemically modified. Hexacarbon neuropathies exhibit a surprising acceleration of the transport of neurofilament proteins,29,150 which narrows the choice among putative pathogenic mechanisms. This effect is difficult to accommodate by hypothesizing energy deficits or toxicant-induced cross-linking, but it supports the idea of abnormal intermolecular interactions. Sayre and colleagues199 have proposed that the central mechanism of all filamentous neuropathies may be chemical alterations that disrupt the organization of neurofilaments, whether by cross-linking, pyrrole formation, or abnormal phosphorylation.

Zinc Pyridinethione Abnormalities in the turnaround of rapidly transported proteins were demonstrated first in the experimental neuropathy induced by ZPT. Using Bisby’s method of delayed nerve ligation after injecting radioactive amino acid into the spinal cord, Sahenk and Mendell196 showed a reduced return of labeled protein by retrograde transport in motor nerves of rats fed ZPT for 15 to 20 days. Further analysis indicated that ZPT caused both a delay in onset of retrograde transport and a reduction in the amount of returning protein.196 In parallel experiments, measurements of fast anterograde transport velocity failed to detect any consistent effect on this phase of transport, and the occasional reductions of velocity did not correlate with the symptomatic severity of neuropathy. Sahenk and Mendell concluded that ZPT selectively impaired the turnaround of protein from anterograde to retrograde transport. This impairment provided an explanation for the accumulation of branched tubular structures in distal axons.195 Later data implicated axonal proteases in transport turnaround and suggested that ZPT interferes with this crucial mechanism. Thus Schroer and colleagues showed that protease treatment of synaptic vesicles from squid axoplasm converts their in vitro transport from anterograde to retrograde.205 Martz and co-workers obtained evidence for proteolysis during turnaround in vivo, modifying the apparent mass of specific, vesicular proteins.145 Significant in the present context is the demonstration by Sahenk and Lasek that protease inhibitors profoundly impair transport turnaround when applied directly to axonal tips and also

400

Neurobiology of the Peripheral Nervous System

induce organellar accumulations that closely resemble the lesions in ZPT neuropathy.194 Whether ZPT is itself a protease inhibitor, and what its spectrum of activity might be, has never been established.

p-Bromophenylacetylurea One other disorder in which impaired turnaround has been intensively studied is an experimental neuropathy induced in rats by p-bromophenylacetylurea (BPAU). This distal axonopathy is convenient for analyzing the onset of peripheral nerve disease because it is produced by a single dose of compound and becomes symptomatic only after a latent period of several days.55,66 Work in the author’s laboratory showed that BPAU induces limb weakness accompanied by selective deficit of retrograde transport. When the sciatic nerve is ligated a few hours after injection of 3 H-leucine into dorsal root ganglia, accumulation of radiolabeled protein above the ligature is normal but accumulation below the ligature is less than half of control. There are several key features of this impairment in transport reversal. First, the defect occurs without any change in the velocity or total flux of labeled protein or enzymes moving by rapid anterograde axonal transport. Second, the abnormality appears in a matter of hours and precedes obvious structural damage. Third, in animals treated with BPAU, the decrease in retrograde accumulation correlates well with the clinical severity of the neuropathy115 and with the reduced amplitude of compound action potentials in the hind limb musculature.117 Because of the pronounced motor effects, further studies of axonal transport in BPAU neuropathy were performed in motor nerve after intraspinal injection of radioactive amino acid.157 Again the velocity of anterograde transport was unaffected while the recirculation of labeled protein by retrograde transport was reduced in a dose-dependent manner. The reduction was greatest during the later phases of accumulation. Therefore, the defect involves a net deficit in turnaround rather than a simple delay. Electron microscopy further strengthens the link between the transport defect and the characteristic tubulomembranous axonopathy of BPAU.19 Statistical analysis of autoradiograms shows that the lesions in BPAU-treated motor nerves are associated with rapidly transported, radiolabeled proteins. When examined at appropriate times after intraspinal radioisotope injection, most of the label is located over lesions and most of the lesions are heavily labeled.175 Thus the evidence in hand for BPAU neuropathy resembles that for acrylamide neuropathy. In both, the membranous axonal lesions appear to be local deposits of material trapped at the end stage of fast anterograde transport or at the onset of retrograde transport. The common pathogenic element may be local stasis of fast-transported organelles in the most distal reaches of the neuron.

Another interesting abnormality in BPAU neuropathy is a shortened lag between injection of protein precursor in ganglia and onset of transport in axons. In a structureactivity study of seven analogues, only the two neurotoxic agents, BPAU and its close relative p-chlorophenylacetylurea, were able to shorten transport lag.158 This information suggests that BPAU causes abnormal processing of proteins destined for transport. The possibility even arises that the neuropathy results from that effect. A working hypothesis for pathogenesis is as follows: (1) shortened lag time suggests accelerated transit of the Golgi apparatus; (2) accelerated transit could affect the posttranslational modification of proteins on the organellar surface; (3) organelles with abnormal surface chemistry may interact inappropriately with axonal translocation factors and be subject to local stasis; and (4) growing deposits of organelles could plug off distal axons and cause dying-back degeneration.

GENETICALLY BASED ANIMAL NEUROPATHIES A number of genetically mediated neuromuscular disorders of animals have attracted attention as potential models of the inherited neuropathies of humans. Chief among these are the muscular dystrophies occurring in mutant strains of mice, chickens, and hamsters, which resemble their human counterparts in many of their clinical and pathologic features.8,111,153 Axonal transport has been examined in all these disorders.

Murine Dystrophy Mice of the Jackson Laboratories ReJ/129 strain suffer from a hereditary autosomal recessive dystrophy with severe progressive limb weakness. There is minimal loss of axons but striking failure of myelination in the ventral and dorsal roots as well as in the proximal sciatic plexus.25 Early attempts to test the “neurogenic hypothesis” of muscular dystrophy148 indicated that the synthesis and axonal transport of proteins in dystrophic motor nerves may be abnormal. Bradley and Jaros, analyzing transported radioactivity by autoradiography, observed an increase in the amount of fast-transported label and a decrease in the amount of slow-transported label.24 Jablecki and Brimijoin112 noted a 40% decrease in the distal flux of choline acetyltransferase activity despite a 20% increase in the local content of enzyme activity. These findings are in reasonable accord because choline acetyltransferase is thought to move by slow axonal transport.71,198 However, in a later ligation study, dystrophic sciatic nerve also showed a deficient flux of the rapidly transported enzymes DBH and AChE.43

Axonal Transport: Properties, Mechanisms, and Role in Nerve Disease

Others reported conflicting data on the flow of labeled protein and lipid along the sciatic nerve after injection of radioactive precursors into dystrophic spinal cord or spinal ganglia. One group found normal velocities of transport but increased amounts of labeled material in dystrophic nerves.127,134,224 The largest difference was in the rapid component, which had three to five times the normal labeling, though the waves of radioactivity were indistinct and may have included blood-borne radioactivity. It was argued that murine dystrophy involves increased synthesis and rapid transport of selected proteins. Nearly opposite results were obtained by McLane and McClure, who used collection ligatures to distinguish transported label from local uptake.150 They noted a 40% decrease in the accumulation of rapidly transported protein labeled by injection of radioactive amino acid into the dystrophic spinal cord. When spinal ganglia were incubated with the same precursor in vitro, the results were similar and the impairment of transport progressively increased to a near-total block at 60 days of age. These conflicting observations have never been resolved. It should be worthwhile to determine whether the flux of rapidly transported proteins is actually increased or decreased. Reassessments of this problem should include controls for (1) amino acid pool sizes in the spinal cord, (2) blood-borne label, (3) alterations in the delay in initiation of transport, and (4) abnormalities of turnaround.

Related Dystrophies Axonal transport also has been examined in the dystrophic chicken and hamster. Although the neuromuscular abnormalities are superficially similar to those of the dystrophic mouse, there is no demyelination and transport seems unaffected. In dystrophic chickens, the anterograde and retrograde fluxes of AChE are normal, as is the pattern of this enzyme’s molecular forms.67 Likewise, the transport of 3H-leucine– or 3H-fucose–labeled protein in radial nerves does not differ in rate or amount from values observed in normal birds.223 These findings support the qualitative conclusions of an earlier study on retrograde labeling of motor neuron cell bodies in the spinal cord after injection of peroxidase into the latissimus dorsi muscles.65 In the dystrophic hamster, the anterograde flux of choline acetyltransferase and the velocity of fast transport of labeled protein along the sciatic nerve are apparently normal.21 Evidently, disturbances of axonal transport are not a universal feature of genetically mediated dystrophies.

Other Genetically Based, Transgenic, and Cell Culture Models There are fewer reports concerning axonal transport in other natural, genetically based neuromuscular diseases of

401

animals. Some early studies investigated axonal transport of radiolabeled protein in nerves of wobbler mice, which develop a motor neuron disease. Forelimb nerves in these mice, which are affected more severely than hind limb nerves, showed reduced fluxes of protein moving by fast and slow anterograde transport.14,47 Such results were in line with expectations for a model of a primary neuronal disorder. Studies on other natural animal models have investigated axonal transport in myelin-deficient mice such as the trembler mouse. These animals, in which cross-transplantation experiments localized the primary disorder to the Schwann cell,1 show severe demyelination in the sciatic nerve. An early study on trembler mice found that rapid axonal transport was at least grossly normal.20 This result was in line with expectations for a model in which the primary deficit was extrinsic to the neuron. The result was also consistent with observations of normal transport after focal demyelination by diphtheria toxin122 or antiserum to galactocerebroside.7 Surprisingly, however, more recent work has shown that a failure of myelination based solely on gene expression in glial cells can affect transport substantially. Thus a new study of trembler mice has found that the Schwann cell abnormality leads to a number of changes in the axons of peripheral nerve. These changes include a reduced density of axonal neurofilaments, a lowered stability of microtubules, and a slowing of axonal transport, among other phenomena.124 Effects in an opposite direction are seen in a model of central demyelination, the shiverer mouse, which is defective in myelin basic protein and expresses little compact myelin in the central nervous system. These animals show clear increases in the density of microtubules and the rate of slow axonal transport in affected brain regions.124 Such findings should be relevant to demyelinating peripheral neuropathies in humans, including those arising from genetic defects in glial cell protein expression and those arising from effects of toxicants and heavy metals on oligodendroglia. Studies of axonal transport in transgenic animals began relatively recently, but they already promise fresh insights into the mechanisms of neurologic disease in humans. Efforts to date have concentrated on degenerative disorders of central and alpha motor neurons. Mouse models of ALS have been particularly informative (for a recent review, see Rao and Nixon184). For example, it has been shown that transgenic mice expressing high levels of the human neurofilament light and heavy subunit proteins (NF-L and NF-H) develop an ALS-like pathology with filamentous aggregates in cell somata and proximal axons.58,243 These same mice exhibit marked retardation in the rates of transport of neurofilament protein and tubulin.57,146 A lesson may be that abnormal processing of cytoskeletal elements is central to ALS. Another likely key to ALS pathology, suggested by genetic studies of patients

402

Neurobiology of the Peripheral Nervous System

and their relations, is the enzyme superoxide dismutase 1 (SOD1), which is mutated in certain familial versions of ALS.186,242 Significantly, mice expressing mutant human SOD1 develop the classic pathology of motor neuron disease.22,230 As one might expect, the transport of neurofilament protein in these animals is disturbed.241,245 More surprisingly, reductions in fast anterograde transport and in axonal levels of kinesin precede the onset of pathology and motor weakness.236 A most recent innovation in the experimental approach to human neurologic disease is the use of Drosophila, an especially powerful model for studies of genetically based disorders. Insights into the neurodegeneration of Huntington’s disease have been obtained from transgenic Drosophila expressing a pathogenic version of the human huntingtin protein with an expanded polyglutamine motif. These flies were found to develop protein aggregates in distal regions of the nerve cell leading to disruption of axonal transport and a host of associated deficits, including loss of motor coordination and decreased lifespan.139

CONCLUSION The stages and major components of the system for axonal transport are indicated diagrammatically in Figure 18–6,

which also summarizes information about the sites where transport is vulnerable to attack by toxicants that induce neuropathy in humans or animals. One clear result of the past decade of studies is the increasing sophistication of our ideas about the relationship between axonal transport and peripheral nerve disease. Outright failure of transport would surely cause rapid degeneration and death of the nerve cell, but it is an unlikely cause of most forms of neuropathy. However, less dramatic abnormalities in the supply and disposition of transported materials occur early in certain nerve diseases and probably contribute to their evolution. Especially in toxic neuropathies, the pathogenic roles of transport abnormalities are beginning to be understood at the physiologic level. The most important pathology of rapid transport appears to involve defects operating in the most distal regions of the axon, during transport reversal or at stages just before or after. In the case of slow transport, there is an emerging consensus that chemical modifications of neurofilament subunits and deficits of filament transport contribute to the development of filamentous neuropathies. The challenge now is to develop fuller biochemical and molecular explanations of all these phenomena and link them to the rapidly expanding understanding of the diversity of transport motors and partner proteins that define the array of phenomena involved in axonal transport.

FIGURE 18–6 Stages in rapid axonal transport and suspected or proven sites of attack by agents capable of inducing peripheral neuropathy. Numbered arrows refer to the following stages: (1) loading of transported protein via the Golgi apparatus; (2) loading that bypasses the Golgi apparatus; (3) microtubule-based anterograde (kinesindriven) transport; (4) turnaround; (5) retrograde (dynein-driven) transport; and (6) lysosomal digestion. Gluc ⫽ glucose; Mitoch ⫽ mitochondrion; MT ⫽ microtubule; MVB ⫽ multivesicular body; NF ⫽ neurofilament; Pyr ⫽ pyruvate; SER ⫽ smooth endoplasmic reticulum; VES ⫽ synaptic vesicle.

Axonal Transport: Properties, Mechanisms, and Role in Nerve Disease

REFERENCES 1. Aguayo, A. J., Attiwell, M., and Trecarten, J.: Abnormal myelination in transplanted trembler mouse Schwann cells. Nature 265:73, 1977. 2. Allen, R. D., Travis, J. L., and Hayden, J. H.: Cytoplasmic transport: moving ultrastructural elements common to many cell types revealed by video-enhanced microscopy. Cold Spring Harb. Symp. Quant. Biol. 46:85, 1982. 3. Allen, R. D., Weiss, D. G., and Hayden, J. H.: Gliding movement of and bidirectional transport along single native microtubules from squid axoplasm: evidence for an active role of microtubules in cytoplasmic transport. J. Cell. Biol. 100:1736, 1985. 4. Almenar-Queralt, A., and Goldstein, L. S.: Linkers, packages and pathways: new concepts in axonal transport. Curr. Opin. Neurobiol. 11:550, 2001. 5. Anderson, D. R.: Axonal transport in the retina and optic nerve. In Glaser, J. S. (ed.): Neuro-Ophthalmology. St. Louis, C. V. Mosby, p. 140, 1977. 6. Anthony, D. C., Boekelheide, K., and Graham, D. G.: The effect of 3,4-dimethyl substitution on the neurotoxicity of 2,5-hexanedione. I. Accelerated clinical neuropathy is accompanied by more proximal axonal swellings. Toxicol. Appl. Pharmacol. 71:362, 1983. 7. Armstrong, R., Toews, A. D., and Morell, P.: Rapid axonal transport in focally demyelinated sciatic nerve. J. Neurosci. 7:4044, 1987. 8. Asmundson, V. S., Kratzer, F. H., and Julian, L. M.: Inherited myopathy in the chicken. Ann. N. Y. Acad. Sci. 138:49, 1966. 9. Baas, P. E.: Microtubule transport in the axon. Int. Rev. Cytol. 212:41, 2002. 10. Bajada, S., Sharma, A. K., and Thomas, P. K.: Axoplasmic transport in vagal afferent fibers in normal and alloxandiabetic rabbits. J. Neurol. Sci. 47:365, 1980. 11. Banks, P., Mayor, D., and Tomlinson, D.: Further evidence for the involvement of microtubules in the intra-axonal movement of noradrenaline storage vesicles. J. Physiol. (Lond.) 219:755, 1971. 12. Banks, P., and Till, R.: A correlation between the effects of antimitotic drugs on microtubule assembly in vitro and the inhibition of axonal transport in noradrenergic neurons. J. Physiol. (Lond.) 252:283, 1975. 13. Behrens, M. I., Torrealba, G., Court J., et al.: Axonal transport dysfunction in dystrophia myotonica. Acta Neuropathol. (Berl.) 62:157, 1983. 14. Bird, M. T., Shuttleworth, J. R., Koestner, A., and Reinglass, J.: The wobbler mouse mutant: an animal model of hereditary motor system disease. Acta Neuropathol. (Berl.) 19:39, 1971. 15. Bisby, M. A.: Orthograde and retrograde axonal transport of labeled protein in motoneurons. Exp. Neurol. 50:628, 1976. 16. Bisby, M. A.: Retrograde axonal transport of endogenous protein: differences between motor and sensory axons. J. Neurochem. 28:249, 1977. 17. Bisby, M. A.: Axonal transport of labelled protein and regeneration rate in nerves of streptozotocin-diabetic rats. Exp. Neurol. 69:74, 1980.

403

18. Bisby, M. A.: Retrograde axonal transport. Adv. Cell Neurobiol. 1:69, 1980. 19. Blakemore, W. F., and Cavanagh, J. B.: “Neuroaxonal dystrophy” occurring in the experimental “dying back” process in the rat. Brain 92:789, 1969. 20. Boegman, R. J., Aguayo, A. J., and Bray, G. M.: Axoplasmic transport in (trembler mouse) nerves with a widespread disorder of myelination. J. Neuropathol. Exp. Neurol. 36:590, 1977. 21. Boegman, R. J., and Wood, P. L.: Axonal transport in dystrophic hamsters. Can. J. Physiol. 59:202, 1981. 22. Borchelt, D. R., Wong, P. C., Becher, M. W., et al.: Axonal transport of mutant superoxide dismutase 1 and focal axonal abnormalities in the proximal axons of transgenic mice. Neurobiol. Dis. 5:27, 1998. 23. Bowen, D. M., Smith, C. B., White, P., and Davison, A. N.: Neurotransmitter-related enzymes and indices of hypoxia in senile dementia and other abiotrophies. Brain 99:459, 1976. 24. Bradley, W. G., and Jaros, E.: Axoplasmic flow in axonal neuropathies. II. Axoplasmic flow in mice with motor neuron: disease and muscular dystrophy. Brain 96:247, 1973. 25. Bradley, W. G., and Jenkinson, M.: Abnormalities of peripheral nerves in murine muscular dystrophy. J. Neurol. Sci. 18:227, 1973. 26. Bradley, W. G., and Williams, M. H.: Axoplasmic flow in axonal neuropathies. I. Axoplasmic flow in cats with toxic neuropathies. Brain 96:235, 1973. 27. Brady, S. T., and Lasek, R. J.: Fast axonal transport in extruded axoplasm from squid giant axon. Science 218:1129, 1982. 28. Brady, S. T., Lasek, R. J., and Allen, R. D.: Video microscopy of fast axonal transport in extruded axoplasm: A new model for study of molecular mechanisms. Cell Motil. 5:81, 1985. 29. Braendgaard, H., and Sidenius, P.: The retrograde fast component of axonal transport in motor and sensory nerves of the rat during administration of 2,5-hexanedione. Brain Res. 378:1, 1986. 30. Brat, D., and Brimijoin, S.: Acrylamide and glycidamide impair neurite outgrowth in differentiating N1E.115 neuroblastoma without disturbing rapid bidirectional transport of organelles observed by video microscopy. J. Neurochem. 60:2145, 1993. 31. Brat, D., Windebank, A., and Brimijoin, S.: Emulsifier for intravenous cyclosporin inhibits neurite outgrowth, causes deficits in rapid axonal transport and leads to structural abnormalities in differentiating N1E.115 neuroblastoma. J. Pharmacol. Exp. Ther. 261:803, 1992. 32. Bray, J. J., Kon, E. M., and Breckenridge, B. M.: Reversed polarity of rapid axonal transport in chicken motoneurons. Brain Res. 33:560, 1971. 33. Breuer, A. C., and Atkinson, M. B.: Fast axonal transport alterations in amyotrophic lateral sclerosis (ALS) and in parathyroid hormone (PTH)-treated axons. Cell Motil. Cytoskeleton 10:321, 1988. 34. Breuer, A. C., Lynn, M. B., and Atkinson, M. B., et al.: Fast axonal transport in amyotrophic lateral sclerosis: an intraaxonal organelle traffic analysis. Neurology 37:738, 1987. 35. Brimijoin, S.: Stop-flow: a new technique for measuring axonal transport, and its application to the transport of dopamine-␤-hydroxylase. J. Neurobiol. 6:379, 1975.

404

Neurobiology of the Peripheral Nervous System

36. Brimijoin, S.: A histofluorescence study of events accompanying accumulation and migration of norepinephrine within locally cooled nerves. J. Neurobiol. 8:251, 1977. 37. Brimijoin, S.: On the kinetics and maximal capacity of the system for rapid axonal transport in mammalian neurones. J. Physiol.(Lond.) 292:325, 1979. 38. Brimijoin, S.: Axonal transport in autonomic nerves: views on its kinetics. In Kalsner, S. (ed.): Trends in Autonomic Pharmacology. Baltimore, Urban and Schwarzenberg, p. 17, 1982. 39. Brimijoin, S.: Microtubules and the capacity of the system for rapid axonal transport. Fed. Proc., 41:2312, 1982. 40. Brimijoin, S., Capek, P., and Dyck, P. J.: Axonal transport of dopamine-beta-hydroxylase by human sural nerves in vitro. Science 180:1295, 1973. 41. Brimijoin, S., and Dyck, P. J.: Axonal transport of dopamine-beta-hydroxylase and acetylcholinesterase in human peripheral neuropathy. Exp. Neurol. 66:467, 1979. 42. Brimijoin, S., Dyck, P. J., Jakobsen, J., and Lambert, E. H.: Axonal transport in human nerve disease and in the experimental neuropathy induced by p-bromophenylacetylurea. In Weiss, D.G., and Gorio, A. (eds.): Axoplasmic Transport in Physiology and Pathology. Berlin, Springer-Verlag, p. 124, 1982. 43. Brimijoin, S., and Jablecki, C. K.: Reduced axonal transport of dopamine-beta-hydroxylase in dystrophic mice: evidence for abnormality of adrenergic nerve cells. Exp. Neurol. 53:454, 1976. 44. Brimijoin, S., Lundberg, J. M., Brodin, E., et al.: Axonal transport of substance P in the vagus and sciatic nerves of the guinea pig. Brain Res., 191:443, 1980. 45. Brimijoin, S., Olsen, J., and Rosenson, R.: Comparison of the temperature-dependence of rapid axonal transport and microtubules in nerves of the rabbit and bullfrog. J. Physiol. (Lond.), 287:303, 1979. 46. Brimijoin, S., and Wiermaa, M. J.: Direct comparison of the rapid axonal transport of norepinephrine and dopaminebeta-hydroxylase activity. J. Neurobiol. 8:239, 1977. 47. Brimijoin, S., and Wiermaa, M. J.: Rapid axonal transport of tyrosine hydroxylase in rabbit sciatic nerves. J. Physiol. (Lond.) 120:77, 1977. 48. Brimijoin, S., and Wiermaa, M. J.: Rapid orthograde and retrograde axonal transport of acetylcholinesterase as characterized by the stop-flow technique. J Physiol (Lond.) 285:129, 1978. 49. Brown, A.: Axonal transport of membranous and nonmembranous cargoes: a unified perspective. J. Cell Biol. 160:817, 2003. 50. Brownlees, J., Ackerley, S., Grierson, A. J., et al.: CharcotMarie-Tooth disease neurofilament mutations disrupt neurofilament assembly and axonal transport. Hum. Mol. Genet. 11:2837, 2002. 51. Burton, P. R., and Paige, J. L.: Polarity of axoplasmic microtubules in the olfactory nerve of the frog. Proc. Natl. Acad. Sci. U. S. A. 78:3269, 1981. 52. Cai, F., and Helke, C. J.: Abnormal PI3 kinase/Akt signal pathway in vagal afferent neurons and vagus nerve of streptozotocin-diabetic rats. Mol. Brain. Res. 110:234, 2003. 53. Carpenter, S.: Proximal axonal enlargement in motor neuron disease. Neurology, 18:841, 1968.

54. Cavanagh, J. B.: The significance of the “dying back” process in experimental and human neurological diseases. Int. Rev. Exp. Pathol. 3:219, 1964. 55. Cavanagh, J. B., Chen, F. C. K., Kyu, M., H., and Ridley, A.: The experimental neuropathy in rats caused by p-bromophenylacetylurea. J. Neurol. Neurosurg. Psychiatry 31:471, 1968. 56. Chretien, M., Patey, G., Souyri, F., and Droz, B.: “Acrylamide-induced” neuropathy and impairment of axonal transport of proteins. II. Abnormal accumulations of smooth endoplasmic reticulum at sites of focal retention of fast transported proteins: electron microscope radioautographic study. Brain Res. 205:15, 1981. 57. Collard, J. F., Cote, F., and Julien, J. P.: Defective axonal transport in a transgenic mouse model of amyotrophic lateral sclerosis. Nature 375:61, 1995. 58. Cote, F., Collard, J. F., and Julien, J. P.: Progressive neuronopathy in transgenic mice overexpressing the human neurofilament heavy gene: a mouse model of amyotropic lateral sclerosis. Cell 73:35, 1993. 59. Dahlström, A.: Effect of colchicine on transport of amine storage granules in sympathetic nerves of the rat. Eur. J. Pharmacol. 5:111, 1968. 60. Dahlström, A.: Axoplasmic transport (with particular respect to adrenergic neurons). Philos. Trans. R. Soc. Lond. [Biol.] 261:325, 1971. 61. De Jonghe, P., Mersivanova, I., Nelis, E., et al.: Further evidence that neurofilament light chain gene mutations can cause Charcot-Marie-Tooth disease type 2E. Ann. Neurol. 49:245, 2001. 62. Delcroix, J. D., Michael, G. J., Priestley, J. V., et al.: Effect of nerve growth factor treatment on p75NTR gene expression in lumbar dorsal root ganglia of streptozocin-induced diabetic rats. Diabetes 47:1779, 1998. 63. Delcroix, J. D., Tomlinson, D. R., and Fernyhough, P.: Diabetes and axotomy-induced deficits in retrograde axonal transport of nerve growth factor correlate with decreased levels of p75lntr protein in lumbar dorsal root ganglia. Mol. Brain Res. 51:82, 1997. 64. Dent, W. W., Calloway, J. L., Szebenyi, G., et al.: Reorganization and movement of microtubules in axonal growth cones and developing interstitial branches. J. Neurosci. 19:8894, 1999. 65. DeSantis, M., Hoekman, T., and Limwongse, V.: Retrograde transport of peroxidase in motor neurons innervating slow and fast twitch muscles: absence of influence between normal and dystrophic chickens. Brain Res. 119:454, 1977. 66. Diezel, P. B., and Quadbeck, G.: Nervenschudigung durch p-Bromophenylacetyl-Harnstoff. Naunyn Schmiedebergs Arch. Exp. Pathol. Pharmakol. 238:534, 1960. 67. DiGiamberardino, L., Couraud, J. Y., and Barnard, E. A.: Normal axonal transport of acetylcholinesterase forms in peripheral nerves of dystrophic chickens. Brain Res. 160:196, 1979. 68. Droz, B., Brunetti, M., and Giamberardino, L. D.: Selective distribution of axonally transported phospholipids to nerve endings and/or myelin: the case of ethanolamine glycerophospholipids. In Horrocks, L. A. (ed.): Phospholipids in the Nervous System. New York, Raven Press, p. 315, 1985.

Axonal Transport: Properties, Mechanisms, and Role in Nerve Disease 69. Droz, B., Giamberardino, L. D., and Koenig, H. L.: Contribution of axonal transport to the renewal of myelin phospholipids in peripheral nerves. I. Quantitative radioautographic study. Brain Res. 219:57, 1981. 70. Dustin, P., Microtubules. In Dustin, P. (ed.): Microtubules, 2nd ed. New York, Springer-Verlag, p. 1, 1984. 71. Dyck, P. J., Conn, D. L., and Okazaki, H.: Necrotizing angiopathic neuropathy: Three dimensional morphology of fiber degeneration related to sites of occluded vessels. Mayo Clin. Proc. 47:461, 1972. 72. Edstrom, A., and Hanson, M.: Retrograde axonal transport of proteins in vitro in frog sciatic nerves. Brain Res. 6:311, 1973. 73. Edström, A., Kanje, M., and Rusovan, A.: Orthograde and retrograde axonal transport in the regenerating frog sciatic nerve show different sensitivities to vanadate. Acta Physiol. Scand. 134:437, 1988. 74. Erdmann, G., Wiegand, H., and Wellhoner, H. H.: Intraaxonal and extraaxonal transport of 125I-tetanus toxin in early local tetanus. Naunyn Schmiedebergs Arch. Pharmacol. 290:357, 1975. 75. Ernfors, P., Lee, K. F., Kucera, J., and Jaenisch, R.: Lack of neurotrophin-3 leads to deficiencies in the peripheral nervous system and loss of limb proprioceptive afferents. Cell 77:503, 1994. 76. Eyer, J., McLean, W. G., and Leterrier, J.-F.: Effect of a single dose of ␤,␤⬘-iminodipropionitrile in vivo on the properties of neurofilaments in vitro: comparison with the effect of iminodipropionitrile added directly to neurofilaments in vitro. J. Neurochem. 52:1759, 1989. 77. Fariñas, I., Jones, K. R., Backus, C., et al.: Severe sensory and sympathetic deficits in mice lacking neurotrophin-3. Nature 369:658, 1994. 78. Fernyhough, P., Diemel, L. T., Brewster, W. J., and Tomlinson, D. R.: Deficits in sciatic nerve neuropeptide content coincide with a reduction in target tissue nerve growth factor mRNA in streptozotocin-diabetic rats: effects of insulin treatment. Neuroscience 62:337, 1994. 79. Fernyhough, P., Diemel, L. T., and Tomlinson, D. R.: Target tissue production and axonal transport of neurotrophin-3 are reduced in streptozotocin-diabetic rats. Diabetologia 41:300, 1998. 80. Forman, D. S.: A permeabilized cell model of saltatory organelle movement. J. Cell Biol. 91:414a, 1981. 81. Forman, D. S., Brown, K. J., and Livengood, D. R.: Fast axonal transport in permeabilized lobster giant axons is inhibited by vanadate. J. Neurosci. 3:1279, 1983. 82. Forman, D. S., Brown, K. J., and Promersberger, M. E.: Selective inhibition of retrograde axonal transport by erythro9[3-(2-hydroxy-nonyl)]adenine. Brain Res. 272:194, 1983. 83. Forman, D. S., Padjen, A. L., and Siggins, G. R.: Axonal transport of organelles visualized by light microscopy: cinemicrographic and computer analysis. Brain Res. 136:197, 1977. 84. Ghetti, B., and Ochs, S.: On the relation between microtubule density and axoplasmic transport in nerves treated with maytansine. In Canal, N., and Pozza, G. (eds.): Peripheral Neuropathies. Amsterdam, Elsevier, p. 177, 1978. 85. Goldberg, D. J., Schwartz, J. H., and Sherbany, A. A.: Kinetic properties of normal and perturbed axonal transport of serotonin in a single identified axon. J. Physiol. (Lond.) 281:559, 1978.

405

86. Goldstein, S. B., and Yang, Z. H.: Microtubule-based transport systems in neurons: the roles of kinesins and dyneins. Annu. Rev. Neurosci. 23:39, 2000. 87. Grafstein, B., and Forman, D.: Intracellular transport in neurons. Physiol. Rev. 60:1167, 1980. 88. Graham, D. G., Anthony, D. C., and Boekelheide K.: Studies of the molecular pathogenesis of hexane neuropathy II. Evidence that pyrrole derivatization of lysyl residues leads to protein crosslinking. Toxicol. Appl. Pharmacol. 64:629, 1982. 89. Graham, D. G., Szakal-Quin, G., Priest, J. W., and Anthony, D. C.: In vitro evidence that covalent crosslinking of neurofilaments occurs in gamma-diketone neuropathy. Proc. Natl. Acad. Sci. U. S. A. 81:4979, 1984. 90. Green, L. S., Donoso, J. A., Heller-Bettinge, I. E., and Samson, F. E.: Axonal transport disturbances in vincristineinduced peripheral neuropathy. Ann. Neurol. 1:255, 1977. 91. Griffin, J. W., Pahnestock, K. E., Price, D. L., and Hoffman, P. N.: Microtubule-neurofilament segregation produced by ␤,␤⬘-iminodipropionitrile: evidence for association of fast axonal transport with microtubules. J. Neurosci. 3:557, 1983. 92. Griffin, J. W., Gold, B. G., and Cork, L. C.: IDPN neuropathy in the cat: coexistence of proximal and distal axonal swellings. Neuropathol. Appl. Neurobiol. 8:351, 1982. 93. Griffin, J. W., Hoffman, P. N., and Clark, A. W.: Slow axonal transport of neurofilament proteins: impairment by ␤,␤⬘iminodipropionitrile administration. Science 202:633, 1978. 94. Griffin, J. W., and Price, D. L.: Proximal axonopathies induced by toxic chemicals. In Spencer, P. S., and Schaumberg, H. H. (eds.): Experimental and Clinical Neurotoxicology. Baltimore, Williams & Wilkins, p. 161, 1980. 95. Griffin, J. W., Price, D. L., Hoffman, P. N., and Cork, L. C.: The axonal cytoskeleton: alterations of organization and axonal transport in models of neurofibrillary pathology [abstract]. J. Neuropathol. Exp. Neurol. 40:316, 1981. 96. Griffin, J. W., Price, D. L., and Spencer, P. S.: Fast axonal transport through giant axonal swellings in hexacarbon neuropathy. J. Neuropathol. Exp. Neurol. 36:603, 1977. 97. Gupta, R. P., Abdel-Rahman, A., Wilmarth, K. W., and Abou-Donia, M. B.: Alteration in neurofilament axonal transport in the sciatic nerve of the diisopropyl phosphorofluoridate (DFP)-treated hen. Biochem. Pharmacol. 53:1799, 1997. 98. Hahnenberger, R. W.: Effects of pressure on fast axoplasmic flow: an in vitro study in the vagus nerve of rabbits. Acta Physiol. Scand. 104:299, 1978. 99. Hahnenberger, R. W.: Effect of a pressure barrier on retrograde axoplasmic transport in vitro: a study in the motor neurons of the rabbit vagus. Acta Physiol. Scand. 108:133, 1980. 100. Hammerschlag, R.: The role of calcium in the initiation of fast axonal transport. Fed. Proc. 39:2809, 1980. 101. Hammerschlag, R., Bakhit, C., and Chiu, A. Y.: Role of calcium in the initiation of fast axonal transport of protein-effects of divalent cations. J. Neurobiol. 8:439, 1977. 102. Hammerschlag, R., Bolen, F. A., and Stone, G. C.: Inhibition of fast axonal transport by a sodium ionophore. Trans. Am. Soc. Neurochem. 11:143, 1980. 103. Hammerschlag, R., Chiu, A. Y., and Dravid, A. R.: Inhibition of fast axonal transport of [3H]-protein by cobalt ions. Brain Res. 114:353, 1976.

406

Neurobiology of the Peripheral Nervous System

104. Hammerschlag, R., Stone, G. C., and Bolen, F. A.: Evidence that all newly synthesized proteins destined for fast axonal transport pass through the Golgi apparatus. J. Cell Biol. 93:568, 1982. 105. Hansen, T., Andersen, C. B., Echwald, S. M., et al.: Identification of a common amino acid polymorphism in the p85alpha regulatory subunit of phosphatidylinositol 3-kinase: effects on glucose disappearance constant, glucose effectiveness, and the insulin sensitivity index. Diabetes 46:494, 1997. 106. Harris, C. H., Gulati, A. K., Friedman, M. A., and Sickles, D. W.: Toxic neurofilamentous axonopathies and fast anterograde axonal transport. Part V. Reduced bi-directional vesicle transport in cultured neurons by acrylamide and glycidamide. J. Toxicol. Environ. Health 42:343, 1994. 107. Heidemann, S. R., and McIntosh, J. R.: Visualization of the structural polarity of microtubules. Nature 286:517, 1980. 108. Helland, L.: Rapid retrograde transport of dopamine␤-hydroxylase as examined by the stop-flow technique. Brain Res. 102:217, 1976. 109. Hirano, A., Donnenfeld, H., Sasaki, S., and Nakano, I.: Fine structural observations of neurofilamentous changes in amyotrophic lateral sclerosis. J. Neuropathol. Exp. Neurol. 43:461, 1984. 110. Hoffman, P. N., and Lasek, R. J.: The slow component of axonal transport: identification of major structural polypeptides of the axon and their generality among mammalian neurons. J. Cell Biol. 66:351, 1975. 111. Homburger, F., Nixon, C. W., Eppenberger, M., and Baker, J.: Hereditary myopathy in the Syrian hamster: studies on pathogenesis. Ann. N. Y. Acad. Sci. 138:14, 1966. 112. Jablecki, C. K., and Brimijoin, S.: Reduced axoplasmic transport of choline acetyltransferase activity in dystrophic mice. Nature 250:151, 1974. 113. Jakobsen, J.: Axonal dwindling in early experimental diabetes. Diabetologia 12:539, 1976. 114. Jakobsen, J.: Axonal dwindling in early experimental diabetes. II. A study of isolated nerve fibers. Diabetologia 12:547, 1976. 115. Jakobsen, J., and Brimijoin, S.: Axonal transport of enzymes and labeled proteins in experimental axonopathy induced by p-bromophenylacetylurea. Brain Res. 229:103, 1981. 116. Jakobsen, J., Brimijoin, S., Skau, K., and Wells, D.: Retrograde axonal transport of transmitter enzymes, fucose labeled proteins and nerve growth factor in streptozotocindiabetic rats. Diabetes 30:797, 1981. 117. Jakobsen, J., Lambert, E., Carlson, G., and Brimijoin, S.: Clinical and electrophysiological characteristics of the experimental neuropathy caused by p-bromophenylacetylurea. Exp. Neurol. 75:158, 1982. 118. Jakobsen, J., and Sidenius, P.: Decreased axonal flux of retrogradely transported glycoproteins in early experimental diabetes. J. Neurochem. 33:1055, 1979. 119. Jakobsen, J., and Sidenius, P.: Decreased axonal transport of structural proteins in streptozotocin diabetic rats. J. Clin. Invest. 66:292, 1980. 120. Jakobsen, J., and Sidenius, P.: Early and dose-dependent decrease in retrograde axonal transport in acrylamide-intoxicated rats. J. Neurochem. 40:447, 1983.

121. Kamal, A., Stokin, G. B., Yang, Z., et al.: Axonal transport of amyloid precursor protein is mediated by direct binding to the kinesin light chain subunit of kinesin-1. Neuron 28:449, 2000. 122. Kidman, A., Hanwell, M., and Cooper, N.: Failure of diphtheritic demyelination to block slow axonal transport in chicken sciatic nerve. J. Neurochem. 33:357, 1979. 123. Kirkpatrick, J. B., and Stern, L. Z.: Axoplasmic flow in human sural nerve. Arch. Neurol. 28:308, 1973. 124. Kirkpatrick, L. L., Witt, A. S., Payne, H. R., et al.: Changes in microtubule stability and density in myelin-deficient shiverer mouse CNS axons. J. Neurosci. 21:2288, 2001. 125. Kirpekar, S. M., Prat, J. C., and Wakade, A. R.: Metabolic requirements for the intraaxonal transport of noradrenaline in the cat hypogastric nerve. J. Physiol. (Lond.) 228:173, 1973. 126. Klein, R., Silos-Santiago, I., and Smeyne, R. J.: Disruption of the neurotophin-3 receptor gene TrkC eliminates 1a muscle afferents and results in abnormal movements. Nature 368:249, 1994. 127. Komiya, Y., Cooper, N. A., and Kidman, A. D.: The long-term effects of a single injection of ␤,␤⬘-iminodipropionitrile on slow axonal transport in the rat. J. Biochem. 100:1241, 1986. 128. Komiya, Y., and Austin, L.: Axoplasmic flow of protein in the sciatic nerve of normal and dystrophic mice. Exp. Neurol. 43:1, 1974. 129. Korthals, J. K., Korthals, M. A., and Wisniewski, H. M.: Peripheral nerve ischemia: part 2. Accumulation of organelles. Ann. Neurol. 4:487, 1978. 130. Kreutzberg, G. W.: Neuronal dynamics and axonal flow IV. Blockage of intra-axonal enzyme transport by colchicine. Proc. Natl. Acad. Sci. U. S. A. 62:722, 1969. 131. Kreutzberg, G. W., and Emmert, H.: Glucose utilization of motor nuclei during regeneration: a [14C] 2-deoxyglucose study. Exp. Neurol. 70:712, 1980. 132. Kristensson, K.: Retrograde transport of macromolecules in axons. Ann. Rev. Pharmacol. Toxicol. 18:97, 1978. 133. Kristensson, K., Lycke, E., and Sjostrand, J.: Spread of herpes simplex virus in peripheral nerves. Acta Neuropathol. (Berl.) 19:44, 1971. 134. Kuffer, A. D., Komiya, Y., and Austin, L.: Proteins of fast axoplasmic transport in the sciatic nerve of the dystrophic mouse. Exp. Neurol. 55:74, 1977. 135. Larsen, J. R., and Sidenius, P.: Slow axonal transport of structural polypeptides in rat, early changes in streptozocin diabetes, and effect of insulin treatment. J. Neurochem. 52:390, 1989. 136. Lasek, R. J.: Bidirectional transport of radioactively labelled axoplasmic components. Nature 216:1212, 1967. 137. Lasek, R. J., and Hoffman, P. N.: The neuronal cytoskeleton, axonal transport and axonal growth. In Goldman, R., Pollard, T., and Rosenbaum, J. (eds.): Cell Motility (Cold Spring Harbor Conferences on Cell Proliferation Series). Cold Spring Harbor, NY, Cold Spring Harbor Press, p. 1021, 1976. 138. Lee, P. G., Hohman, T. C., Cai, F., et al.: Streptozotocininduced diabetes causes metabolic changes and alterations in neurotrophin content and retrograde transport in the cervical vagus nerve. Exp. Neurol. 170:149, 2001. 139. Lee, W. C., Yoshihara, M., and Littleton, J. T.: Cytoplasmic aggregates trap polyglutamine-containing proteins and block axonal transport in a Drosophila model of Huntington’s disease. Proc. Natl. Acad. Sci. U. S. A. 101:3224, 2004.

Axonal Transport: Properties, Mechanisms, and Role in Nerve Disease 140. Leone, J., and Ochs, S.: Anoxic block and recovery of axoplasmic transport and electrical excitability of nerve. J. Neurobiol. 9:229, 1978. 141. Liwnicz, B. H., Kristensson, I., and Wisniewski, H. M.: Observations on axoplasmic transport in rabbits with aluminum-induced neurofibrillary tangles. Brain Res. 80:413, 1974. 142. LoPachin, R. M.: The role of fast axonal transport in acrylamide pathophysiology: mechanism of epiphenomenon? Neurotoxicology 23:253, 2002. 143. Lubinska, L.: Axoplasmic streaming in regenerating and in normal nerve fibres. Prog. Brain Res. 13:1, 1964. 144. Maeda, K., Fernyhough, P., and Tomlinson, D. R.: Regenerating sensory neurones of diabetic rats express reduced levels of mRNA for GAP-43, gamma-preprotachykinin and the nerve growth factor receptors, trkA and p75NGFR. Mol. Brain Res. 37:166, 1996. 145. Martz, D., Gamer, J., and Lasek, R. J.: Protein changes during anterograde-to-retrograde conversion of axonally transported vesicles. Brain Res. 476:199, 1989. 146. Marszalek, J. R., Williamson, T. L., Lee, M. K., et al.: Neurofilament subunit H modulates axonal diameter by affecting the rate of neurofilament transport. J. Cell Biol. 135:711, 1996. 147. Mayer, J. H., and Tomlinson, D. R.: Prevention of defects of axonal transport and nerve conduction velocity by oral administration of myo-inositol or an aldose reductase inhibitor in streptozotocin-diabetic rats. Diabetologia 25:433, 1983. 148. McComas, A. J., Sica, R. E. P., and Currie, S.: Muscular dystrophy: evidence for a neural factor. Nature 226:1263, 1970. 149. McEwen, B. S., Forman, D. S., and Grafstein, B.: Components of fast and slow axonal transport in the goldfish optic nerve. J. Neurobiol. 2:361, 1971. 150. McLane, J., and McClure, W. O.: Rapid axoplasmic transport in dystrophic mice. J. Neurochem. 29:863, 1977. 151. McNiven, M. A., and Ward, J. B.: Calcium regulation of pigment transport in vitro. J. Cell Biol. 106:111, 1988. 152. Mendell, J. R., Sahenk, Z., and Saida, K.: Alterations of fast axoplasmic transport in experimental methyl n-butyl ketone neuropathy. Brain Res. 133:107, 1977. 153. Michelson, A. M., Russell, E. S., and Pinckney, J. H.: Dystrophia muscularis: a hereditary primary myopathy in the house mouse. Proc. Natl. Acad. Sci. U. S. A. 41:1079, 1955. 154. Miller, C. C., Ackerley, S., Brownlees, J., et al.: Axonal transport of neurofilaments in normal and disease states. Cell. Mol. Life Sci. 59:323, 2002. 155. Moretto, A., Capodicasa, E., Peraica, M., and Lotti, M.: Age sensitivity to organophosphate-induced delayed polyneuropathy: Biochemical and toxicological studies in developing chicks. Biochem. Pharmacol. 41:1497, 1991. 156. Morgan, G., and Wilbourn, A. J.: Cervical radiculopathy and coexisting distal entrapment neuropathies: doublecrush syndromes? Neurology 50:78, 1998. 157. Nagata, H., and Brimijoin, S.: Axonal transport in the motor neurons of rats with neuropathy induced by p-bromophenylurea. Ann. Neurol. 19:458, 1986. 158. Nagata, H., and Brimijoin, S.: Neurotoxicity of halogenated pheylacetylureas is linked to abnormal onset of rapid axonal transport. Brain Res. 385:136, 1986.

407

159. Nagata, H., Brimijoin, S., Low, P., and Schmelzer, J. D.: Slow axonal transport in experimental hypoxia and in neuropathy induced by p-bromophenylacetylurea. Brain Res. 422:319, 1987. 160. Nagata, H., Ohkoshi, N., Kanazawa, J., et al.: Rapid axonal transport velocity is reduced in experimental ethylene oxide neuropathy. Mol. Chem. Neuropathol. 17:209, 1992. 161. Nordberg, A., Hellstrom-Lindahl, E., Lee, M., et al.: Chronic nicotine treatment reduces b-amyloidosis in the brain of a mouse model of Alzheimer’s disease (APPsw). J. Neurochem. 81:655, 2002. 162. Ochs, S.: Local supply of energy to the fast axoplasmic transport mechanism. Proc. Natl. Acad. Sci. U. S. A. 68:1279, 1971. 163. Ochs, S.: Rate of fast axoplasmic transport in mammalian nerve fibres. J. Physiol. (Lond.) 227:627, 1972. 164. Ochs, S.: Effect of maturation and aging on the rate of fast axoplasmic transport in mammalian nerve. Prog. Brain Res. 40:349, 1973. 165. Ochs, S.: Axoplasmic transport—energy metabolism and mechanism. In Hubbard, J. I. (ed.): The Vertebrate Peripheral Nervous System. New York, Plenum Press, p. 47, 1974. 166. Ochs, S.: Retention and redistribution of proteins in mammalian nerve fibres by axoplasmic transport. J. Physiol. (Lond.) 253:459, 1975. 167. Ochs, S.: A unitary concept of axoplasmic transport based on the transport filament hypothesis. In Bradley, W. G., Gardner-Medwin, D., and Walton, J. N., (eds.): Third International Congress on Muscle Diseases. Amsterdam, Excerpta Medica, p. 189, 1975. 168. Ochs, S.: Axoplasmic Transport and Its Relation to Other Nerve Functions. New York, Wiley Interscience, 1982. 169. Ochs, S., and Hollingsworth, D.: Dependence of fast axoplasmic transport in nerve on oxidative metabolism. J. Neurochem. 18:107, 1971. 170. Ochs, S., and Jersild, R. A. Jr.: Fast axoplasmic transport in nonmyelinated mammalian nerve fibers shown by electron microscopic radioautography. J. Neurobiol. 5:373, 1974. 171. Ochs, S., and Smith, C. B.: Fast axoplasmic transport in mammalian nerve in vitro after block of glycolysis with iodoacetic acid. J. Neurochem. 18:833, 1971. 172. Ochs, S., and Smith, C. B.: Low temperature slowing and cold-block of fast axoplasmic transport in mammalian nerves in vitro. J. Neurobiol. 6:85, 1975. 173. Ochs, S., Worth, R. M., and Chan, S. Y.: Calcium requirement for axoplasmic transport in mammalian nerve. Nature 270:748, 1977. 174. Ohnishi, A., Schilling, K., Brimijoin, S., et al.: Lead neuropathy. I. Morphometry, nerve conduction and choline acetyltransferase transport: new finding of endoneurial edema associated with segmental demyelination. J. Neuropathol. Exp. Neurol. 36:499, 1977. 175. Oka, N., and Brimijoin, S.: Tubulomembranous lesions in BPAU neuropathy reflect local stasis of fast axonal transport: evidence from electronmicroscope autoradiography. Mayo Clin. Proc. 67:341, 1992. 176. Otsuka, A. J., Jeyaprakash, A., Garcia-Anoveros, J., et al.: The C. elegans unc-104 gene encodes a putative kinesin heavy chain-like protein. Neuron 6:113, 1991. 177. Padilla, S., Atkinson, M. B., and Breuer, A. C.: Direct measurement of fast axonal organelle transport in the sciatic

408

178.

179.

180.

181.

182.

183. 184.

185.

186.

187.

188.

189.

190.

191.

192.

193.

194.

195.

Neurobiology of the Peripheral Nervous System nerve of rats treated with acrylamide. J. Toxicol. Environ. Health 39:429, 1993. Paschal, B. M., and Vallee, R. B.: Retrograde transport by the microtubule associated protein MAP IC. Nature 330:181, 1987. Pierce, D. W., Hom-Booher, N., Otsuka, A. J., and Vale, R. D.: Single-molecule behavior of monomeric and heteromeric kinesins. Biochemistry 38:5412, 1999. Pleasure, D. E., Mishler, K. C., and Engel, W. K.: Axonal transport of proteins in experimental neuropathies. Science 166:524, 1969. Price, D. L., Griffin, J., and Young, A.: Tetanus toxin: direct evidence for retrograde intraaxonal transport. Science 88:945, 1975. Ramon y Cajal, S.: Degeneration and Regeneration of the Nervous System (R. M. May, transl.). London, Oxford University Press, 1928. Ranvier, M. L. L.: Leçons sur I’Histologie du Systeme Nerveux. Paris, Librairie F. Savy, p. 68, 1878. Rao, M. V., and Nixon, R. A.: Defective neurofilament transport in mouse models of amyotrophic lateral sclerosis: a review. Neurochem. Res. 28:1041, 2003. Reynolds, A. J., Bartlett, S. E., and Hendry, I. A.: Signaling events regulating the retrograde axonal transport of 125I-beta nerve growth factor in vivo. Brain Res. 798:67, 1998. Rosen, D. R., Siddique, T., Patterson, D., et al.: Mutation in Cu/Zn superoxide dismutase gene are associated with familial amyotrophic lateral sclerosis. Nature 362:59, 1993. Roy, S., Coffee, P., Smith, G., et al.: Neurofilaments are transported rapidly but intermittently in axons: implications for slow axonal transport. J. Neurosci. 20:6849, 2000. Rydevik, B., McLean, W. G., Sjöstrand, J., and Lundborg, G.: Blockage of axonal transport induced by acute graded compression of the rabbit vagus nerve. J. Neurol. Neurosurg. Psychiatry 43:690, 1980. Sabri, M. I.: Further observations on in vitro and in vivo effects of 2,5-hexanedione on giyceraldehyde-3-phosphate dehydrogenase. Arch. Toxicol. 55:191, 1984. Sabri, M. I.: In vitro effect of n-hexane and its metabolites on selected enzymes in glycolysis, pentose phosphate pathway and citric acid cycle. Brain Res. 297:145, 1984. Sabri, M. I., Moore, C., and Spencer, P. S.: Studies on the biochemical basis of distal axonopathies. I. Inhibition of glyceraldehyde-3-phosphate dehydrogenase by neurotoxic hexacarbon compounds. J. Neurochem. 32:683, 1979. Sabri, M. I., and Ochs, S.: Relation of ATP and creatine phosphate to fast axoplasmic transport in mammalian nerve. J. Neurochem. 19:2821, 1972. Sabri, M. I., and Spencer, P. S.: Toxic distal axonopathy: biochemical studies and hypothetical mechanisms. In Spencer, P. S., and Schaumberg, H. H. (eds.): Experimental and Clinical Neurotoxicology. Baltimore, Williams & Wilkins, p. 206, 1980. Sahenk, Z., and Lasek, R. J.: Inhibition of proteolysis blocks anterograde-to-retrograde conversion of axonally transported vesicles. Brain Res. 460:199, 1988. Sahenk, Z., and Mendell, J. R.: Ultrastructural study of zinc pyridinethione-induced peripheral neuropathy. J. Neuropathol. Exp. Neurol. 38:532, 1979.

196. Sahenk, Z., and Mendell, J. R.: Axoplasmic transport in zinc pyridinethione neuropathy: evidence for an abnormality in distal turn-around. Brain Res. 186:343, 1980. 197. Sahenk, Z., and Mendell, J. R.: Acrylamide and 2,5-hexanedione neuropathies: abnormal bidirectional transport rate in distal axons. Brain Res. 219:397, 1981. 198. Saunders, N. R., Dziegielewska, K., Haggendal, C. J., and Dahlstrom, A. B.: Slow accumulation of choline acetyltransferase in crushed sciatic nerves of the rat. J. Neurobiol. 4:95, 1973. 199. Sayre, L. M., Autilio-Gambetti, L., and Gambetti, P.-L.: Pathogenesis of experimental giant neurofilamentous axonopathies: a unified hypothesis based on chemical modification of neurofilaments. Brain Res. 357:69, 1985. 200. Schmidt, M. L., Martin, J. A., Lee, V. M., and Trojanowski, J. Q.: Convergence of Lewy bodies and neurofibrillary tangles in amygdala neurons of Alzheimer’s disease and Lewy body disorders. Acta Neuropathol (Berl.) 91:475, 1996. 201. Schmidt, R. E., Beaudet, L. N., Plurad, S. B., and Dorsey, D. A.: Axonal cytoskeletal pathology in aged and diabetic human sympathetic autonomic ganglia. Brain Res. 769:375, 1997. 202. Schmidt, R. E., Grabau, G. G., and Yip, H. K.: Retrograde axonal transport of [125I] nerve growth factor in ileal mesenteric nerves in vitro: effect of streptozotocin diabetes. Brain Res. 378:325, 1986. 203. Schmidt, R. E., Matschinsky, F. M., and Godfrey, D. A.: Fast and slow axoplasmic flow in sciatic nerve of diabetic rats. Diabetes 24:1081, 1975. 204. Schmidt, R. E., Plurad, S. G., and Saffitz, J. E.: Retrograde axonal transport of [l25I]-nerve growth factor in rat ileal mesenteric nerves: effect of streptozotocin diabetes. Diabetes 34:1230, 1985. 205. Schroer, T. A., Brady, S. T., and Kelly, R. B.: Fast axonal transport of foreign synaptic vesicles in squid axoplasm. J. Cell Biol. 101:568, 1985. 206. Schwab, M. E., Agid, Y., Glowinski, J., and Thoenen, H.: Retrograde axonal transport of 125I-tetanus toxin as a tool for tracing fiber connections in the central nervous system: connections of the rostral part of the rat neostriatum. Brain Res. 126:211, 1977. 207. Schwab, M. E., Heumann, R., and Thoenen, H.: Communication between target organs and nerve cells: retrograde axonal transport and site of action of nerve growth factor. Cold Spring Harb. Symp. Quant. Biol. 46:125, 1982. 208. Schwab, M. E., and Thoenen, H.: Selective binding uptake and retrograde transport of tetanus toxin by nerve terminals in the rat iris. J. Cell Biol. 77:3, 1978. 209. Scott, F. H.: On the relation of nerve cells to fatigue of their nerve fibers. J. Physiol. (Lond.) 34:145, 1906. 210. Seiler, W. A., Schelgel, R., Mackinnon, S., and Dellon, A. L.: Double crush syndrome: experimental model in the rat. Surg. Forum 34:596, 1983. 211. Shah, J. V., and Cleveland, D. W.: Slow axonal transport: fast motors in the slow lane. Curr. Opin. Cell Biol. 14:58, 2002. 212. Sickles, D. W.: Toxic neurofilamentous axonopathies and fast anterograde axonal transport. Part I. The effects of single doses of acrylamide on the rate and capacity of transport. Neurotoxicology 10:91, 1989.

Axonal Transport: Properties, Mechanisms, and Role in Nerve Disease 213. Sickles, D. W., Stone, J. D., and Friedman, M. A.: Fast axonal transport: a site of acrylamide neurotoxicity. Neurotoxicology 23:223, 2002. 214. Sidenius, P.: The effect of doxorubicin on slow and fast components of the axonal transport system in rats. Brain 109:885, 1986. 215. Sidenius, P., and Jakobsen, J.: Axonal transport in early experimental diabetes. Brain Res. 173:315, 1979. 216. Sidenius, P., and Jakobsen, J.: Retrograde axonal transport: a possible role in the development of neuropathy. Diabetologia 20:110, 1981. 217. Sidenius, P., and Jakobsen, J.: Anterograde axonal transport in rats during intoxication with acrylamide. J. Neurochem. 40:697, 1983. 218. Smith, R. S.: The short term accumulation of axonally transported organelles in the region of localized lesions of single myelinated axons. J. Neurocytol. 9:39, 1980. 219. Southam, E., Thomas, P. K., King, R. H., et al.: Experimental vitamin E deficiency in rats: morphological and functional evidence of abnormal axonal transport secondary to free radical damage. Brain 114:915, 1991. 220. Souyri, F., Chretien, M., and Droz, B.: “Acrylamideinduced” neuropathy and impairment of axonal transport of proteins. I. Multifocal retention of fast transported proteins at the periphery of axons as revealed by light microscope radioautography. Brain Res. 205:1, 1981. 221. Spencer, P. S., Sabri, M. I., Schaumburg, H. H., and Moore, C.: Does a defect in energy metabolism in the nerve fiber cause axonal degeneration in polyneuropathies? Ann. Neurol. 5:501, 1979. 222. Stone, J. D., Peterson, A. P., Eyer, J., et al.: Axonal neurofilaments are non-essential elements of toxicant-induced reductions in fast axonal transport. Part II. Video-enhanced differential interference microscopy in PNS axons. Toxicol. Appl. Pharmacol. 161:50, 1999. 223. Stromska, D., and Ochs, S.: Patterns of slow transport in the sensory nerves. J. Neurobiol. 12:441, 1981. 224. Tang, B. Y., Komiya, Y., and Austin, L.: Axoplasmic flow of phospholipids and cholesterol in the sciatic nerve of normal and dystrophic mice. Exp. Neurol. 43:13, 1974. 225. Tetzlaff, W., and Kreutzberg, G. W.: Ornithine decarboxylation in motoneurons during regeneration. Exp. Neurol. 89:679, 1985. 226. Tomlinson, D. R., and Mayer, J. H.: Reversal of deficits in axonal transport and nerve conduction velocity by treatment of streptozotocin-diabetic rats with myo-inositol. Exp. Neurol. 89:420, 1985. 227. Tomlinson, D. R., Sidenius, P., and Larsen, J. R.: Slow component-a of axonal transport, nerve myo-inositol, and aldose reductase inhibition in streptozocin-diabetic rats. Diabetes 35:398, 1985. 228. Tomlinson, D. R., Townsend, J., and Fretten, P.: Prevention of defective axonal transport in streptozocin-diabetic rats by treatment with “Statil” (ICI 128436), an aldose reductase inhibitor. Diabetes 34:970, 1985. 229. Tsukita, S., and Ishikawa, H.: The movement of membranous organelles in axons: electron microscopic identification of anterogradely and retrogradely transported organelles. J. Cell Biol. 84:513, 1980.

409

230. Tu, P.-H., Raju, P., Robinson, K. A., et al.: Transgenic mice carrying a human mutant superoxide dismutase transgene develop neuronal cytoskeletal pathology resembling human amyotrophic lateral sclerosis lesions. Proc. Natl. Acad. Sci. U. S. A. 93:3155, 1996. 231. Upton, A. R., and McComas, A. J.: The double crush in nerve entrapment syndromes. Lancet 2:359, 1973. 232. Wakabayashi, M., Araki, K., and Takahashi, Y.: Increased rate of fast axonal transport in methylmercury-induced neuropathy. Brain Res. 117:524, 1976. 233. Waller, A.: Experiments on the section of glossopharyngeal and hypoglossal nerves of the frog and observations of the alternatives produced thereby in the structure of their primitive fibers. Philos. Trans. R. Soc. Lond. 140:423, 1850. 234. Wang, L., and Brown, A.: Rapid intermittent movement of axonal neurofilaments observed by fluorescence photobleaching. Mol. Biol. Cell 12:3257, 2001. 235. Wang, L., and Brown, A.: Rapid movement of microtubules in axons. Curr. Biol. 12:1496, 2002. 236. Warita, H., Itoyama, Y., and Abe, K.: Selective impairment of fast anterograde axonal transport in the peripheral nerves of asymptomatic transgenic mice with a G93A mutant SOD1 gene. Brain Res. 819:120, 1999. 237. Weiss, H. D., Walker, M. D., and Wiernik, P. H.: Neurotoxicity of commonly used antineoplastic drugs. N Engl. J. Med. 291:127, 1974. 238. Whitely, S. J., Townsend, J., Tomlinson, D. R., and Brown, A. M.: Fast orthograde axonal transport in sciatic motoneurones and nerve temperature in streptozotocindiabetic rats. Diabetologia 28:8471, 1985. 239. Willard, M., Cowan, W. M., and Vagelos, P. R.: The polypeptide composition of intra-axonally transported proteins: evidence for four transport velocities. Proc. Natl. Acad. Sci. U. S. A., 71:2183, 1974. 240. Willard, M. B., and Hulebak, K. L.: The intra-axonal transport of polypeptide H: evidence for a fifth (very slow) group of transported proteins in the retinal ganglion cells of the rabbit. Brain Res. 36:289, 1977. 241. Williamson, T. L., and Cleveland, D. W.: Slowing of axonal transport is a very early event in the toxicity of ALS-linked SOD1 mutants to motor neurons. Nat. Neurosci. 2:50, 1999. 242. Wong, P. C., Pardo, C. A., Borchelt, D. R., et al.: An adverse property of a familial ALS-linked SOD1 mutation causes motor neuron disease characterized by vacuolar degeneration of mitochondria. Neuron 14:1104, 1995. 243. Xu, Z., Cork, L. C., Griffin, J. W., and Cleveland, D. W.: Increased expression of neurofilament subunit NF-L produces morphological alterations that resemble the pathology of human motor neuron disease. Cell 73:23, 1993. 244. Yitadello, M., Filliatreau, G., and Dupont, J. L.: Altered axonal transport of cytoskeletal proteins in the mutant diabetic mouse. J. Neurochem. 45:860, 1985. 245. Zhang, F., Eckman, C., Younkin, S., et al.: Increased susceptibility to ischemic brain damage in transgenic mice overexpressing the amyloid precursor protein. J. Neurosci. 17:7655, 1997. 246. Zhao, C., Takita, J., Tanaka, Y., et al.: Charcot-Marie-Tooth disease type 2A caused by mutation in a microtubule motor KIF1B␤. Cell 105:587, 2001.

19 Myelination UELI SUTER AND RUDOLF MARTINI

Morphology of Peripheral Nervous System Myelination Establishing the Myelin Internode Architecture of the Node of Ranvier

Molecular Biology of PNS Myelination Proteins Involved in Myelination Regulation of Myelin Genes Regulation of Myelin Protein Biosynthesis

MORPHOLOGY OF PERIPHERAL NERVOUS SYSTEM MYELINATION Establishing the Myelin Internode Myelination is a pivotal prerequisite for the rapid and saltatory conduction of action potentials, but also for the maintenance of the axonal structure.109 Two principle myelinating glial cells are found in the nervous system of higher vertebrates, the oligodendrocytes in the central nervous system (CNS) and the Schwann cells in the peripheral nervous system (PNS). Being predominantly derivatives of the neural crest,75,90,127 embryonic Schwann cells travel along axon bundles during development.16,17 These cells are most probably Schwann cell precursors, that is, prospective Schwann cells that still lack the ability to survive in the absence of axons.117 In the rat, Schwann cell precursors have been identified up to embryonic days 14 and 15 and do not yet express significant levels of the glial marker S100. Surprisingly, only scant attention has been paid so far to the morphologic characteristics and developmental changes of these cells, although they have been viewed as pivotal cellular components for axon survival.148 A role in guiding axons to their targets during development has been proposed earlier.120 This appears less probable because, during development, Schwann cell precursors follow pioneer axons in the embryo.16,17,154 In addition, Schwann cell precursors cannot survive when contact with axons is precluded,117 and axons find their targets in the absence of Schwann cell precursors.148

Role of Lipids in Myelination Role of the Extracellular Matrix, Its Cellular Receptors, and the Schwann Cell Cytoskeleton in Myelination

Schwann cell precursor cells are characteristically devoid of a basal lamina, intermingling and extending processes between the axon bundles (Fig. 19–1A). These cells are additionally found at the margin of the prospective nerve, facing the mesenchyme with their abaxonal surface (Fig. 19–1A).12,17,34,35,108,135,209 When development proceeds, Schwann cell precursors start to form a basal lamina, proliferate, and collectively ensheathe fasciculating axons, forming so-called Schwann cell families as initially described by Webster and colleagues.136,195 By then such glial cells most probably have reached the stage of immature Schwann cells, a term characterized by the ability to survive independently as a result of an autocrine survival mechanism.99,117 In the rat, all glial cells of the sciatic nerve are immature Schwann cells by embryonic day 17, whereas in the mouse, immature Schwann cells are the main population by embryonic day 15.43,117 The next step in development is related to an increase in caliber of some axons before becoming myelinated. They acquire a “peripheral” position within the axon bundles and achieve a so-called 1:1 ratio with immature Schwann cells (Fig. 19–1B). This stage of Schwann cell development is called “promyelin,” a morphologic starting point for myelin formation. In rodents, this sorting of many larger caliber axons into the promyelin stage occurs at around birth. However, myelin formation in the PNS is not a highly synchronized event. Axons with large calibers become myelinated earlier than those with smaller diameters, so that some promyelin stages can still be found in mice of approximately 3 weeks of age. 411

412

Neurobiology of the Peripheral Nervous System

B

A

FIGURE 19–1 Developing peripheral nerve of the mouse. A, Sciatic nerve of a mouse at embryonic day 13. Schwann cell precursors (SP) are found at the periphery of the nerve and protrude slender processes in between the embryonic axon bundles (A). A Schwann cell precursor in a more central position of the nerve is also visible (SP). Note the absence of Schwann cell basal lamina. M ⫽ mesenchymal cell. B, Femoral nerve of a 4-day-old mouse. Three axons of larger caliber display a thin sheath of compacted myelin. The upper fiber shows a redundant myelin profile (arrow). The axon demarcated by an asterisk is positioned at the periphery of a bundle of prospective nonmyelinating axons and is separated from them by a process of the immature Schwann cell. Double asterisks mark a similar situation at the bottom of the micrograph. A1 and A2 indicate promyelin stages. The Schwann cell process of A1 shows a more progressed stage of spiraling than A2. Bars: 1 ␮m.

The promyelin stage is followed by spiral formation of one of the Schwann cell processes engulfing the axon (Fig. 19–2; see also Fig. 19–1B). The apposition of Schwann cell membranes at the inner, axon-related side of the Schwann cell process is called the “inner mesaxon,” whereas the contact of Schwann cell membranes at the endoneurial side is termed the “outer mesaxon” (Fig. 19–2). These cell contacts are characterized by the formation of adherens (desmosomelike) junctions (Fig. 19–2). In an elegant study, Bunge and colleagues25 investigated the question of which end of the Schwann cell, the inner or outer one, turns around the axon during myelination. For this purpose, living rat Schwann cells co-cultured with dorsal root ganglion neurons were first investigated at the light microscopic level and the movements of the Schwann cell nuclei were recorded. After having monitored the behavior of the Schwann cells for up to 70 hours, the cultures were fixed and the axon–Schwann cell units in question were examined by electron microscopy (Fig. 19–3). These studies revealed that it is basically the inner, axon-related Schwann cell process that turns around the axon, and that the cell soma containing the nucleus is dragged behind at a much slower rate than the inner lip of the axon-related end of the Schwann cell process. This view is in line with older models stating that insertion of myelin components (e.g., radiolabeled lipids)136 occurs along the

entire extension of the developing spiral. Thus relatively rapid turning of the inner lip of a relatively slow-moving Schwann cell body and the simultaneous overall insertion of myelin membrane components appear to be characteristic events during myelin formation in the PNS. In vivo studies revealed that the myelinating process appears more complicated than just a spiral formation of the inner Schwann cell process around the axon. For instance, irregular forms with thin, redundant myelin loops are often observed (see Fig. 19–1B). In addition, the myelinating process in the region of the Schwann cell nucleus turns at a higher rate around the axon than the same process at the level of the Schwann cell edges (i.e., near the prospective node of Ranvier). A comprehensive overview of the morphologic events during myelin formation is summarized by Webster and colleagues.136,195 A unique subcellular feature of myelinating glial cells is the compaction of the turning membranes. In the case of rodent Schwann cells, this occurs when the myelinating process has turned around the axon a few times. Two processes occur simultaneously: (1) the narrowing of the spiraling surface membranes from approximately 12 to approximately 2 nm, and (2) the “squeezing out” of cytoplasm. The collapsed cytoplasmic sites of the Schwann cell membranes fuse and form a 3.5-nm-wide electron-dense

Myelination

413

A

FIGURE 19–2 Internodal segment of the mature myelin sheath. A, The majority of this internodal segment consists of compact myelin (M), whereas cytoplasmic aspects are sparse. Inner (arrow) and outer mesaxons (arrowhead) are indicated. A ⫽ axon. B, Highpower micrograph of A focusing on the inner mesaxon (arrow). Note adherens-like junction and layers of compact myelin consisting of major dense lines and intraperiod lines. Bars: 0.5 ␮m (A); 0.25 ␮m (B).

B

band called the “major dense line”; the membrane leaflets facing the extracellular space of the spiral form the “intraperiod line,” which is double-layered as a result of the 2-nm gap separating the extracellular leaflets (see Fig. 19–2B).136 Immature myelin sheaths are characterized by relatively expanded sites of still uncompacted myelin containing cytoplasm. During maturation, however, uncompacted myelin becomes restricted to distinct sites, such as the periaxonal collar (see Fig. 19–2), the outer Schwann cell loop, the paranodal loops, and the Schmidt-Lanterman incisures. The

latter are funnel-like clefts traversing the internodal myelin sheath and forming a helical cytoplasmic band that connects the periaxonal with the perinuclear (abaxonal) Schwann cell cytoplasm. In longitudinal sections of osmium tetroxide– fixed tissue, they can be identified on light microscopy as slim, bright lines obliquely traversing the myelin. In teased fiber preparations labeled for markers of noncompacted myelin (e.g., myelin-associated glycoprotein [MAG], connexin 32 [Cx32]), Schmidt-Lanterman incisures appear as funnel-like profiles.5,6,159 Electron microscopy reveals that

414

Neurobiology of the Peripheral Nervous System

FIGURE 19–3 Light microscopy (A) and electron micrographs (B and C) of a profile of a myelinated nerve fiber grown in vitro. A, Three internodes (1 through 3) are visible. These internodes have been investigated for 34 hours and nuclear movements have been recorded. B and C, The insets show the position of the nuclei of internodes 1 (B) and 3 (C) at distinct time points. The electron micrographs show the direction of the inner lip of the myelinating Schwann cells (arrows). Note that movement of Schwann cell nuclei (insets) and of the corresponding inner lip (electron micrograph) are always in the same direction. Bars: 10 ␮m (A), 0.5 ␮m (B and C). (From Bunge, R. P., Bunge, M. B., and Bates, M.: Movements of the Schwann cell nucleus implicate progression of the inner (axon-related) Schwann cell process during myelination. J. Cell Biol. 109:273, 1989, with copyright permission of The Rockefeller University Press.)

they consist of slender pockets of cytoplasm that form a helical funnel. Based on the observation that these cytoplasmic domains are Cx32 positive, it has been speculated that they form a rapid, radial cytoplasmic pathway for ions and small molecules between the adaxonal and abaxonal Schwann cell cytoplasm.160 Interestingly, recent immunohistochemical studies on teased fiber preparations revealed that Caspr1/paranodin and the voltage-gated K⫹ channels Kv1.1 and 1.2 demarcate the axonal domain underlying the inner loop of the Schmidt-Lanterman incisures.5,6,159 An interesting question is the regulation of myelin thickness. In particular, a striking feature is the positive correlation between axonal caliber and myelin thickness. This is

clearly reflected by the observation that the quotient between axon diameter and fiber diameter (taken at the outer aspect of the myelin sheath) is generally constant when axons of different diameters are compared.64 A similarly constant interrelationship exists between internodal length and axon diameter in the adult nerve. Quantitative studies in various species have revealed that a normal internode of a mature peripheral nerve is usually 100 times as long as the diameter of the corresponding axon.64 However, developmental studies in nerves with a robust growth in length during maturation in the absence of Schwann cell proliferation modified this simple view, as revealed by studies in phrenic nerves in rabbits and several

Myelination

peripheral nerves in humans.64,165 In young rabbits, myelin sheaths are relatively thin for axon diameter, followed by an increase in myelin thickness.64 An extreme case is found in various human peripheral nerves. Although radial growth of axons is completed at 5 years of age, the corresponding myelin continues to grow in thickness until the age of 17.165 In tibial nerves of rats, maximal growth in axonal diameter occurs between 3 weeks and 3 months. Here, the developmental change in relative myelin thickness is not uniform because myelin sheaths of fibers of different caliber behave in a slightly different way during development.60

mental in rapid restoration of the resting potential after an action potential has occurred. This chapter focuses on the structural characteristics of the node of Ranvier. The molecular components of the node of Ranvier and their functional roles are discussed below and in Chapter 24. The outer cytoplasmic aspect of the Schwann cells ends in the nodal region with microvilli-like protrusions that abut against the nodal plasmalemma (Fig. 19–4A). These protrusions contain actin filaments189 and typical microvilli-related cytoskeletal components (i.e., ezrin, radixin, and moesin, the ERM proteins).114,161 The ERM proteins may bind to the actin filaments linking the cytoskeleton to integral membrane components, possibly adhesion molecules. The functional role of the microvilli is not yet known, but it is assumed that they organize or stabilize distinct domains of the axolemma via their ERM-linked adhesion molecules in a trans-specific manner.11,114,161 The nodal microvilli and the nodal axolemma are surrounded by an extracellular matrix consisting of tenascin-C and the proteoglycan NG2.5,6,107,134 The nodal axolemma is characterized by a typical, electron-dense undercoating (see Fig. 19–4A). In this region, cytoskeletal components such as ankyrin G and spectrin are highly enriched and may organize the accumulation of the adhesion molecules Nr-CAM and neurofascin 186 and voltage-gated Na⫹ channels.11,159 It is thus plausible to assume that the accumulation of cytoskeletal elements is at

Architecture of the Node of Ranvier The node of Ranvier is highly organized both structurally and molecularly. Ontogenetically, it develops from the cell borders of neighboring Schwann cells that form the insulating myelin sheath around axons of larger caliber.14 From the physiologic point of view, the node of Ranvier is the component of the fiber responsible for the generation and propagation of action potentials. Two important prerequisites for this functional role are of note: (1) the accessibility of the axolemma for currents and (2) the presence of a high density of voltage-gated sodium channels that allow the generation of the action potential. In addition, voltagegated potassium channels and Na⫹,K⫹-ATPases are instru-

A

415

B

FIGURE 19–4 Nodal complex of the peripheral nerve of the adult mouse. A, Cross section of a node of Ranvier. Note nodal microvilli (arrows) and electron-dense undercoating of the axolemma (arrowheads). B, Longitudinal section of a paranodal region of a smaller caliber nerve fiber. Note “regular” organization of paranodal loops and that each loop abuts the axolemma. Septate-like junctions are indicated by arrows. Figure continued on following page

416

Neurobiology of the Peripheral Nervous System

C

D FIGURE 19–4 Continued C and D, Longitudinal section of a nodal region of a smaller (C) and a larger (D) caliber nerve fiber. In D, the organization of paranodal loops appears less regular than in C. Also, the juxtaparanodal aspects are conspicuous, as is typical for larger caliber axons, with interdigitations between inner Schwann cell loops and protruding axolemma (arrows). Bars: 0.5 ␮m (A), 0.25 ␮m (B), 1 ␮m (C and D).

least partially related to the electron-dense undercoating of the nodal axolemma. Other cytoskeletal components are neurotubules and neurofilaments. Characteristically, the latter components are of low phosphorylation stage as opposed to their internodal counterparts.5 The low phosphorylation level may be related to the diminution of axon diameter at the nodal region.39

Other striking compartments of the nodal region are the paranodes. In longitudinal sections, myelin lamellae end as cytoplasmic pockets that abut the paranodal axolemma (Fig. 19–4B and C). Typically, those myelin lamellae that are next to the axon end earlier than those closer to the node (Fig. 19–4B and C). Similar to the SchmidtLanterman incisures, they form a helical cytoplasmic band

417

Myelination

that connects the periaxonal with the perinuclear (abaxonal) Schwann cell cytoplasm. Forming reflexive Cx32positive gap junctions, both paranodal pockets and Schmidt-Lanterman incisures may function as rapid pathways between periaxonal and abaxonal cytoplasm.8 The paranodal pockets appear to be closely associated with the paranodal axolemma. With conventional aldehyde fixation, the axolemma appears “scalloped” as a result of a slight protrusion of the convex axon-associated surfaces of the paranodal pockets into the axon surface (see Fig. 19–4B). In addition, septate junctions form from the axonal surface and abut the paranodal Schwann cell membrane (Fig. 19–4B). These septate-like junctions are the place where Neurexin-Caspr-Paranodin is expressed and are involved in the molecular organization of the nodal environment.48,115,133 A detailed description of the functional role of paranodal molecules is found in Chapter 24. Other subcellular components of the paranodal loops are microtubules, which appear crosscut in longitudinal nerve sections as a result of their spiral orientation with respect to the longitudinal axis of the axon. Other compartments of interest are adherens (desmosome-like) junctions, which connect the paranodal pockets at their lateral sites. These junctions are characterized by the presence of the Ca2⫹-dependent adhesion molecule E-cadherin, which might be involved in their formation (see below).50 Depending on the size of the axon, there are in principle two different forms of paranodal region14,136,137 (see Fig. 19–4C and D). In smaller fibers, the cytoplasmic pockets approach the axon tangentially, and each pocket reaches the axolemma (Fig. 19–4B and C). The general appearance is “orderly.” In larger caliber fibers, the pockets approximate the nodal axolemma in a steep fashion and not all of them reach the axolemma (Fig. 19–4D). They are usually smaller and often appear much more electron dense than the pockets of the smaller axons. As a result of these characteristic appearances and the fact that myelin structures of larger fibers are more prone to show fixation artifacts, the paranodes of thick fibers appear less organized (Fig. 19–4D). Another issue of interest is the node-related compartment proximal to the paranode, the juxtaparanodal region. Some authors add this region to the paranodal compartment,14 but based on recent molecular findings, we prefer to treat this region as a separate compartment. It is particularly interesting from the physiologic point of view because this is the expression site of the axon- and Schwann cell–related, voltage-dependent K⫹ channels (see Chapter 24). From the morphologic point of view, the juxtaparanodal zone is particularly conspicuous in larger caliber fibers. As a result of a tapering of larger caliber axons, the myelin sheath appears fluted. In cross sections, the myelin sheath and the axon are cross-shaped or acquire the form of a trefoil. Here, the outer Schwann cell cytoplasm “fills” the grooves in the folded myelin and shows accumulations of mitochondria. The axon–Schwann cell

apposition often shows a complicated organization related to a profound interdigitation of inner Schwann cell loop and axolemma, called the axon–Schwann cell network (formerly paranodal network) (see Fig. 19–4D). In longitudinal sections, a striking asymmetry is detectable in that the axon–Schwann cell network is much more conspicuous at the distal side of the nodal region than at the proximal side. In addition, axonal lysosomes containing acid phosphatase are more abundant in the distal regions.68

MOLECULAR BIOLOGY OF PNS MYELINATION Proteins Involved in Myelination (Fig. 19–5 and Table 19–1) P0 P0 (myelin protein zero [MPZ]) is the major protein of the peripheral nerve myelin sheath, accounting for 50% to 60% of myelin proteins.93,94 It is a 30-kDa adhesion molecule belonging to the immunoglobulin (Ig) superfamily. Its single Ig domain bears a glycosylation site that can carry the HNK-1 carbohydrate epitope that is common to many cell adhesion and recognition molecules of the nervous system.156 In vitro studies suggested a pivotal function in myelin compaction as a result of its homophilic adhesive properties when transgenically expressed in various cell types.45,58,164 This view received support from the determination of the three-dimensional structure of the extracellular domain of P0 by x-ray crystallography.167 A simplistic model, however, does not, at present, explain all aspects of P0-mediated adhesion, because the intracellular domain of the molecule appears to be crucial for the adhesive properties of the extracellular domain,66,198 possibly by providing phosphorylation sites for protein kinase C.200 In addition, the intracellular domain of P0 contains predominantly basic residues, which have been suggested to interact with negatively charged phospholipids of the adjacent cytoplasmic parts of the Schwann cell membrane, leading to the formation of the major dense line.41,85,95 Expression of antisense-P0 messenger RNA (mRNA) blocks myelination in vitro at the promyelin stage,123 and studies in P0-deficient mice clearly demonstrate that P0 has a pivotal function during myelination, such as spiral formation, compaction, and myelin maintenance (see Chapter 24). Although P0 has been identified as a major component of peripheral myelin, it is expressed much earlier than initially anticipated. Its first expression has been found on neural crest cells of chicken and rat,15,32,75,92,127 but the functional roles at these early developmental stages are still elusive. Upregulation of P0 expression is temporally linked to the myelination process, and axon–Schwann cell units having acquired the promyelin stage show some weak but substantial immunoreactivity.110 The majority of P0 immunoreactivity, however, is found in compact

FIGURE 19–5 Schematic representation of various myelin-related proteins of the peripheral nerve. Note that distinct domains of myelin (e.g., compact vs. noncompact myelin) are associated with their specific set of molecules. For the specific functional roles of the molecules, see text.

419

Myelination

Table 19–1. Proteins of the Peripheral Nerve Myelin Sheath Molecule

Structure

Localization in Myelin

Function

Comments

P0, MPZ

Adhesion molecule with one Ig-like domain

Compact myelin (minor expression in neural crest cells)

Extra- and intracellular compaction, promotion of myelination

PMP22

Membrane protein of the EMP family, tetraspan topology questionable Cytoplasmic, membrane-associated protein

Compact myelin (minor expression in neural crest cells) Compact myelin (minor expression in neural crest cells) Noncompacted myelin

Cell cycle proliferation, apoptosis, spreading; regulation of myelination Supports P0 in intracellular compaction

Major PNS myelin protein, mutations lead to inherited myelin disorders Mutations lead to inherited myelin disorders

MBP

MAG

Adhesion molecule with 5 Ig-like domains

Cx32

Tunnel protein of gap junctions

Noncompacted myelin

E-cadherin

Ca2⫹-dependent adhesion molecule

Neurofascin 155

Adhesion molecule with 6 Ig-like domains

Adherens junctions of noncompacted myelin and of inner and outer mesaxon Adaxonal site of paranodal loops

TAG-1

Adhesion molecule with 6 Ig-like domains (GPI-linked) Tetraspan protelipid

MAL Periaxin

Cytoskeleton-associated protein with PDZ motif; L-periaxin, S-periaxin

Juxtaparanodal myelin

Compact myelin, nonmyelinating fibers Periaxonal cytoplasm during development (mainly S-form), abaxonal in mature fibers (mainly L-form)

myelin.110,188,205 Interestingly, P0 expression on nonmyelinating Schwann cells appears to be actively suppressed by the corresponding small-caliber axons.92 Peripheral Myelin Protein 22 Peripheral myelin protein 22 (kDa) (PMP22), the first identified culprit gene for inherited neuropathies of CharcotMarie-Tooth (CMT) type, is an integral membrane glycoprotein containing a single N-linked carbohydrate moiety. It is expressed mainly in compact myelin of peripheral nerves.177 It accounts for approximately 2% to 5% of myelin proteins. In contrast to P0, expression in other neural structures and even non-neural organs and tissues has been described.7,121,130,178 Initially, PMP22 was believed to be a tetraspan membrane molecule, but recent studies suggest that an alternative organization is conceivable.183 By biochemical approaches, PMP22 and P0 could be co-purified from peripheral nerve myelin, suggesting that they form complexes in myelin.46 In vitro experiments suggest a func-

Maintenance of myelin and axonal integrity Communication of non-compacted myelin domains Stabilization of adherens junctions

Glial partner for NCP-1

Lack of MBP leads to minor abnormalities in PNS In vitro experiments suggested involvement in myelin formation Mutations lead to inherited myelin disorders Downregulation in P0 mutants is associated with abnormal junctions Dysregulated in mice with abnormal NCP-1 expression

May mediate distribution of K⫹ channels Component of lipid rafts L-form interacts indirectly with laminin of SC basal lamina via DRP2–␣␤–dystroglycan complex

Mutations lead to inherited myelin disorders

tion of PMP22 in cell growth, differentiation, and apoptosis,22,210 but mice lacking PMP22 or overexpressing the gene show a normal number of Schwann cells perinatally.153 Although it is predominantly a component of compact myelin, PMP22 expression in rat embryos starts long before the onset of myelination, in neural crest–derived progenitor cells at embryonic day 14. By contrast, such an early expression has not been found in the mouse, so the role of the early expression remains elusive.127 PMP22 appears to play pivotal roles during myelination, because in the absence of the gene, myelination is severely affected.2 At present the exact role of PMP22 during myelination is difficult to derive from the pathologic changes in nerves of knockout mutants (see Chapter 24). Myelin Basic Protein Myelin basic proteins (MBPs) are products from a large gene that encodes two families of proteins, the Golli proteins and the “classic” (i.e., myelin-related) MBP isoforms.27 The

420

Neurobiology of the Peripheral Nervous System

latter appear to be confined to differentiated myelinforming cells, whereas Golli proteins are more widely distributed. Substantial levels are expressed in neurons and oligodendrocyte precursors at embryonic stages as well as in the mature immune system.27,76,89 Five isoforms of the classic proteins ranging from 14 to 21.5 kDa are expressed in myelin-forming cells, but the 17.5-kDa isoform alone can obviously fulfill the functional tasks of all isoforms in the nervous system.84 In the PNS, MBP accounts for 5% to 15% of myelin proteins. As a component of the cytoplasm-related aspect of the compact myelin sheath (i.e., the major dense line), the highly positively charged domains of the protein interact with the negatively charged phospholipids and may mediate compaction.173 Morphologic evidence for this hypothesis is the absence of major dense lines in CNS myelin of shiverer mice,144,150 a spontaneous functional null mutant for classic MBPs.146,149 In the PNS of shiverer mice, absence of MBP does not lead to loss of major dense lines.85,151 This is due to the compensatory function of the cytoplasmic part of the P0 protein.41,111 However, subtle changes have been noted in the PNS of shiverer mice, such as an increase of Schmidt-Lanterman incisures and posttranscriptional modulation of Cx32 protein expression.72,171 Myelin-associated Glycoprotein MAG is a minor component of myelin, accounting for only 0.1% of other PNS myelin proteins. It is a transmembrane glycoprotein with five Ig-like extracellular domains, a transmembrane domain, and an intracellular part. Two isoforms differing in their intracellular domains have been found: a 67-kDa (S-MAG) and a 72-kDa (L-MAG) isoform.88,192 The major form of the PNS is S-MAG. The glycoprotein is a typical adhesion molecule with plasma membrane binding partners.140,203 Recently, the neurotrophin receptor p75 and the Nogo66 receptor have been identified as axon-related partners that transmit the growth-inhibitory activities of MAG to CNS neurons.42,202 Additionally, there is compelling evidence that sialic acid–containing glycoproteins and complex gangliosides are good candidates for axonal MAG receptors.38,203 Thus MAG can also been viewed as a member of the siglecs.83 As opposed to P0, PMP22, and MBP, MAG is confined to noncompacted myelin. Because of its early expression during myelination and its strategic location at the axon–Schwann cell interface,112,190,191 and from perturbation experiments using MAG antisense mRNA,124 the molecule has been considered to play pivotal roles during myelination. Mice deficient in MAG showed an unexpectedly normal myelination in the PNS and MAG-deficient Schwann cells myelinated normally in vitro (see Chapter 24).29,96,118 However, there was prominent axon and myelin degeneration in motor nerves.28,65,204 Thus

MAG is dispensable for myelin formation but is vital for the maintenance of the axon–Schwann cell entity. Connexins 32 and 29 Similar to MAG, Cx32 is another component of noncompacted regions of peripheral nerve myelin,160 but its structure is completely different from that of MAG. Cx32 is a tunnel protein with four transmembrane domains. It belongs to a still-enlarging multigene family comprising more than 15 other gap junction components.169 Cx32 is regulated similarly to a typical myelin gene in that its expression in Schwann cells is dependent on contact with larger caliber axons.160 In the mature myelin sheath, Cx32 is considered to connect each individual noncompacted loop by diffusion channels, thus creating a radial pathway between the periaxonal cytoplasmic collar and the outer Schwann cell cytoplasm.8,160 The blockade of this pathway by the engineered null mutation in the mouse or by various spontaneous mutations in humans1,147 leads to typical pathologic hallmarks such as widened periaxonal collars and abnormal noncompacted zones.4,163 A recently identified myelin-related connexin is connexin 29 (Cx29).3,97,172 Its distribution overlaps with, but is also distinct from, that of Cx32. In addition to the paranodal region and the Schmidt-Lanterman incisures, Cx29 is primarily found at the innermost aspect of the myelin sheath and in the juxtaparanodal regions (i.e., in the immediate neighborhood of K⫹ channels).3 Although its functional role remains to be established, it is plausible to assume that it contributes to the reflexive junctions between noncompacted myelin loops. This would explain why, in Cx32deficient mice, dye-coupling between abaxonal and adaxonal Schwann cell cytoplasm is not interrupted.3,8 E-cadherin E-cadherin is a member of the very large group of Ca2⫹dependent adhesion molecules comprising more than 80 members in mammals.201 It is expressed in many different tissues of epithelial organization.181,182 In the myelin sheath, E-cadherin is associated with autotypic adherens junctions of the inner and outer mesaxons, the SchmidtLanterman incisures, and the paranodal loops.50 Indirect evidence for the role of E-cadherin in myelinating Schwann cells comes from studies in P0-deficient mice in which E-cadherin is downregulated and adherens junctions are only sparsely developed.116 However, tissuespecific gene ablation in Schwann cells revealed that E-cadherin is not required for the integrity of myelinated PNS fibers but does affect the correct formation of autotypic adherens junctions of the outer mesaxon.206 Neurofascin 155 and TAG-1 The neurofascins are members of the Ig superfamily closely related to L1-like adhesion molecules.155,156 In myelinated fibers, the larger isoform, designated neuro-

421

Myelination

fascin 186, is a molecular component of the nodal axolemma, whereas the smaller isoform, neurofascin 155, has been detected as an adaxonal, paranodal membrane component of oligodendrocytes and Schwann cells.21,139,180 Although direct proof is lacking, it is probable that neurofascin 155 is the trans-interacting partner of the paranodal, axon-related NCP1-contactin complex. Indirect evidence for this is the observation that, in mice with an abnormal NCP1-contactin complex expression, neurofascin 155 is altered in its distribution21,44,139 (see also Chapter 24). TAG-1, a glycosylphosphatidylinositol-linked member of the Ig superfamily, has recently been localized at the juxtaparanodal region of developing and mature myelinated fibers. Its role at that position is still obscure, but the co-localization with voltage-gated K⫹ channels suggests participation in the distribution of these channels.187 Myelin and Lymphocyte Protein and Plasmolipin The myelin and lymphocyte protein MAL is a tetraspan proteolipid component of both PNS and CNS myelin.61 In the cell membrane, it is associated with glycosphingolipids possibly forming microdomains in the form of lipid rafts. Apart from its expression in myelin, MAL is expressed in various epithelial tissues, such as kidney and stomach, but also by immune cells.61 In the peripheral nervous system, it is found in compact myelin, but also on nonmyelinating and immature Schwann cells. The early expression argues in favor of a functional role during segregation of axons leading to the promyelin stage. Interestingly, overexpression of MAL leads to an unusual separation of nonmyelinated axons, which might be an argument in favor of a segregating function of MAL in the PNS.62 A MAL-related and myelin-associated proteolipid protein is plasmolipin, which shares approximately 30% amino acid identity with MAL.61,78,102 Its functional roles in the myelinating process are presently not known. Periaxin Periaxin is a Schwann cell protein with two different sizes: 147 kDa (L-periaxin) and 16 kDa (S-periaxin). As a product of a single gene, each form contains a PDZ (postsynaptic density protein 95, Drosophila discs large tumor suppressor, zonula occludens-1) motif, suggesting involvement in cell-cell contact and cell signaling.47,168 The name of the protein is derived from its localization within the periaxonal Schwann cell loop in young mice,70 whereas the expression of the predominantly large form of the protein shifts to the abaxonal compartment in mature nerves.162 The function of periaxin has been investigated by the generation of null mutants. A striking feature was the formation of myelin tomacula at 6 weeks of age followed by myelin degeneration.71 Thus the protein is a nonredundant component of the myelinating Schwann cell stabilizing the myelinated axon–Schwann cell unit. This stabilization is

most probably mediated by the interaction of L-periaxin with dystroglycan-dystrophin–related protein 2 (DRP2), which in turn forms a complex with ␣,␤-dystroglycan that interacts with laminin in the Schwann cell basal lamina.168 Thus periaxin and DRP2 form an essential link between the extracellular matrix of the Schwann cell and the Schwann cell itself, which eventually stabilizes the mature myelin sheath. Mutations in the periaxin gene lead to distinct forms of inherited peripheral neuropathies in humans.18,74

Regulation of Myelin Genes Myelination during development and after demyelination demands an extremely high synthesis rate of myelin proteins and lipids within a short period of time. To accomplish this precisely regulated task, the myelination process is guided by the coordinated expression of genes that encode myelin components.185 The principal control of the system is governed by transcription factors. Some of these are present ubiquitously in many cell types, including Schwann cells, whereas others are more specifically expressed. Together, these factors are responsible for the cell type–specific and differentiation stage–specific gene expression, including the regulation of myelination.197 In addition to the pivotal roles of transcription factors in orchestrating the myelination process, additional regulation at the posttranscriptional level by altering mRNA stability is likely, but this possibility has not yet been thoroughly explored experimentally in the PNS. Recent evidence suggests a particularly important role for the close interaction between neurons (axons) and the ensheathing Schwann cells in the regulatory network governing myelination.104,157 This finding may be part of the reason why the use of transcription factors by Schwann cells is quite different from that by oligodendrocytes, the myelinating counterpart in the CNS.197 Several parallel strategies have been employed to examine myelin gene regulation in the PNS. Major efforts have focused on the identification of transcription factors expressed by Schwann cells that might be master regulators of myelin gene expression, as has been found during muscle development (MyoD).20 This hunt has proved to be difficult, but some candidates have been identified. In particular, suppressed cyclic AMP (cAMP)–inducible POU protein (SCIP, Tst-1, Oct-6) plays a critical role in the regulation of PNS myelination because SCIP-deficient mice are characterized by severe congenital hypomyelination of peripheral nerves. However, the defect is transient and the nerves are close to normal at 3 months of age.81 Similarly, the zinc-finger family protein Egr2 (Krox20) appears to be required for the normal development of the myelinating Schwann cell phenotype. Transgenic mice carrying a null mutation in the egr2 gene display severe defects in Schwann cell development resulting in hyperproliferation

422

Neurobiology of the Peripheral Nervous System

and presumed differentiation arrest at the premyelination stage.186 Egr2 also regulates the expression of genes involved in PNS myelination.119 In line with these findings, different mutations affecting Egr2 are associated with CMT disease type 1, the Déjèrine-Sottas syndrome, and congenital hypomyelinating neuropathy (see Chapter 72).119,194 Additional transcription factors that may affect late steps in the differentiation of Schwann cells include Pax3, Krox24, and Brn5.157,199 Of particular importance is the transcription factor Sox10, which has been identified as a common transcriptional modulator for SCIP, Krox20, and Pax3. It was suggested that Sox10 is responsible for celltype specificity by interacting with these transcription factors in developing and mature glia.87 Furthermore, a crucial role of Sox10 in early Schwann cell development has been demonstrated23,126 and reviewed by Lobsiger and colleagues.100 With regard to PNS myelination, Sox10 regulates MPZ (P0) and GJB1 (Cx32) gene expression.19,132 Mutations in each of these genes are responsible for distinct forms of CMT disease (see Chapter 71). Sox10 mutations are associated with Shah-Wardenburg syndrome (WS4), a neurocristopathy with intestinal aganglionosis, pigmentation defects, sensorineural deafness, and, in specific cases, alterations in myelination in the PNS and CNS.80,125,138 These Sox10 mutants are unable to activate the Cx32 promoter and, conversely, a specific mutation in the Cx32 promoter, previously described in a patient with CMT disease, impairs Sox10 function directly.19 This elegant series of studies provides a direct genetic link between the regulation of myelin gene expression by the transcription factors Egr2 and Sox10 and myelin deficiencies in inherited peripheral neuropathies. It suggests that factors involved in the regulation of myelin gene expression should be generally considered as candidates involved in the etiology of peripheral neuropathies. Furthermore, a general concept emerges that correct myelination, myelin maintenance, and axonal maintenance are crucially dependent on the correct expression levels of myelin proteins.13,104 Besides studying the role of transcription factors and their modulators, myelination-related, cell type–specific control can also be elucidated by the identification and characterization of cis-acting control elements of genes encoding myelin components.196 Transfection experiments in cultured Schwann cells have been used as a standard assay in this regard, followed by classic footprinting and bandshift assays to identify the sequences that are bound by transcription factors.24 However, there are limitations with this assay system because Schwann cells that have been kept in culture do not express myelin genes up to the rates that are observed in vivo during development and in regeneration. Co-culturing with neurons is required for myelination, but because of technical restrictions, this system is not well suited for gene transfer analysis. Transgenic mice, however, provide an excellent assay system to examine myelin gene regulation because numerous developmental

and physiologic signals for correct interactions are present. Such a strategy has recently led to the identification of Schwann cell–specific enhancers that mediate axonal regulation in SCIP,106 MBP,59 Egr2,69 and PMP22103 genes. Initial results of studies with similar aims have been reported for the 2⬘,3⬘-cyclic nucleotide 3⬘-phosphodiesterase gene,30,73 the P0 gene,52 and the proteolipid protein gene.105 Apart from providing important insights into the molecular basis of the regulation of myelin genes, these studies have also generated important tools to target transgenic expression of genes of choice to Schwann cells.53,143 PMP22, the dosage-sensitive culprit gene of CMT disease type 1A and hereditary neuropathy with liability to pressure palsies, was reviewed by Suter and Snipes.176 Studies on PMP22 gene regulation during myelination may also provide the basis for future therapies.77,207 What are the signal transduction events that regulate Schwann cell myelination, and by which transcriptional effectors? Examination of the role of different signaling pathways in Schwann cell differentiation using Schwann cell–neuron co-cultures revealed that, at early stages, inhibition of phosphatidylinositol 3 (PI3)-kinase, but not myelin-associated protein (MAP) kinase, blocked Schwann cell elongation and subsequent myelination.113 After Schwann cells established a one-to-one relationship with axon segments, inhibition of PI3-kinase did not block myelin formation, but the myelin segments were shorter and the rate of myelin protein accumulation was decreased. PI3-kinase inhibition had no detectable effect on the maintenance of myelin sheaths in mature myelinated co-cultures. Interestingly, glial growth factor (GGF), a neuregulin-1 isoform, significantly inhibited myelination in the same system by preventing axonal segregation and ensheathment.208 Treatment of established myelinated cultures with GGF resulted in striking demyelination. The neuregulin receptors ErbB2 and ErbB3 are expressed on ensheathing and myelinating Schwann cells and are rapidly activated by GGF treatment. GGF treatment of myelinating cultures also induced phosphorylation of PI3-kinase, MAP kinase, and a 120-kDa protein. Thus neuronal mitogens, including neuregulins, may inhibit myelination during development, and activation of mitogen signaling pathways may contribute to the initial demyelination and subsequent Schwann cell proliferation observed in various pathologic conditions. Surprisingly, functional information about the downstream transcriptional effectors mediating the events described above is still scarce.196 Several AP-1 (dimeric transcription factors of the Jun, Fos, and ATF family of basic leucine zipper proteins) binding sites have been found in putative regulatory regions of myelin genes. However, although c-Jun is expressed by Schwann cells, a direct functional role of AP-1 transcription factors in regulating PNS myelination remains unclear.174 Similarly, the transcription factor CREB (which becomes activated

423

Myelination

through phosphorylation) is found in Schwann cells, and a protein kinase A–dependent increase of CREB phosphorylation was observed after axonal stimulation and the elevation of cAMP, intracellular calcium, platelet-derived growth factor, or endothelins. Furthermore, ␤-neuregulin treatment causes sustained CREB phosphorylation, and this effect appears to be mainly dependent on the MAP kinase pathway.179 Although this is strong evidence for a function of CREB in the regulation of Schwann cells, a direct link to myelination has not yet been described. Such a link has been established for progesterone that promotes the myelination of sciatic nerves during regeneration after a cryolesion.86 Progesterone triggers a strong upregulation of Egr1 (Krox24), Egr2, Egr3, and FosB in cultured Schwann cells, most likely at the transcriptional level via the interaction of the hormone with its cognate receptor. Furthermore, neuroactive steroids are able to upregulate the mRNA levels of PMP22 and P0 and, thus, specific receptor ligands or antagonists may be promising candidates for therapeutic approaches in demyelinating inherited neuropathies.101,166

Regulation of Myelin Protein Biosynthesis Relatively little is known about the regulation of myelin proteins at the posttranslational level, although a thorough understanding of intracellular transport mechanisms, protein turnover regulation, and membrane and myelin insertion processes are of utmost importance to the regulation of myelination in health and disease. The PMP22 protein provides an illustrative example. Schwann cells express low levels of PMP22 in the absence of neurons. Only if myelin is formed do these levels increase significantly. The exact amount of PMP22 protein produced is critical, because peripheral neuropathies result from either its underexpression or its overexpression. Most of the newly synthesized PMP22 in Schwann cells is rapidly degraded in the endoplasmic reticulum.128,129 Only a small proportion of the total PMP22 acquires complex glycosylation and accumulates in the Golgi compartment. This material is translocated to the Schwann cell membrane in detectable amounts only when axonal contact and myelination occur. However, myelination does not alter the rapid turnover of PMP22 in Schwann cells. Inhibition of the proteosome pathway results in marked accumulation of PMP22 in the perinuclear cytoplasm, forming unique intracellular inclusions called aggresomes.122,152 Moreover, overexpression of PMP22 in Schwann cells can induce the perinuclear accumulation of the protein. Thus degradation of PMP22 by the proteosome pathway might be critical for the regulation of PMP22 protein levels and may be involved in the pathogenesis of PMP22-associated peripheral neuropathies. Recently, association of the chaperone calnexin with PMP22 has also been reported.40 To understand the dynamics of the biosynthesis of PMP22 and other myelin

protein further, sophisticated novel techniques need to be developed that allow direct observation of how proteins travel in living myelinating Schwann cells.26,131

Role of Lipids in Myelination Myelinogenesis requires precise coordination not only of gene expression of the proteins involved in myelination but also of those enzymes associated with the synthesis of myelin lipids. This is not surprising given that myelin membranes contain 70% lipids and 30% proteins, both of which are highly organized within the myelin sheath.67 Myelin membranes resemble apical membranes in their lipid composition and contain high levels of cholesterol and glycosphingolipids such as galactocerebroside and sulfatide.175 In addition, myelin-forming cells are also polarized cells with distinct sorting pathways. Thus it has been speculated that the model of sphingolipid-cholesterol rafts to generate distinct transport and sorting pathways, particular to apical membranes, may also apply to myelinating cells.170 In the CNS, a number of studies have provided support for such a model, and membrane proteins appear to be delivered to the myelin sheath of an oligodendrocyte on rafts with a distinctive lipid composition.91 Not much of this important process has been explored with regard to PNS myelination. Recent data suggest that the myelin proteolipid MAL, which has been shown in other systems to be required for apical transport, and the glycosylphosphatidylinositol-anchored protein CD59 are specific components of PNS myelin rafts.49 The structural myelin proteins PMP22 and P0 are, at least partially, associated with glycosphingolipid rafts,49,79 but the functional role of the association remains to be determined. It appears that the role of rafts in myelination is likely to be more complex than in epithelial cells. Myelinating cells establish several structurally and functionally distinct membrane compartments within the myelin sheath, and each of these myelin compartments is characterized by a unique protein and lipid composition.159 This diversity may preclude the simple assignment of an apical compartment in myelinating cells. Sorting of proteins by different types of rafts may mediate different transport pathways to direct specific proteins to their respective target membranes within the myelin sheath. This process may also guide the assembly of microdomains in myelin membranes that might be important for signal transduction. The role of galactolipids in PNS myelination has recently been examined using mice deficient in the UDPgalactose:ceramide:galactosyltransferase gene.44 These mice lack galactolipids. In contrast to the CNS, PNS myelin sheaths appear normal with regard to myelin stability, thickness, and compaction. The paranodal morphology is also not disrupted in a major way, but transverse bands are absent. This is accompanied by compromised function of the axon–Schwann cell junctions because

424

Neurobiology of the Peripheral Nervous System

potassium channels that are normally shielded by myelin membrane are now exposed to the extracellular milieu. These observations and other recently published data suggest that galactolipids may play an important role in axon–Schwann cell interactions and node of Ranvier formation.141 A potential molecular explanation for these findings might be that glycosphingolipids are important components of intracellular lipid rafts. Thus alteration of these lipids may alter protein trafficking and result in a disruption of the highly ordered cellular distribution of membrane proteins and lipids that are crucial for myelinated nerves. In general, CNS nerves appear to be much more dependent on galactolipids than PNS nerves. The basis for this observation is currently not clear but may be related to the fact that myelinated nerve fibers are stabilized and protected by basal laminae. Recent evidence using genetically altered mice suggests that sialylated glycosphingolipids (gangliosides) are critical components of myelinated peripheral nerves. Mice lacking complex gangliosides show axonal degeneration and demyelination in the PNS.193 This phenotype is similar to MAG-deficient mice and, indeed, MAG is downregulated in complex ganglioside-deficient mouse nerves. The proposal that this is because complex gangliosides are MAG ligands in the nerve remains to be substantiated because an alternative explanation, such as altered MAG trafficking to the cell membrane resulting from altered lipid composition, remains a possibility. The first step of sphingolipid synthesis involves the enzyme serine palmitoyltransferase.98 In mammals, this enzyme is composed of two subunits, encoded by SPTLC1 and SPTLC2, both of which are necessary for activity. Mutations in SPTLC1 are associated with hereditary sensory neuropathy type 1.10,37 Furthermore, a ganglioside-induced differentiation-associated gene (GDAP1) is mutated in the demyelinating peripheral neuropathy CMT type 4A.9,36 In both cases, the corresponding disease mechanisms remain to be determined. Finally, aberrant sphingolipid turnover can also affect myelinated peripheral nerves considerably.184 These diseases are described in Chapter 69.

Role of the Extracellular Matrix, Its Cellular Receptors, and the Schwann Cell Cytoskeleton in Myelination Proper interactions between Schwann cells and axons leading to the formation of myelin are crucially dependent on the extracellular matrix. This has first become evident in classic in vitro myelination studies by Bunge’s laboratory demonstrating that, if the generation of a proper basal lamina is prevented, myelination and the accompanying expression of myelin proteins do not take place.54 The concept emerged that Schwann cells are polarized cells that separate

themselves from mesenchyme by forming a basal lamina. The adaxonal (apical) surface of the Schwann cell contacts the axon while the abaxonal (basal) side is in contact with the basal lamina, likely to provide mechanical support during the myelination process and to permit signal transduction to the cytoskeleton and back. This regulation is critically important to allow the major Schwann cell cytoskeletal reorganization that is required during myelination.55 The importance of the basal lamina in myelination is further highlighted by the mouse mutant dystrophia muscularis and merosin-deficient patients with congenital muscular dystrophy. Both are affected by muscular dystrophy and a dysmyelinating peripheral neuropathy resulting from mutations in the Lama2 gene resulting in defective laminin-2.82 The connection between myelinating Schwann cells and the basal lamina is molecularly maintained by interactions with laminin receptors. There are mainly two sorts of laminin receptors. First, two dystroglycan complexes link the basal lamina to the actin cytoskeleton of Schwann cells through dystrophins or utrophin. A third dystroglycan complex is found exclusively in myelinating Schwann cells and contains DRP2 instead of dystrophin and utrophin.168 Disruption of this complex by mutation in the intracellularly associated periaxin protein, the culprit gene of CMT type 4F, causes demyelination of both mouse and human peripheral nerves.13,71 Intriguingly, Schwann cell dystroglycan complexes also act as receptors for Mycobacterium leprae, which invades nonmyelinating Schwann cells by binding of the M. leprae glycolipid PGL-1 to dystroglycan receptors in the Schwann cell plasma membrane.145 Simultaneously, the bacillus also attacks myelinating Schwann cells and induces demyelination. Dedifferentiation and proliferation of formerly myelinating Schwann cells provides additional carrier cells for colonization by M. leprae, while the loss of myelin sheaths appears to cause axonal degeneration as has been observed in several demyelinating neuropathies.104,109 A second set of laminin receptors found on Schwann cells is integrins.33,142 On myelinating Schwann cells, these include mainly ␣6/␤1 and ␣6/␤4 integrin dimers. Functionblocking anti–␤1 integrin antibodies block Schwann cell attachment to laminin as well as myelination in vitro.56 Furthermore, gene ablation of ␤1 in Schwann cells (␤1 was eliminated at E17.5 in the mouse) leads to markedly delayed myelination. This appears to be due to a partial failure of Schwann cells to subdivide axons as well as impaired adhesion of the promyelinating Schwann cells to their basal laminae.53 However, myelination per se is possible in these mutants, albeit with a long delay. One possible interpretation is that Schwann cells switch their laminin integrin receptors during development. Immature and promyelinating Schwann cells express ␣6/␤1 integrin while myelinating Schwann cells predominantly express ␣6/␤4 integrin.51 Analysis of ␤4 integrin–deficient mice has shown that initial steps of myelination are not dependent on the expression of the ␤4 subunit and in vitro myelin-

Myelination

ation can proceed without ␤4 integrin expression.63 Alternatively, dystroglycan receptors might be predominantly important for myelination.158 Only limited information is currently available with regard to the signal transduction mechanisms related to integrin signaling in Schwann cell myelination. Initial experiments have demonstrated that ␤1 integrin, focal adhesion kinase, paxillin, and fyn kinase form an actin-associated complex in Schwann cells adhering to basal lamina in the presence of axons.31 The adaptor protein paxillin also binds directly to schwannomin, and this interaction mediates the membrane localization of schwannomin to the plasma membrane, where it associates with ␤1 integrin and ErbB2.57 Given the complex structural changes that Schwann cells undergo during myelination, the further elucidation of potential integrin-mediated inside-out and outside-in signaling is likely to contribute significantly to our understanding of the myelination process.

REFERENCES 1. Abrams, C. K., Oh, S., Ri, Y., and Bargiello, T. A.: Mutations in connexin 32: the molecular and biophysical bases for the X-linked form of Charcot-Marie-Tooth disease. Brain Res. Rev. 32:203, 2000. 2. Adlkofer, K., Martini, R., Aguzzi, A., et al.: Hypermyelination and demyelinating peripheral neuropathy in Pmp22-deficient mice. Nat. Genet. 11:274, 1995. 3. Altevogt, B. M., Kleopa, K. A., Postma, F. R., et al.: Connexin29 is uniquely distributed within myelinating glial cells of the central and peripheral nervous systems. J. Neurosci. 22:6458, 2002. 4. Anzini, P., Neuberg, D. H. H., Schachner, M., et al.: Structural abnormalities and deficient maintenance of peripheral nerve myelin in mice lacking the gap junction protein connexin 32. J. Neurosci. 17:4545, 1997. 5. Arroyo, E. J., and Scherer, S. S.: On the molecular architecture of myelinated fibers. Histochem. Cell Biol. 113:1, 2000. 6. Arroyo, E. J., Xu, Y.-T., Zhou, L., et al.: Myelinating Schwann cells determine the internodal localization of Kv1.1, Kv1.2, Kv␤2, and Caspr. J. Neurocytol. 28:333, 1999. 7. Baechner, D., Liehr, T., Hameister, H., et al.: Widespread expression of the peripheral myelin protein-22 gene (pmp22) in neural and non-neural tissues during murine development. J. Neurosci. Res. 42:733, 1995. 8. Balice-Gordon, R. J., Bone, L. J., and Scherer, S. S.: Functional gap junctions in the Schwann cell myelin sheath. J. Cell Biol. 142:1095, 1998. 9. Baxter, R. V., Ben Othmane, K., Rochelle, J. M., et al.: Ganglioside-induced differentiation-associated protein-1 is mutant in Charcot-Marie-Tooth disease type 4A/8q21. Nat. Genet. 30:21, 2002. 10. Bejaoui, K., Wu, C., Scheffler, M. D., et al.: SPTLC1 is mutated in hereditary sensory neuropathy, type 1. Nat. Genet. 27:261, 2001.

425

11. Bennett, V., Lambert, S., Davis, J. Q., and Zhang, X.: Molecular architecture of the specialized axonal membrane at the node of Ranvier. Soc. Gen. Physiol. Ser. 52:107, 1997. 12. Bentley, C. A., and Lee, K. F.: p75 is important for axon growth and Schwann cell migration during development. J. Neurosci. 20:7706, 2000. 13. Berger, P., Young, P., and Suter, U.: Molecular cell biology of Charcot-Marie-Tooth disease. Neurogenetics 4:1, 2002. 14. Berthold, C. H.: Development of nodes of Ranvier in feline nerves: an ultrastructural presentation. Microsc. Res. Tech. 34:399, 1996. 15. Bhattacharyya, A., Frank, E., Ratner, N., and Brackenbury, R.: P0 is an early marker of the Schwann cell lineage in chickens. Neuron 7:831, 1991. 16. Billings-Gagliardi, S.: Mode of locomotion of Schwann cells migrating in vivo. Am. J. Anat. 150:73, 1977. 17. Billings-Gagliardi, S., Webster, H. d., and O’Connel, M. F.: In vivo and electron microscopic observations on Schwann cells in developing tadpole nerve fibers. Am. J. Anat. 141:375, 1974. 18. Boerkoel, C., Takashima, H., Stankiewicz, P., et al.: Periaxin mutations cause recessive Dejerine-Sottas neuropathy. Am. J. Hum. Genet. 68:325, 2001. 19. Bondurand, N., Girard, M., Pingualt, V., et al.: Human connexin 32, a gap junction protein altered in CharcotMarieTooth disease, is directly regulated by the transcription factor SOX10. Hum. Mol. Genet. 10:2783, 2001. 20. Borycki, A. G., and Emerson, C. P.: Muscle determination: another key player in myogenesis? Curr. Biol. 7:R620, 1997. 21. Boyle, M. E., Berglund, E. O., Murai, K. K., et al.: Contactin orchestrates assembly of the septate-like junctions at the paranode in myelinated peripheral nerve. Neuron 30:385, 2001. 22. Brancolini, C., Edomi, P., Marzinotto, S., and Schneider, C.: Exposure at the cell surface is required for gas3/PMP22 to regulate both cell death and cell spreading: implication for the Charcot-Marie-Tooth type 1A and Dejerine-Sottas diseases. Mol. Biol. Cell 11:2901, 2000. 23. Britsch, S., Goerich, D. E., Riethmacher, D., et al.: The transcription factor Sox10 is a key regulator of peripheral glial development. Genes Dev. 15:66, 2001. 24. Brown, A. M., and Lemke, G.: Multiple regulatory elements control transcription of the peripheral myelin protein zero gene. J. Biol. Chem. 272:28939, 1997. 25. Bunge, R. P., Bunge, M. B., and Bates, M.: Movements of the Schwann cell nucleus implicate progression of the inner (axon-related) Schwann cell process during myelination. J. Cell Biol. 109:273, 1989. 26. Caduff, J., Sansano, S., Bonnet, A., et al.: Characterization of GFP-MAL expression and incorporation in rafts. Microsc. Res. Tech. 52:645, 2001. 27. Campagnoni, A. T., and Skoff, R. P.: The pathobiology of myelin mutants reveal novel biological functions of the MBP and PLP genes. Brain Pathol. 11:74, 2001. 28. Carenini, S., Montag, D., Cremer, H., et al.: Absence of myelin-associated glycoprotein (MAG) and the neural cell adhesion molecule (N-CAM) interferes with the maintenance, but not with the formation of peripheral myelin. Cell Tissue Res. 287:3,1997.

426

Neurobiology of the Peripheral Nervous System

29. Carenini, S., Montag, D., Schachner, M., and Martini, R.: MAG-deficient Schwann cells myelinate dorsal root ganglion cells in culture. Glia 22:213, 1998. 30. Chandross, K. J., Cohen, R. I., Paras, P. Jr., et al.: Identification and characterization of early glial progenitors using a transgenic selection strategy. J. Neurosci. 19:759, 1999. 31. Chen, L. M., Bailey, D., and Fernandez-Valle, C.: Association of beta 1 integrin with focal adhesion kinase and paxillin in differentiating Schwann cells. J. Neurosci. 20:3776, 2000. 32. Cheng, L., and Mudge, A. W.: Cultured Schwann cells constitutively express the myelin protein P0. Neuron 16:309, 1996. 33. Chernousov, M. A., and Carey, D. J.: Schwann cell extracellular matrix molecules and their receptors. Histol. Histopathol. 15:593, 2000. 34. Ciutat, D., Caldero, J., Oppenheim, R. W., and Esquerda, J. E.: Schwann cell apoptosis during normal development and after axonal degeneration induced by neurotoxins in the chick embryo. J. Neurosci. 16:3979, 1996. 35. Cravioto, H.: The role of Schwann cells in the development of human peripheral nerves: an electron microscopic study. J. Ultrastruct. Res. 12:634, 1965. 36. Cuesta, A., Pedrola, L., Sevilla, T., et al.: The gene encoding ganglioside-induced differentiation-associated protein 1 is mutated in axonal Charcot-Marie-Tooth type 4A disease. Nat. Genet. 30:22, 2002. 37. Dawkins, J. L., Hulme, D. J., Brahmbhatt, S. B., et al.: Mutations in SPTLC1, encoding serine palmitoyltransferase, long chain base subunit-1, cause hereditary sensory neuropathy type I. Nat. Genet. 27:309, 2001. 38. De Bellard, M. E., and Filbin, M. T.: Myelin-associated glycoprotein, MAG, selectively binds several neuronal proteins. J. Neurosci. Res. 56:213, 1999. 39. de Waegh, S. M., Lee, V. M. Y., and Brady, S. T.: Local modulation of neurofilament phosphorylation, axonal caliber, and slow axonal transport by myelinating Schwann cells. Cell 68:451, 1992. 40. Dickson, K. M., Bergeron, J. J., Shames, I., et al.: Association of calnexin with mutant peripheral myelin protein-22 ex vivo: a basis for “gain-of-function” ER diseases. Proc. Natl. Acad. Sci. U. S. A. 99:9852, 2002. 41. Ding, Y., and Brunden, K. R.: The cytoplasmic domain of myelin glycoprotein P0 interacts with negatively charged phospholipid bilayers. J. Biol. Chem. 269:10764, 1994. 42. Domeniconi, M., Cao, Z., Spencer, T., et al.: Myelin-associated glycoprotein interacts with the Nogo66 receptor to inhibit neurite outgrowth. Neuron 35:283, 2002. 43. Dong, Z., Sinanan, A., Parkinson, D., et al.: Schwann cell development in embryonic mouse nerves. J. Neurosci. Res. 56:334, 1999. 44. Dupree, J. L., and Popko, B.: Genetic dissection of myelin galactolipid function. J. Neurocytol. 28:271, 1999. 45. D’Urso, D., Brophy, P. J., Staugaitis, S. M., et al.: Protein zero of peripheral nerve myelin: biosynthesis, membrane insertion, and evidence for homotypic interaction. Neuron 2:449, 1990. 46. D’Urso, D., Ehrhardt, P., and Müller, H. W.: Peripheral myelin protein 22 and protein zero: a novel association in peripheral nervous system myelin. J. Neurosci. 19:3396, 1999.

47. Dytrych, L., Sherman, D. L., Gillespie, C. S., and Brophy, P. J.: Two PDZ domain proteins encoded by the murine periaxin gene are the result of alternative intron retention and are differentially targeted in Schwann cells. J. Biol. Chem. 273:5794, 1998. 48. Einheber, S., Zanazzi, G., Ching, W., et al.: The axonal membrane protein Caspr, a homologue of neurexin IV, is a component of the septate-like paranodal junctions that assemble during myelination. J. Cell Biol. 139:1495, 1997. 49. Erne, B., Sansano, S., Frank, M., and Schaeren-Wiemers, N.: Rafts in adult peripheral nerve myelin contain major structural myelin proteins and myelin and lymphocyte protein (MAL) and CD59 as specific markers. J. Neurochem. 82:550, 2002. 50. Fannon, A. M., Sherman, D. L., Ilynia, G., et al.: Novel E-Cadherin-mediated adhesion in peripheral nerve: Schwann cell architecture is stabilized by autotypic adherens junctions. J. Cell Biol. 129:189, 1995. 51. Feltri, L., Scherer, S. S., Nemni, R., et al.: ␤4 integrin expression in myelinating Schwann cells is polarized, developmentally regulated and axonally dependent. Development 120:1287, 1994. 52. Feltri, M. L., D’Antonio, M., Quattrini, A., et al.: A novel P0 glycoprotein transgene activates expression of lacZ in myelin-forming Schwann cells. Eur. J. Neurol. 11:1577, 1999. 53. Feltri, M. L., Graus Porta, D., Previtali, S. C., et al.: Conditional disruption of beta 1 integrin in Schwann cells impedes interactions with axons. J. Cell Biol. 156:199, 2002. 54. Fernandez-Valle, C., Fregien, N., Wood, P. M., and Bunge, M. B.: Expression of the protein zero myelin gene in axonrelated Schwann cells is linked to basal lamina formation. Development 119:867, 1993. 55. Fernandez-Valle, C., Gorman, D., Gomez, A. M., and Bunge, M. B.: Actin plays a role in both changes in cell shape and gene expression associated with Schwann cell myelination. J. Neurosci. 17:241, 1997. 56. Fernandez-Valle, C., Gwynn, L., Wood, P. W., et al.: Anti␤1 integrin antibody inhibits Schwann cell myelination. J. Neurobiol. 25:1207, 1994. 57. Fernandez-Valle, C., Tang, Y., Ricard, J., et al.: Paxillin binds schwannomin and regulates its density-dependent localization and effect on cell morphology. Nat. Genet. 31:354, 2002. 58. Filbin, M. T., Walsh, F. S., Trapp, B. D., et al.: Role of myelin P0 protein as a homophilic adhesion molecule. Nature 344:871, 1990. 59. Forghani, R., Garofalo, L., Foran, D. R., et al.: A distal upstream enhancer from the myelin basic protein gene regulates expression in myelin-forming Schwann cells. J. Neurosci. 21:3780, 2001. 60. Fraher, J. P., O’Leary, D., Moran, M. A., et al.: Relative growth and maturation of axon size and myelin thickness in the tibial nerve of the rat. 1. Normal animals. Acta Neuropathol. 79:364, 1990. 61. Frank, M.: MAL, a proteolipid in glycosphingolipid enriched domains: functional implications in myelin and beyond. Prog. Neurobiol. 60:531, 2000. 62. Frank, M., Atanasoski, S., Sancho, S., et al.: Progressive segregation of unmyelinated axons in peripheral nerves,

Myelination

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

myelin alterations in the CNS, and cyst formation in the kidneys of myelin and lymphocyte protein-overexpressing mice. J. Neurochem. 75:1927, 2000. Frei, R., Dowling, J., Carenini, S., et al.: Myelin formation by Schwann cells in the absence of ␤4 integrin. Glia 27:269, 1999. Friede, R. L., and Beuche, W.: A new approach toward analyzing peripheral nerve fiber populations. I. Variance in sheath thickness corresponds to different geometric proportions of the internodes. J. Neuropathol. Exp. Neurol. 44:60, 1985. Fruttiger, M., Montag, D., Schachner, M., and Martini, R.: Crucial role for the myelin-associated glycoprotein in the maintenance of axon-myelin integrity. Eur. J. Neurosci. 7:511, 1995. Gao, Y., Li, W., and Filbin, M. T.: Acylation of myelin P0 protein is required for adhesion. J. Neurosci. Res. 60:704, 2000. Garbay, B., Heape, A. M., Sargueil, F., and Cassagne, C.: Myelin synthesis in the peripheral nervous system. Prog. Neurobiol. 61:267, 2000. Gatzinsky, K. P.: Node-paranode regions as local degradative centres in alpha-motor axons. Microsc. Res. Tech. 34:492, 1996. Ghislain, J., Desmarquet-Trin-Dinh, C., Jaegle, M., et al.: Characterisation of cis-acting sequences reveals a biphasic, axon-dependent regulation of Krox20 during Schwann cell development. Development 129:155, 2002. Gillespie, C. S., Sherman, D. L., Blair, G. E., and Brophy, P. J.: Periaxin, a novel protein of myelinating Schwann cells with a possible role in axon ensheathment. Neuron 12:497, 1994. Gillespie, C. S., Sherman, D. L., Fleetwood-Walker, S. M., et al.: Peripheral demyelination and neuropathic pain behavior in periaxin-deficient mice. Neuron 26:523, 2000. Gould, R. M., Byrd, A. L., and Barbarese, E.: The number of Schmidt-Lanterman incisures is more than doubled in shiverer PNS myelin sheaths. J. Neurocytol. 24:85, 1995. Gravel, M., Di Polo, A., Valera, P. B., and Braun, P. E.: Four-kilobase sequence of the mouse CNP gene directs spatial and temporal expression of lacZ in transgenic mice. J. Neurosci. Res. 53:393, 1998. Guilbot, A., Williams, A., Ravise, N., et al.: A mutation in periaxin is responsible for CMT4F, an autosomal recessive form of Charcot-Marie-Tooth disease. Hum. Mol. Genet. 10:415, 2001. Hagedorn, L., Suter, U., and Sommer, L.: P0 and PMP22 mark a multipotent neural crest-derived cell type that displays community effects in response to TGF-␤family factors. Development 126:3781, 1999. Hahn, A. F., Whitaker, J. N., Kachar, B., and Webster, H. d.: P2, P1, and P0 myelin protein expression in developing rat sixth nerve: a quantitative immunocytochemical study. J. Comp. Neurol. 260:501, 1987. Hai, M., Bidichandani, S. I., Hogan, M. E., and Patel, P. I.: Competitive binding of triplex-forming oligonucleotides in the two alternate promoters of the PMP22 gene. Antisense Nucleic Acid Drug Dev. 11:233, 2001. Hamacher, M., Pippirs, U., Köhler, A., et al.: Plasmolipin: genomic structure, chromosomal localization, protein

79.

80.

81. 82.

83.

84.

85.

86.

87.

88.

89.

90. 91. 92.

93. 94. 95.

96.

97.

427 expression pattern, and putative association with BardetBiedl syndrome. Mamm. Genome 12:933, 2001. Hasse, B., Bosse, F., and Muller, H. W.: Proteins of peripheral myelin are associated with glycosphingolipid/cholesterol-enriched membranes. J. Neurosci. Res. 69:227, 2002. Inoue, K., Tanabe, Y., and Lupski, J. R.: Myelin deficiencies in both the central and the peripheral nervous systems associated with a SOX10 mutation. Ann. Neurol. 46:313, 1999. Jaegle, M., and Meijer, D.: Role of Oct-6 in Schwann cell differentiation. Microsc. Res. Tech. 41:372, 1998. Jones, K. J., Morgan, G., Johnston, H., et al.: The expanding phenotype of laminin alpha2 chain (merosin) abnormalities: case series and review. J. Med. Genet. 38:649, 2001. Kelm, S., Schauer, R., and Crocker, P. R.: The sialoadhesins—a family of sialic acid-dependent cellular recognition molecules within the immunoglobulin superfamily. Glycoconj. J. 13:913, 1996. Kimura, M., Sato, M., Akatsuka, A., et al.: Overexpression of a minor component of myelin basic protein isoform (17.2 kDa) can restore myelinogenesis in transgenic shiverer mice. Brain Res. 785:245, 1998. Kirschner, D. A., and Ganser, A. L.: Compact myelin exists in the absence of basic protein in the shiverer mutant mouse. Nature 283:207, 1980. Koenig, H. L., Schumacher, M., Ferzaz, B., et al.: Progesterone synthesis and myelin formation by Schwann cells. Science 268:1500, 1995. Kuhlbrodt, K., Herbarth, B., Sock, E., et al.: Sox10, a novel transcriptional modulator in glial cells. J. Neurosci. 18:237, 1998. Lai, C., Brow, M. A., Nave, K. A., et al.: Two forms of 1B236/myelin-associated glycoprotein, a cell adhesion molecule for postnatal neural development, are produced by alternative splicing. Proc. Natl. Acad. Sci. U. S. A. 84:4337, 1987. Landry, C. F., Ellison, J. P., Skinner, E., and Campagnoni, A. T.: Golli-MBP proteins mark the earliest stages of fiber extension and terminal arboration in the mouse peripheral nervous system. J. Neurosci. Res. 50:265, 1997. Le Douarin, N., Dulac, C., Dupin, E., and Cameron-Curry, P.: Glia cell lineages in the neural crest. Glia 4:175, 1991. Lee, A. G.: Myelin: delivery by raft. Curr. Biol. 11:R60, 2001. Lee, M. J., Brennan, A., Blanchard, A., et al.: P0 is constitutively expressed in the rat neural crest and embryonic nerves and is negatively and positively regulated by axons to generate non-myelin forming and myelin forming Schwann cells, respectively. Mol. Cell. Neurosci. 8:336, 1997. Lemke, G.: Unwrapping the genes of myelin. Neuron 1:535, 1988. Lemke, G.: The molecular genetics of myelination: an update. Glia 7:263, 1993. Lemke, G., Lamar, E., and Patterson, J.: Isolation and analysis of the gene encoding peripheral myelin protein zero. Neuron 1:73, 1988. Li, C., Tropak, M. B., Gerial, R., et al.: Myelination in the absence of myelin-associated glycoprotein. Nature 369:747, 1994. Li, X., Lynn, B. D., Olson, C., et al.: Connexin29 expression, immunocytochemistry and freeze-fracture replica immunogold labelling (FRIL) in sciatic nerve. Eur. J. Neurosci. 16:795, 2002.

428

Neurobiology of the Peripheral Nervous System

98. Linn, S. C., Kim, H. S., Keane, E. M., et al.: Regulation of de novo sphingolipid biosynthesis and the toxic consequences of its disruption. Biochem. Soc. Trans. 29:831, 2001. 99. Lobsiger, C. S., Schweitzer, B., Taylor, V., and Suter, U.: Platelet-derived growth factor-BB supports the survival of cultured rat Schwann cell precursors in synergy with neurotrophin-3. Glia 30:290, 2000. 100. Lobsiger, C. S., Taylor, V., and Suter, U.: The early life of a Schwann cell. Biol. Chem. 383:245, 2002. 101. Magnaghi, V., Cavarretta, I., Galbiati, M., et al.: Neuroactive steroids and peripheral myelin proteins. Brain Res. Brain Res. Rev. 37:360, 2001. 102. Magyar, J., Ebensperger, C., Schaeren-Wiemers, N., and Suter, U.: Myelin and lymphocyte protein (MAL/ MVP17/VIP17) and plasmolipin are members of an extended gene family. Gene 189:269, 1997. 103. Maier, M., Berger, P., Nave, K. A., and Suter, U.: Identification of the regulatory region of the peripheral myelin protein 22 (PMP22) gene that directs temporal and spatial expression in development and regeneration of peripheral nerves. Mol. Cell. Neurosci. 20:93, 2002. 104. Maier, M., Berger, P., and Suter, U.: Understanding Schwann cell-neurone interactions: the key to CharcotMarie-Tooth disease? J. Anat. 200:357, 2002. 105. Mallon, B. S., Shick, H. E., Kidd, G. J., and Macklin, W. B.: Proteolipid promoter activity distinguishes two populations of NG2-positive cells throughout neonatal cortical development. J. Neurosci. 22:876, 2002. 106. Mandemakers, W., Zwart, R., Jaegle, M., et al.: A distal Schwann cell-specific enhancer mediates axonal regulation of the Oct-6 transcription factor during peripheral nerve development and regeneration. EMBO J. 19:2992, 2000. 107. Martin, S., Levine, A. K., Chen, Z. J., et al.: Deposition of the NG2 proteoglycan at nodes of Ranvier in the peripheral nervous system. J. Neurosci. 21:8119, 2001. 108. Martini, R.: Expression and functional roles of neural cell surface molecules and extracellular matrix components during development and regeneration of peripheral nerves. J. Neurocytol. 23:1, 1994. 109. Martini, R.: The effect of myelinating Schwann cells on axons. Muscle Nerve 24:456, 2001. 110. Martini, R., Bollensen, E., and Schachner, M.: Immunocytological localization of the major peripheral nervous system glycoprotein P0 and the L2/HNK-1 and L3 carbohydrate structures in developing and adult mouse sciatic nerve. Dev. Biol. 129:330, 1988. 111. Martini, R., Mohajeri, M. H., Kasper, S., et al.: Mice doubly deficient in the genes for P0 and myelin basic protein show that both proteins contribute to the formation of the major dense line in peripheral nerve myelin. J. Neurosci. 15:4488, 1995. 112. Martini, R., and Schachner, M.: Immunoelectron microscopic localization of neural cell adhesion molecules (L1, N-CAM, and MAG) and their shared carbohydrate epitope and myelin basic protein in developing sciatic nerve. J. Cell Biol. 103:2439, 1986. 113. Maurel, P., and Salzer, J. L.: Axonal regulation of Schwann cell proliferation and survival and the initial events of myelin-

114.

115.

116.

117.

118.

119.

120.

121.

122.

123.

124.

125.

126.

127.

128.

129.

130.

ation requires PI 3-kinase activity. J. Neurosci. 20:4635, 2000. Melendez-Vasquez, C. V., Rios, J. C., Zanazzi, G., et al.: Nodes of Ranvier form in association with ezrin-radixin-moesin (ERM)-positive Schwann cell processes. Proc. Natl. Acad. Sci. U. S. A. 98:1235, 2001. Menegoz, M., Gaspar, P., Le Bert, M., et al.: Paranodin, a glycoprotein of neuronal paranodal membranes. Neuron 19:319, 1997. Menichella, D. M., Arroyo, E. J., Awatramani, R., et al.: Protein zero is necessary for E-cadherin-mediated adherens junction formation in Schwann cells. Mol. Cell. Neurosci. 18:606, 2001. Mirsky, R., Jessen, K. R., Brennan, A., et al.: Schwann cells as regulators of nerve development. J. Physiol. Paris 96:17, 2002. Montag, D., Giese, K. P., Bartsch, U., et al.: Mice deficient for the myelin-associated glycoprotein show subtle abnormalities in myelin. Neuron 13:229, 1994. Nagarajan, R., Svaren, J., Le, N., et al.: EGR2 mutations in inherited neuropathies dominant-negatively inhibit myelin gene expression. Neuron 30:355, 2001. Noakes, P. G., and Bennett, M. R.: Growth of axons into developing muscles of the chick forelimb is preceded by cells that stain with Schwann cell antibodies. J. Comp. Neurol. 259:330, 1987. Notterpek, L., Roux, K. J., Amici, S. A., et al.: Peripheral myelin protein 22 is a constituent of intercellular junctions in epithelia. Proc. Natl. Acad. Sci. U. S. A. 98:14404, 2001. Notterpek, L., Ryan, M. C., Tobler, A. R., and Shooter, E. M.: PMP22 accumulation in aggresomes: implications for CMT1A pathology. Neurobiol. Dis. 6:450, 1999. Owens, G. C., and Boyd, C. J.: Expression antisense P0 RNA in Schwann cells perturbs myelination. Development 112:639, 1991. Owens, G. C., and Bunge, R. P.: Schwann cells infected with a recombinant retrovirus expressing myelin-associated glycoprotein antisense RNA do not form myelin. Neuron 7:565, 1991. Paratore, C., Eichenberger, C., Suter, U., and Sommer, L.: Sox10 haploinsufficiency affects maintenance of progenitor cells in a mouse model of Hirschsprung disease. Hum. Mol. Genet. 11:3075, 2002. Paratore, C., Goerich, D. E., Suter, U., et al.: Survival and glial fate acquisition of neural crest cells are regulated by an interplay between the transcription factor Sox10 and extrinsic combinatorial signaling. Development 128:3949, 2001. Paratore, C., Hagedorn, L., Floris, J., et al.: Cell-intrinsic and cell-extrinsic cues regulating lineage decisions in multipotent neural crest-derived progenitor cells. Int. J. Dev. Biol. 46:193, 2002. Pareek, S., Notterpek, L., Snipes, G. J., et al.: Neurons promote translocation of peripheral myelin protein 22 into myelin. J. Neurosci. 17:7754, 1997. Pareek, S., Suter, U., Snipes, G. J., et al.: Detection and processing of peripheral myelin protein PMP22 in cultured Schwann cells. J. Biol. Chem. 268:10372, 1993. Parmantier, E., Cabon, F., Braun, C., et al.: Peripheral myelin protein-22 is expressed in rat and mouse brain and spinal cord motoneurons. Eur. J. Neurosci. 7:1080, 1995.

Myelination 131. Pedraza, L., and Colman, D. R.: Fluorescent myelin proteins provide new tools to study the myelination process. J. Neurosci. Res. 60:697, 2000. 132. Peirano, R. I., Goerich, D. E., Riethmacher, D., and Wegner, M.: Protein zero gene expression is regulated by the glial transcription factor Sox10. Mol. Cell. Biol. 20:3198, 2000. 133. Peles, E., Nativ, M., Lustig, M., et al.: Identification of a novel contactin-associated transmembrane receptor with multiple domains implicated in protein-protein interactions. EMBO J. 16:978, 1997. 134. Peles, E., and Salzer, J. L.: Molecular domains of myelinated axons. Curr. Opin. Neurobiol. 10:558, 2000. 135. Peters, A., and Muir, A. R.: The relationship between axons and Schwann cells during development of peripheral nerves in the rat. Q. J. Exp. Physiol. 44:117, 1959. 136. Peters, A., Palay, S. L., and Webster, H. d.: The Fine Structure of the Nervous System. New York, Oxford University Press, 1991. 137. Phillips, D. D., Hibbs, R. G., Ellison, J. P., and Shapiro, H.: An electron microscopic study of central and peripheral nodes of Ranvier. J. Anat. 111:229, 1972. 138. Pingault, V., Guiochon-Mantel, A., Bondurand, N., et al.: Peripheral neuropathy with hypomyelination, chronic intestinal pseudo-obstruction and deafness: a developmental “neural crest syndrome” related to a SOX10 mutation. Ann. Neurol. 48:671, 2000. 139. Poliak, S., Gollan, L., Salomon, D., et al.: Localization of Caspr2 in myelinated nerves depends on axon-glia interactions and the generation of barriers along the axon. J. Neurosci. 21:7568, 2001. 140. Poltorak, M., Sadoul, R., Keilhauer, G., et al.: Myelinassociated glycoprotein, a member of the L2/HNK-1 family of neural cell adhesion molecules, is involved in neuronoligodendrocyte and oligodendrocyte-oligodendrocyte interaction. J. Cell Biol. 105:1893, 1987. 141. Popko, B.: Myelin galactolipids: mediators of axon-glial interactions? Glia 29:149, 2000. 142. Previtali, S. C., Feltri, M. L., Archelos, J. J., et al.: Role of integrins in the peripheral nervous system. Prog. Neurobiol. 64:35, 2001. 143. Previtali, S. C., Quattrini, A., Fasolini, M., et al.: Epitopetagged P(0) glycoprotein causes Charcot-Marie-Tooth-like neuropathy in transgenic mice. J. Cell Biol. 151:1035, 2000. 144. Privat, A., Jacque, C., Bourre, J. M., et al.: Absence of the major dense line in the myelin of the mutant mouse “shiverer.” Neurosci. Lett. 12:107, 1979. 145. Rambukkana, A., Zanazzi, G., Tapinos, N., and Salzer, J. L.: Contact-dependent demyelination by Mycobacterium leprae in the absence of immune cells. Science 296:927, 2002. 146. Readhead, C., Popko, B., Takahashi, N., et al.: Expression of a myelin basic protein gene in transgenic shiverer mice: correction of the dysmyelinating phenotype. Cell 48:703, 1987. 147. Ressot, C., and Bruzzone, R.: Connexin channels in Schwann cells and the development of the X-linked form of Charcot-Marie-Tooth disease. Brain Res. Rev. 32:192, 2000. 148. Riethmacher, D., Sonnenberg-Rietmacher, E., Brinkmann, V., et al.: Severe neuropathies in mice with targeted mutations in the ErbB3 receptor. Nature 389:725, 1997. 149. Roach, A., Boylan, K., Horvath, S., et al.: Characterization of cloned cDNA representing rat myelin basic protein:

150. 151. 152.

153.

154.

155. 156.

157. 158. 159.

160.

161.

162.

163.

164.

165.

166.

167.

168.

429 absence of expression in brain of shiverer mutant mice. Cell 34:799, 1983. Rosenbluth, J.: Central myelin in the mouse mutant shiverer. J. Comp. Neurol. 194:639, 1980. Rosenbluth, J.: Peripheral myelin in the mouse mutant shiverer. J. Comp. Neurol. 193:729, 1980. Ryan, M. C., Shooter, E. M., and Notterpek, L.: Aggresome formation in neuropathy models based on peripheral myelin protein 22 mutations. Neurobiol. Dis. 10:109, 2002. Sancho, S., Young, P., and Suter, U.: Regulation of Schwann cell proliferation and apoptosis in PMP22-deficient mice and mouse models of Charcot-Marie-Tooth disease type 1A. Brain 124:2177, 2001. Sanes, J. R., and Lichtman, J. W.: Development of the vertebrate neuromuscular junction. Annu. Rev. Neurosci. 22:389, 1999. Schachner, M.: Neural recognition molecules in disease and regeneration. Curr. Opin. Neurobiol. 4:726, 1994. Schachner, M., and Martini, R.: Glycans and the modulation of neural recognition molecule function. Trends Neurosci. 18:183, 1995. Scherer, S. S.: The biology and pathobiology of Schwann cells. Curr. Opin. Neurol. 10:386, 1997. Scherer, S. S.: Myelination: some receptors required. J. Cell. Biol. 156:13, 2002. Scherer, S. S., and Arroyo, E. J.: Recent progress on the molecular organization of myelinated axons. J. Peripher. Nerv. Syst. 7:1, 2002. Scherer, S. S., Deschênes, S. M., Xu, Y. T., et al.: Connexin32 is a myelin-related protein in the PNS and CNS. J. Neurosci. 15:8281, 1995. Scherer, S. S., Xu, T., Crino, P., et al.: Ezrin, radixin, and moesin are components of Schwann cell microvilli. J. Neurosci. Res. 65:150, 2001. Scherer, S. S., Xu, Y. T., Bannerman, P. G. C., et al.: Periaxin expression in myelinating Schwann cells: modulation by axon-glial interactions and polarized localization during development. Development 121:4265, 1995. Scherer, S. S., Xu, Y.-T., Nelles, E., et al.: Connexin32-null mice develop demyelinating peripheral neuropathy. Glia 24:8, 1998. Schneider-Schaulies, J., von Brunn, A., and Schachner, M.: Recombinant peripheral myelin protein P0 confers both adhesion and neurite outgrowth-promoting properties. J. Neurosci. Res. 27:286, 1990. Schröder, J. M., Bohl, J., and von Bardeleben, U.: Changes of the ratio between myelin thickness and axon diameter in human developing sural, femoral, ulnar, facial, and trochlear nerves. Acta Neuropathol. 76:471, 1988. Schumacher, M., Guennoun, R., Mercier, G., et al.: Progesterone synthesis and myelin formation in peripheral nerves. Brain Res. Brain Res. Rev. 37:343, 2001. Shapiro, L., Doyle, J. P., Hensley, P., et al.: Crystal structure of the extracellular domain from P0, the major structural protein of peripheral nerve myelin. Neuron 17:435, 1996. Sherman, D. L., Fabrizi, C., Gillespie, C. S., and Brophy, P. J.: Specific disruption of a Schwann cell dystrophin-related protein complex in a demyelinating neuropathy. Neuron 30:677, 2001.

430

Neurobiology of the Peripheral Nervous System

169. Shibata, Y., Kumai, M., Nishii, K., and Nakamura, K.: Diversity and molecular anatomy of gap junctions. Med. Electron Microsc. 34:153, 2001. 170. Simons, K., and Ikonen, E.: Functional rafts in cell membranes. Nature 387:569, 1997. 171. Smith-Slatas, C., and Barbarese, E.: Myelin basic protein gene dosage effects the PNS. Mol. Cell. Neurosci. 15:343, 2000. 172. Sohl, G., Eiberger, J., Jung, Y. T., et al.: The mouse gap junction gene connexin29 is highly expressed in sciatic nerve and regulated during brain development. Biol. Chem. 382:973, 2001. 173. Staugaitis, S. M., Colman, D. R., and Pedraza, L.: Membrane adhesion and other functions for the myelin basic proteins. BioEssays 18:13, 1995. 174. Stewart, H. J.: Expression of c-Jun, Jun B, Jun D and cAMP response element binding protein by Schwann cells and their precursors in vivo and in vitro. Eur. J. Neurosci. 7:1366, 1995. 175. Stoffel, W., and Bosio, A.: Myelin glycolipids and their functions. Curr. Opin. Neurobiol. 7:654, 1997. 176. Suter, U., and Snipes, G. J.: Biology and genetics of hereditary motor and sensory neuropathies. Annu. Rev. Neurosci. 18:45, 1995. 177. Suter, U., and Snipes, G. J.: Peripheral myelin protein 22: facts and hypotheses. J. Neurosci. Res. 40:145, 1995. 178. Suter, U., Snipes, G. J., Schoener-Scott, R., et al.: Regulation of tissue-specific expression of alternative peripheral myelin protein-22 (PMP22) gene transcripts by two promoters. J. Biol. Chem. 269:25795, 1994. 179. Tabernero, A., Stewart, H. J. S., Jessen, K. R., and Mirsky, R.: The neuron-glia signal beta neuregulin induces sustained CREB phosphorylation on Ser-133 in cultured rat Schwann cells. Mol. Cell. Neurosci. 10:309, 1998. 180. Tait, S., Gunn-Moore, F., Collinson, J. M., et al.: An oligodendrocyte adhesion molecule at the site of assembly of the paranodal axo-glial junction. J. Cell Biol. 150:657, 2000. 181. Takeichi, M.: The cadherins: cell-cell adhesion molecules controlling animal morphogenesis. Development 102:639, 1988. 182. Takeichi, M.: Cadherins: a molecular family important in selective cell-cell adhesion. Annu. Rev. Biochem. 59:237, 1990. 183. Taylor, V., Zgraggen, C., Naef, R., and Suter, U.: Membrane topology of peripheral myelin protein 22. J. Neurosci. Res. 62:15, 2001. 184. Tifft, C. J., and Proia, R. L.: Stemming the tide: glycosphingolipid synthesis inhibitors as therapy for storage diseases. Glycobiology 10:1249, 2000. 185. Toews, A. D., Hostettler, J., Barrett, C., and Morell, P.: Alterations in gene expression associated with primary demyelination and remyelination in the peripheral nervous system. Neurochem. Res. 22:1271, 1997. 186. Topilko, P., Schneider-Maunoury, S., Levi, G., et al.: Krox20 controls myelination in the peripheral nervous system. Nature 371:796, 1994. 187. Traka, M., Dupree, J. L., Popko, B., and Karagogeos, D.: The neuronal adhesion protein TAG-1 is expressed by Schwann cells and oligodendrocytes and is localized to the

188.

189.

190.

191.

192.

193.

194.

195.

196. 197. 198.

199.

200.

201.

202.

203.

204.

juxtaparanodal region of myelinated fibers. J. Neurosci. 22:3016, 2002. Trapp, B.: Distribution of the myelin-associated glycoprotein and P0 protein during myelin compaction in quaking mouse peripheral nerve. J. Cell Biol. 107:675, 1988. Trapp, B. D., Andrews, S. B., Wong, A., et al.: Co-localization of the myelin-associated glycoprotein and the microfilament components, F-actin and spectrin, in Schwann cells of myelinated nerve fibres. J. Neurocytol. 18:47, 1989. Trapp, B. D., O’Connell, M. F., and Andrews, S. B.: Ultrastructural immunolocalization of MAG and P0 proteins in cryosections of peripheral nerve. J. Cell Biol. 1035:228a, 1986. Trapp, B. D., and Quarles, R. H.: Presence of the myelinassociated glycoprotein correlates with alterations in the periodicity of peripheral myelin. J. Cell Biol. 92:877, 1982. Tropak, M. B., Johnson, P. W., Dunn, R. J., and Roder, J. C.: Differential splicing of MAG transcripts during CNS and PNS development. Brain Res. 464:143, 1988. Vyas, A. A., and Schnaar, R. L.: Brain gangliosides: functional ligands for myelin stability and the control of nerve regeneration. Biochimie 83:677, 2001. Warner, L. E., Mancias, P., Butler, I. J., et al.: Mutations in the early growth response 2 (EGR2) gene are associated with hereditary myelinopathies. Nat. Genet. 18:382, 1998. Webster, H. d.: The geometry of peripheral myelin sheaths during their formation and growth in rat sciatic nerves. J. Cell Biol. 48:348, 1971. Wegner, M.: Transcriptional control in myelinating glia: flavors and spices. Glia 31:1, 2000. Wegner, M.: Transcriptional control in myelinating glia: the basic recipe. Glia 29:118, 2000. Wong, M. H., and Filbin, M. T.: Dominant-negative effect on adhesion by myelin P0 protein truncated in its cytoplasmic domain. J. Cell Biol. 134:1531, 1996. Wu, R., Jurek, M., Sundarababu, S., and Weinstein, D. E.: The POU gene Brn-5 is induced by neuregulin and is restricted to myelinating Schwann cells. Mol. Cell. Neurosci. 17:683, 2001. Xu, W., Shy, M., Kamholz, J., et al.: Mutations in the cytoplasmic domain of P0 reveal a role for PKC-mediated phosphorylation in adhesion and myelination. J. Cell Biol. 155:439, 2001. Yagi, T., and Takeichi, M.: Cadherin superfamily genes: functions, genomic organization, and neurologic diversity. Genes Dev. 14:1169, 2000. Yamashita, T., Higuchi, H., and Tohyama, M.: The p75 receptor transduces the signal from myelin-associated glycoprotein to Rho. J. Cell Biol. 157:565, 2002. Yang, L. J. S., Zeller, C. B., Shaper, N. L., et al.: Gangliosides are neural ligands for myelin-associated glycoprotein. Proc. Natl. Acad. Sci. U. S. A. 93:814, 1996. Yin, X., Crawford, T. O., Griffin, J. W., et al.: Myelinassociated glycoprotein is a myelin signal that modulates the caliber of myelinated axons. J. Neurosci. 18:1953, 1998.

Myelination 205. Yin, X., Kidd, G. J., Wrabetz, L., et al.: Schwann cell myelination requires timely and precise targeting of P0 protein. J. Cell Biol. 148:1009, 2000. 206. Young, P., Boussadia, O., Berger, P., et al.: E-cadherin is required for the correct formation of autotypic adherens junctions of the outer mesaxon but not for the integrity of myelinated fibers of peripheral nerves. Mol. Cell. Neurosci. 21:341, 2002. 207. Young, P., and Suter, U.: Disease mechanisms and potential therapeutic strategies in Charcot-Marie-Tooth disease. Brain Res. Rev. 36:213, 2001.

431

208. Zanazzi, G., Einheber, S., Westreich, R., et al.: Glial growth factor/neuregulin inhibits Schwann cell myelination and induces demyelination. J. Cell Biol. 152:1289, 2001. 209. Ziskind-Conhaim, L.: Physiological and morphological changes in developing peripheral nerves of rat embryos. Dev. Brain Res. 42:15, 1988. 210. Zoidl, G., Blass-Kampmann, S., D’Urso, D., et al.: Retroviral-mediated gene transfer of the peripheral myelin protein PMP22 in Schwann cells: modulation of cell growth. EMBO J. 14:1122, 1995.

20 The Control of Axonal Caliber JOHN W. GRIFFIN AND AHMET HÖKE

Cytology of the Axon Extrinsic Factors That Influence Axonal Caliber Myelination: Alteration of the Axonal Cytoskeleton Signaling Pathways from Myelinating Schwann Cell to Axonal Cytoskeleton Myelin-Associated Glycoprotein: Involvement in Signaling to the Axon

Does Axonal Caliber Influence Myelination? Intrinsic Factors That Influence Axonal Caliber Neurofilaments and Control of Axonal Diameter in Large-Caliber Myelinated Nerve Fibers Neurofilament Transport

One of the cardinal phenotypic features of a neuron is the caliber of its axon. Axonal caliber in the peripheral nervous system (PNS) is precisely regulated, as reflected in the predictable association of fiber sizes with functional categories. The range of normal axonal calibers is noteworthy; a factor of 100 separates the cross-sectional areas of the smallest unmyelinated axons in the human PNS from the largest myelinated ones.7 After maturation caliber can undergo only modest variation, even in disease. For example, an axon 10 ␮m in diameter never withers to 2 ␮m, nor does the reverse occur. Axonal caliber has important functional correlates. In myelinated fibers, the conduction velocity is linearly related to nerve fiber diameter.29 Consequently, in the case of the simplest neuronal circuit, the monosynaptic reflex, the latency (duration) of the reflex changes when axonal caliber is altered.15 The order of recruitment of fibers also correlates with their axonal calibers.44 In addition to these physiologic measures, axonal caliber has obvious structural influences on non-neuronal cells, including the ensheathing Schwann cells. For example, in myelinated fibers myelin sheath thickness and total myelin volume are influenced in some fashion by axonal caliber, and some data suggest that even the initial formation of a myelin sheath may be influenced by caliber.118 Finally, alterations in axonal caliber are prominent and frequent pathologic abnormalities in peripheral nerve diseases and may contribute to such late consequences as secondary demyelination.42

Alterations in Velocity of Neurofilament Transport: Influence on Axonal Neurofilament Content and Axonal Caliber Neurofilament Gene Expression Influences Axonal Caliber Regulation of Neurofilament Gene Expression Conclusions

This chapter considers the mechanisms by which axonal caliber is determined. The influences on axonal caliber are conveniently divided into those extrinsic to the neuron and those intrinsic to it. For example, the synthesis and transport of axonal constituents are considered as intrinsic factors affecting or determining axonal size. However, at some stage of development it is likely that the “decision” of the neuron to grow to its adult proportions is determined by the action of extrinsic factors. Conversely, the presence or absence of myelination around a given axonal segment influences caliber. This “extrinsic” influence is now known to alter posttranslational modifications of intrinsic axonal cytoskeletal proteins.

CYTOLOGY OF THE AXON As viewed in electron micrographs, axoplasm is deceptively simple, containing only a few particulate organelles, mitochondria, and the cytoskeletal elements. Rotaryshadowed, deep-etched preparations give a strikingly different image by emphasizing the innumerable side arms that appear to interlace the cytoskeletal elements to each other as well as to the particulate organelles.12,45,46,116 In both preparations it is apparent that most of the space of the axoplasm appears empty and is occupied by the cytosol. 433

434

Neurobiology of the Peripheral Nervous System

As in other cells, the organization of the cytoplasm appears to be a function of the cytoskeleton.32,62,105 The axonal cytoskeleton includes three elements: microfilaments (MFs), microtubules (MTs), and neurofilaments (NFs). Microfilaments, which are poorly preserved in routine electron micrographs, are prominent in the subaxolemmal cytoskeleton.16,64,68,115 In squid axoplasm, MFs of the inner cytoskeleton are organized in longitudinally oriented columns that are surrounded by a dense matrix and associated with the MT domains.16 Microtubules are present throughout the axoplasm; these tubulin polymers are relatively long, with lengths estimated to be 350 to 750 ␮m.114 Neurofilaments, the major intermediate filaments (IFs) of neurons, deserve special attention in this chapter because of their relationship to axonal caliber. IFs as a class are 10-nm filaments found in the cytoskeletons of eukaryotic cells. Unlike actin filaments and MTs, which are composed of highly conserved protein subunits, the protein composition of IFs varies in different cell types. There are five generally accepted classes of IF that can be

distinguished based on differences in sequence and gene structure.17,32 Class I and class II are composed of type I (acidic) and type II (basic) keratins, respectively, which are expressed in epithelial cells. Class III comprises a diverse group of proteins including vimentin, expressed in mesenchymal cells and embryonic neurons108; glial fibrillary acidic protein, expressed in glial cells97,130; and desmin, expressed in muscle cells. A class III IF especially relevant to the PNS is peripherin, expressed in small sensory and autonomic neurons.82 Class IV comprises the neuronspecific NF proteins, which are reviewed in detail later.97,130 Finally, class V comprises the nuclear lamins. The NFs are composed of three distinct protein subunits, neurofilament light (NF-L, 68 kDa), neurofilament medium (NF-M, 145 kDa), and neurofilament heavy (NF-H, 200 kDa), collectively referred to as the NF triplet.54,67,92 Like all IF proteins, each NF subunit contains a central, structurally conserved, ␣-helical coiled-coil domain divided into three regions (Fig. 20–1).17,32,105 The presence of this conserved rod domain suggests that each of these subunits is associated with the NF core. A unique carboxy-terminal

FIGURE 20–1 Comparison of the structural domains of NF-L, NF-M, and NF-H polypeptides. All three contain an amino-terminal domain approximately 100 amino acids in length and a structurally conserved ␣-helical rod domain, which is approximately 310 amino acids in length and composed of three coiled-coil regions, designated 1a, 1b, and 2. The greatest divergence among NF-L, NF-M, and NF-H resides in the carboxy-terminal tail domains, which in mouse are 140, 435, and 677 amino acids long, respectively.59 The carboxy-terminal tails of NF-M and NF-H contain both glutamic acid–rich regions and multiple repeats (0 to 5 for NF-M and approximately 50 for NF-H) of the KSP tripeptide. The KSP serine residues are NF-specific phosphorylation sites. Locations of mutations in NFs in human disorders are shown by arrows. ALS ⫽ amyotrophic lateral sclerosis; CMT ⫽ Charcot-Marie-Tooth disease. (Adapted from Al-Chalabi, A., and Miller, C. C. J.: Neurofilaments and neurological disease. Bioessays 24:346, 2003.)

The Control of Axonal Caliber

tail contains multiple repeats of the lysine-serine-proline (KSP) sequence. There are 0 to 7 repeats, depending on the species, for NF-M and approximately 50 repeats for NF-H.30,59,63,65,66,74 There are also head domain serine phosphorylation sites.102 Immunocytochemical and in vitro reassembly studies suggest that the tail domain of NF-M and NF-H is associated with side arms that extend from the NF core.30,31,46,47,60,96 As discussed later, these side arms are likely to influence interfilament spacing.

EXTRINSIC FACTORS THAT INFLUENCE AXONAL CALIBER Myelination: Alteration of the Axonal Cytoskeleton Of the extrinsic factors that have been shown to affect axonal caliber, the extent of Schwann cell ensheathment and myelination is the most important in normal nerves. In cultured sensory neurites ensheathment by Schwann cells results in some enlargement of the axons.121 Myelination provides another and greater increment in caliber. This myelination-mediated effect on axonal size is segmental, affecting only the axon beneath the sheath.58 For example, in the unmyelinated initial segment and stem process of dorsal root ganglion neurons, the axon is smaller. It increases in caliber at the first myelinated heminode

435

(Fig. 20–2). Caliber decreases again at the nodes of Ranvier. Demyelination of one or a few segments of a myelinated fiber is associated with loss of axonal caliber in the demyelinated segments,2,86 as initially recognized by Gombault in the 1870s.37 In individual nerve fibers, the myelinated segments have different cytoskeletal organization than the unmyelinated segments, such as the stem process, nodes of Ranvier, and axon terminals. The greater caliber in myelinated segments correlates with greater spacing of neighboring NFs and with greater phosphorylation of NF-H and NF-M.11,58,131 In the past the predominant hypothesis was that negative charges on the phosphorylated carboxy-terminal tails of NF-H molecules might be repulsive, leading to greater separation of filaments, thereby reducing NF density and hence axonal caliber. Many data seemed to confirm this relationship. Heavy phosphorylation of NF-H is not seen in the neuron cell bodies, the dendrites, the initial segment, or the nodes of Ranvier,106 and in these sites the interfilament distances are less than in myelinated segments.46 These data predict that elimination of the KSP repeats on NF-H would prevent widely spaced NFs. Surprisingly, removal by genetic engineering of the 150 carboxy-terminal amino acids of NF-H, thus removing all of the KSP sites, had no effect on NF spacing, radial axonal growth, ultimate axonal caliber, or conduction velocity.87,88 In contrast, eliminating the seven KSP repeats on the murine NF-M molecule produced partially collapsed neurofilament

FIGURE 20–2 Myelination results in increased phosphorylation of neurofilaments at the first heminode. A, Schematic diagram of a primary sensory neuron showing the perikaryon, the glomerular stem process, the myelinated internode, and the bifurcation into central and peripheral processes. The dotted line describes a plane of section similar to that shown in B, with the perikaryon (P), nonmyelinated stem process (S), and myelinated internode (m) of the same neuron shown in a single section. C, Graph of the ratio of phosphorylated to phosphorylation-independent epitopes in the stem process compared to the myelinated axon. Myelination is associated with a marked increase in the proportion of phosphorylated epitopes. (Modified from Hsieh, S.-T., Kidd, G. J., Crawford, T. O., et al.: Regional modulation of neurofilament organization by myelination in normal axons. J. Neurosci. 14:6392, 1994.)

436

Neurobiology of the Peripheral Nervous System

spacing, slowed growth and reduced calibers, and slower conduction velocities.28 These results appear to exclude simple charge repulsion, because in that model the effect of NF-H phosphorylation would necessarily be greater than that of NF-M.28 Instead, a more rigid side arm may result from phosphorylation of NF-M.28

Signaling Pathways from Myelinating Schwann Cell to Axonal Cytoskeleton The mechanisms by which myelination signals NF phosphorylation involve axonal kinases. Current data suggest that the most important is extracellular signal–related protein kinase 1/2 (ERK1/2). Cdk5, a member of the cyclindependent protein kinase family, can also phosphorylate NF-H and NF-M in vitro. However, the role of Cdk5 in vivo is complex. In mice lacking cdk5 or its activator (p35), the neurofilament proteins, especially NF-M, are hyperphosphorylated, not hypophosphorylated.95 Normally, Cdk/p35 phosphorylates mitogen-activated protein kinase-1 (MEK-1), resulting in a reduction in MEK-1 activity. A recent hypothesis suggests that the myelination-associated signal from Schwann cell to axon may activate the small GTPase Raf, activating MEK-1 and in turn activating ERK1/2, leading to phosphorylation of NFs but also upregulating the Cdk activator p35, resulting in increased p35/Cdk activity and thus feedback inhibition of MEK-1.95 Thus Cdk5 may be more important in limiting than in promoting NF phosphorylation.95 This hypothesis is further complicated by a recent study demonstrating that overexpression of Cdk5 in chick dorsal root ganglion neurons leads to increased NF phosphorylation and reduced axonal transport of NFs.98 Conversely, inhibition of Cdk5 activity reduces NF phosphorylation and enhances NF axonal transport. These findings suggest that Cdk5 activity is tightly linked to NF phosphorylation, transport, and distribution within the axoplasm.

Myelin-Associated Glycoprotein: Involvement in Signaling to the Axon How does the myelin-forming Schwann cell signal to these enzymes? At least part of this signaling involves myelinassociated glycoprotein (MAG). MAG is an intrinsic membrane glycoprotein that is not found in compact myelin, but is found in myelinated internodes. It is localized in several Schwann cell membranes, including the inner and outer mesaxons, the terminal loops, and the site relevant to the present discussion, the adaxonal Schwann cell. MAG is not necessary for myelination,79 but myelinated internodes in MAG⫺/⫺ mice have the smaller caliber, lower NF phosphorylation, and lesser NF spacing that are characteristic of unmyelinated axons.61,132 Thus MAG may represent at least one Schwann cell ligand capable of initiating ERK1/2 activation and thus NF phosphorylation.

The nature of the MAG receptor on the axon has been controversial. The gangliosides GT1b and GD1a appear to interact with MAG,129 and mice lacking complex gangliosides have many of the axonal features of the MAG⫺/⫺ mice, including closer NF spacing.99 However, increasing evidence suggests that there is also a protein receptor, or coreceptor. This might be the p75 low-affinity neurotrophin receptor119,124,128 and the gangliosides may be organized in membrane rafts, so that they cooperate in facilitating MAG signaling and activation of ERK1/2.128 It is likely that Schwann cells also have MAGindependent signaling mechanisms, but the sequence outlined above represents a satisfying if incomplete synthesis of the available data regarding NF spacing. To complete the story of the relationship of myelination to axonal caliber, it is necessary to add that myelinated internodes may contain more NFs than unmyelinated segments. This in turn appears to reflect greater retention of transported NF proteins in myelinated segments. In this aspect of caliber control, phosphorylation of NF-H may play a role; several data suggest that NF-H phosphorylation slows the aggregate transport of NF proteins.1 In contrast, removing the phosphorylation sites of NF-M has no effect on NF transport.87 In sum, an attractive hypothesis is that MAG and perhaps other Schwann cell membrane ligands interact with p75 neurotrophin receptor in ganglioside-containing membrane rafts of the axolemma, initiating the signaling cascade that leads to ERK1/2-mediated phosphorylation of NF-M and NF-H carboxy-terminal domains. NF-M phosphorylation promotes an extension and perhaps a “stiffening” of the NF side arms and thus greater NF spacing. NF-H phosphorylation promotes increased retention of NFs within the myelinated segments, and thus greater NF numbers. Taken together, these processes substantially increase the caliber of myelinated axonal segments. This is reflected in greater conduction velocity. The magnitude of these effects is modeled in Figure 20–3.

DOES AXONAL CALIBER INFLUENCE MYELINATION? The focus so far has been on the effect of myelination on axonal caliber. Conversely, both classic and modern investigators have devoted attention to the effect of axonal caliber on myelination. Two related questions have been the subjects of analyses: to what extent does axonal caliber determine whether Schwann cells initiate myelination of an axon, and to what degree is the amount of myelin formed influenced by axonal caliber? These possibilities are often contrasted with the alternative that specific molecular signals that trigger myelination are present in some axons, and absent in those that do not become myelinated. At the moment much of the available evidence,

The Control of Axonal Caliber

437

FIGURE 20–3 The effects of modification of neurofilament organization by myelination on nerve fiber diameters and conduction velocities. The unmyelinated axon at the top represents the mean diameter of stem processes in L5 dorsal root ganglion of the rat. The open arrow to the left illustrates the consequence of simply adding a myelin sheath to that stem process, with values calculated so that the axonal diameter/fiber diameter ratio (the G ratio) equals 0.6. The conduction velocity (in meters per second) was calculated as the total fiber diameter (axon ⫹ myelin sheath) ⫻ 6.109,120 The solid arrow to the right illustrates the features that were observed by electron microscopic morphometry.58 Note that myelination is associated with increased neurofilament spacing and number, so that the diameter of the axon is 6 ␮m, the fiber diameter is 10 ␮m, and the calculated conduction velocity is 2.7-fold greater than without the modification of the axon by myelination. (Adapted from Hsieh S.-T., Kidd, G. J., Crawford, T. O., et al.: Regional modulation of neurofilament organization by myelination in normal axons. J. Neurosci. 14:6392, 1994.)

reviewed below, can be interpreted in favor of either hypothesis, so that the influence of axonal caliber on myelination should be regarded as an unresolved issue. The onset of myelin formation correlates with radial growth of the axon. This was an early argument in favor of the “caliber” hypothesis—that a critical diameter triggers myelination. This concept has been supported20,118 and critiqued43 (see Chapter 19). Two recent studies examined models in which normally unmyelinated axons were found to myelinate. In the first study, a murine salivary gland was partially denervated by a partial injury of the external carotid branch of the superior cervical ganglion. As a result, the uninjured, normally unmyelinated axons in this branch sprouted to reinnervate denervated salivary gland cells, and the individual unmyelinated axons thereby took on extended fields of innervation.118 The unmyelinated axons in this branch were observed to increase in caliber. An attractive speculation is that this axonal enlargement occurred because the same amount of target-derived growth factors were being retrogradely

transported to fewer nerve cells, so that each neuron had a relative increase in its trophic support. The pivotal finding was that the enlarged unmyelinated axons became segregated into a 1:1 relationship with their Schwann cells, and many subsequently myelinated. This result is consistent with the possibility that increasing axonal caliber may be sufficient to initiate myelin formation. However, it is possible that the same stimulus that induced axonal hypertrophy also changed the neuronal phenotype in other ways, so that a new molecular signal for myelination was expressed. The second study involved exogenous administration of large doses of the growth factor glial cell line–derived neurotrophic factor (GDNF).56 A population of unmyelinated axons, including some C-fiber nociceptor axons, normally express the low-affinity GDNF receptor GFR␣-1 and the high-affinity receptor c-Ret. Giving high daily doses of GDNF resulted in proliferation of Schwann cells (Fig. 20–4), followed by segregation of some unmyelinated axons into 1:1 relationships with Schwann cells and subsequent

438

Neurobiology of the Peripheral Nervous System

FIGURE 20–4 Administration of exogenous glial cell line–derived neurotrophic factor (GDNF) results in proliferation of Schwann cells. Toluidine blue–stained 1-␮m-thick sections of intact sciatic nerves of control vehicle (A) and high-dose recombinant human GDNF (rhGDNF)–injected (B) rats. In rhGDNFinjected animals the myelinated fibers were widely separated and the intervening space contained an increased number of unmyelinated Schwann cells. Examples of Schwann cell nuclei are marked by arrows in B. Mitotic figure (arrow) in a Schwann cell in a rhGDNF-injected animal is shown among other Schwann cell nuclei (asterisks) in the inset (original magnification, ⫻1000). (From Hoke, A., Ho, T., Crawford, T. O., et al.: Glial cell line-derived neurotrophic factor alters axon Schwann cell units and promotes myelination in unmyelinated nerve fibers. J. Neurosci. 23:561, 2003.)

myelination (Fig. 20–5). In GDNF-treated animals there was abundant myelination of very small fibers, and some fibers were clearly undergoing early stages of myelination, with only a few wraps of noncompacted myelin. That these fibers were originally axons of unmyelinated fibers was suggested by their presence within a single continuous basal lamina that also encircled unmyelinated fibers; no such profiles occur normally. Notably, these myelinating profiles occur in intact nerves with neither nerve fiber degeneration nor, more importantly, axonal sprouting. The segregated axons were larger, consistent with the possibility that the process was driven by axonal hypertrophy. However, the same caveat—that GDNF might have altered the neuronal phenotype in more than one fashion—remains. In addition, the possibility of a direct effect of GDNF on Schwann cells cannot be excluded; although the Schwann cells lacked c-Ret, GDNF signaling can occur though GFR␣-1.84,113 Nevertheless, the role of caliber in myelination is very complex. In tissue culture, Windebank et al.121 found that axons increase in size as they segregate into a 1:1 relationship, and there is an additional increase in caliber as they myelinate. Nonmyelinated segments of axons are smaller, have fewer neurofilaments, and have lower levels of phosphorylation of NF-H than myelinated segments of the same axons.58 Demyelination results in loss of caliber and NF-H phosphorylation.11 Does axonal caliber influence the amount of internodal myelin? In a given nerve, the ratio of axonal diameter to total nerve fiber diameter (the outer surface of the myelin sheath) is relatively constant.35 This ratio, termed the G ratio, is typically around 0.6 (see Fig. 20–3). However, this value is not invariant normally,19,102 and it is altered in a variety of pathologic conditions. In remyelinated or regenerated

fibers, it is low (the myelin is relatively thin).18,22,90,93,101 In axonal atrophy the G ratio is higher, and in axonal swelling it is reduced.24,25 It is clear that axonal circumference does not in itself dictate the number of myelin lamellae or the total myelin volume, although it appears to have an influence. Although differing in details, two lines of argument have shown that internodal distance is also an important factor. Thus internodes on regenerated and remyelinated fibers are relatively short, as are internodes near the dorsal root ganglion perikarya and in other areas where longitudinal growth is not as great (e.g., the sciatic nerve).21,23,101 These short internodes have relatively thin sheaths. Thus Smith and colleagues concluded that the determinant of total myelin surface area or myelin volume is that total axolemma surface area covered by the internode.101 There are no data indicating how this measure is read by the myelinating Schwann cell, or how signaling to alter myelin production might occur. The changes in G ratio that occur with atrophy or swelling of axons have been interpreted to reflect “slippage” of myelin lamellae, such that, in swelling, lamellae slide over each other radially to produce a thinner sheath, and in atrophy, the sliding produces a thicker sheath.25 However, the data are better explained by constancy of number of lamellae as the axon changes, with “passive” changes in the G ratio, except in paranodal regions. In these areas lamellae appear to slide over each other longitudinally, so that the sheath gets progressively thin as it approaches the node. This pattern reflects the slippage of the innermost myelin terminal loops into the old internode, so that only the outermost myelin lamellae still reach their original attachment sites next to the node (see Fig. 20–7C later).

The Control of Axonal Caliber

439

FIGURE 20–5 Segregation of axon–Schwann cell units and myelination of unmyelinated axons. A, Electron micrograph of an intact sciatic nerve of a control rat at 4 weeks. An example of an unmyelinated fiber is outlined by arrowheads on the basal lamina. This fiber contains 14 axons (examples identified by “a”; SC ⫽ Schwann cell nucleus). (Original magnification, ⫻5000.) B, Electron micrograph of an intact sciatic nerve of a high-dose GDNF-treated rat at 4 weeks. Note the marked decrease in the axon–Schwann cell ratio and numerous Remak bundles with one or two axons per Schwann cell basal lamina (examples outlined by arrowheads). (Original magnification, ⫻5000.) C, Low-power electron micrograph of the sciatic nerve of GDNF-treated rat shows examples of newly myelinated small axons (original magnification, ⫻3000). D, Some of the small myelinated fibers were still within Remak bundles that also contained unmyelinated axons. Arrowheads point to portions of the basal lamina that are continuous between a myelinated axon and an unmyelinated one. (Original magnification, ⫻25,000.) E, Histograms of the number of axons within a single Schwann cell basal lamina showed a dramatic shift to a higher percentage of Remak bundles attaining a 1:1 relationship when the rats were treated with rhGDNF for 4 weeks (top) compared to controls (bottom). (Adapted from Höke, A., Ho, T., Crawford, T. O., et al.: Glial cell line-derived neurotrophic factor alters axon Schwann cell units and promotes myelination in unmyelinated nerve fibers. J. Neurosci. 23:561, 2003.)

INTRINSIC FACTORS THAT INFLUENCE AXONAL CALIBER Neurofilaments and Control of Axonal Diameter in Large-Caliber Myelinated Nerve Fibers Several lines of evidence indicate that, in large fibers, NF content is closely related to axonal caliber. During radial growth and maturation in normal postnatal development, the composition of the axonal cytoskeleton changes to

include a higher proportion of NFs. In early development axons are thin (⬍1 ␮m in diameter), their cytoskeletons have high proportions of MTs and peripherin filaments but few NFs, and a greater proportion of their cross-sectional area is occupied by membranous elements, including mitochondria and agranular reticulum. The relatively few IFs in these axons are often organized in discrete bundles.6,83,107 Even in adults the cross-sectional areas of unmyelinated axons correlate closely with the content of MTs and membranous elements.107 With radial growth in development, NFs become the predominant cytoskeletal

440

Neurobiology of the Peripheral Nervous System

organelle, outnumbering MTs by more than 10:1 in large axons.55 In large myelinated fibers NFs are the most numerous cytoskeletal elements, and axonal crosssectional area is linearly related to their number.26,27 In large axons MTs and membranous organelles occupy only a small fraction of axonal cross-sectional area.6 The extent to which the content of MTs and membranous organelles correlates with caliber represents a major difference between myelinated and unmyelinated axons. The relationship between NF number and axon crosssectional area could be interpreted to suggest that NF density is fixed and immutable. In fact, even in normal nerves, modest differences in axonal NF density (approximately 25%) have been reported among different populations of neurons85 and in different regions of small-caliber nerve fibers.75,80 During developmental growth, as well as in injury-related atrophy of axons (see later), NF densities stay relatively constant.51,52 In contrast, in some types of axonal swelling, marked increases in NF densities can be found (Fig. 20–6). A number of genetically engineered models have shown that, for altered NF protein expression to alter axonal caliber, the normal stoichiometric relationships between NF-L, NF-H, and NF-M need to be maintained.34,72,123 For example, the content of the NF core protein, NF-L, can be changed substantially without affecting axonal caliber. NF density can increase two- to threefold without compensatory increases in axonal

caliber in transgenic mice overexpressing NF-L alone.72 Because this increase in NF-L occurs in the absence of comparable increases in either NF-M or NF-H, increased NF density in these transgenic axons correlates with reductions in the relative proportions of NF-M and NF-H, which presumably reflect a decrease in the relative number of NF side arms. In mice in which NF entry into the axons is defective, marked reductions in caliber are found,14 and a similar finding is present in Japanese quail deficient in NF proteins.126,127

Neurofilament Transport Because axons lack cytoplasmic ribosomes, most axonal proteins are synthesized in the neuron cell body (soma). However, some pol(A) messenger RNA (mRNA) is present in axons, including message for NF proteins. The total contribution of local axonal synthesis to the total protein economy of the axon has been estimated to be 0.5% to 5%,57 so that perikaryally derived proteins must represent most of the axonal proteins. Pulse-labeled cytoskeletal proteins, including the NF triplet, are slowly transported somatofugally within the axon at rates of several millimeters per day.54 The mechanism of slow transport is incompletely understood, but appears to be MT and kinesin based. The slow aggregate rate represents stationary periods. The state of polymerization of the transported NF proteins remains to be determined.4,78

FIGURE 20–6 Electron micrographs illustrating factors that influence NF density. A, Axoplasm in the sciatic nerve of a normal mouse; this nontransgenic mouse was a littermate of the transgenic mouse whose axoplasm is shown in B. B, Axoplasm from a transgenic mouse overexpressing NF-L. The relative overabundance of NF-L (compared to NF-M and NF-H) correlates with an increase in NF density and greater variability in the orientation of NF. C, An even greater increase in NF density is seen in axons of rat sciatic nerve after systemic intoxication with IDPN. Also note that MT and membranous organelles have become segregated from regions of axoplasm highly enriched in NF. Bars: 0.5 ␮m.

The Control of Axonal Caliber

NF synthesis or the amount of newly synthesized NF protein entering the axon.81 After IDPN intoxication, newly synthesized NFs are unable to move distally after entering the axon. This leads to massive accumulation of NFs in the proximal swellings8–10,39,40 (Fig. 20–7; see also Fig. 20–6). The rapidity with which these axonal swellings form is related to velocity of NF transport,

Alterations in Velocity of Neurofilament Transport: Influence on Axonal Neurofilament Content and Axonal Caliber Systemic intoxication with the neurotoxin ␤,␤⬘iminodipropionitrile (IDPN) severely impairs the transport of NF in axons41 without altering either the level of

A

441

B

FIGURE 20–7 Diagrammatic representation of the cell body and proximal axon of motor neurons in normal (A) and ␤,␤⬘-iminodipropionitrile (IDPN)–intoxicated (B) animals. A, The normal motor neuron is shown giving rise to a thin axon that is myelinated over the first three internodes by oligodendroglia; the last two internodes diagrammed are ensheathed by Schwann cells. The subpial astrocytes (stippled) give rise to a basal lamina, the glial limitans, which is continuous with that of the Schwann cells (dashed line) in the roots. A few pial cells are shown outside the glial limitans. B, In the early stage of IDPN intoxication, the first few internodes undergo fusiform enlargement. The initial segment and the nodes of Ranvier are normal or only moderately enlarged. The internodal axoplasm, shown in cross section, contains a central channel surrounded by whorled NF. The axonal swelling results in attenuation of the internodal myelin sheath and paranodal demyelination. The subpial astrocytes show proliferation of glial filaments. (From Griffin, J. W., Cork, L. C., Hoffman, P. N., and Price, D. L.: Experimental models of motor neuron degeneration. In Dyck, P. J., Thomas, P. K., Lambert, E. H., and Bunge, R. [eds.]: Peripheral Neuropathy, 2nd ed. Philadelphia, W. B. Saunders, p. 621, 1984.) C, Slippage of paranodal myelin lamella in IDPN-induced swelling of the axon. Ventral roots from IDPN-intoxicated animals were teased and immunostained with anti-Caspr antibody. Arrows point to retraction of the terminal myelin loops from the paranodal region. Note the larger size of the ventral roots from IDPN-treated animals demonstrating axonal swelling. (Original magnification ⫻630.) (Courtesy of Thien Nguyen.)

442

Neurobiology of the Peripheral Nervous System

which declines with age. Swellings form more rapidly in young than in old animals. No swellings occur in the NF-deficient Japanese quail when given IDPN,69 and the degree of IDPN-induced axonal swelling is substantially less in the proximal stumps of axotomized nerves than in control nerves (as described later, axotomy reduces the level of NF gene expression). These data suggest that the amount of new NF synthesis by neurons also affects their vulnerability to IDPN. The greater the level of NF gene expression and the greater the velocity of NF transport, the greater the size of the neurofilamentous masses induced by IDPN. Neurofilament-rich axonal swellings, similar to those produced by IDPN intoxication, are also found in a number of other pathologic situations. Examples include the human motor neuron disease amyotrophic lateral sclerosis.7 Structurally similar swellings occur with a different distribution in the heritable disorder giant axonal neuropathy,5 and in the toxic neuropathies caused by 2,5-hexanedione,103,104 carbon disulfide,71 aluminum,33,110 and paclitaxel (Taxol).89 In these disorders the swellings tend to begin in distal regions of the nerves.103,104 In experimental 2,5-hexanedione neuropathy the swellings have been correlated with a speeding of NF transport.70,71

Neurofilament Gene Expression Influences Axonal Caliber The level of NF gene expression in mature neurons correlates with axonal NF content and axonal caliber. Similarly, small spinal sensory neurons contain low levels of NF mRNAs and give rise to unmyelinated axons.50 In contrast, neurons that contain relatively high levels of NF mRNAs (e.g., large spinal sensory neurons) give rise to largecaliber myelinated nerve fibers, which are highly enriched in NF.50 Given the correlation between the level of NF gene expression and axonal caliber in mature neurons, it is not surprising that postnatal increases in NF gene expression correlate closely with the radial growth and myelination of developing axons.49,73,91 In developing sensory neurons, postnatal increases in NF gene expression also correlate with increases in perikaryal size.73 Studies of NF gene expression in axotomized neurons also indicate that axonal caliber correlates closely with the level of NF gene expression. Selective reductions in the abundance of NF mRNAs in axotomized neurons50 correlate with declines in (1) NF synthesis (as reflected by reduced amino incorporation),38 (2) the amount of pulselabeled NF protein undergoing axonal transport,55 (3) axonal NF content,52 and (4) axonal caliber in the proximal stumps of transected nerve fibers.52 These reductions in caliber start proximally near the cell body (soma) and proceed somatofugally along nerve fibers at a rate equal to the velocity of NF transport, a process referred to as somatofugal axonal atrophy.48,52,53 Thus alterations in NF

gene expression have a profound influence on both axonal NF content and axonal caliber. Similar reductions in NF content also occur in the axonal atrophies associated with hereditary motor and sensory neuropathy type I in humans76 and streptozotocin-induced diabetic neuropathy in experimental animals.125 It should be noted that axotomy-induced reductions in axonal caliber are confined primarily to large-caliber myelinated nerve fibers (e.g., alpha motor fibers). There is little, if any, change after axotomy in the calibers of relatively small myelinated axons (e.g., gamma motor fibers).52 In fact, substantial changes in axonal NF content are associated with only modest changes in the crosssectional areas of small-caliber myelinated axons such as those in the optic nerve,75,80 indicating that factors other than NF content play important roles in the control of axonal caliber in these nerve fibers. Neurofilaments, class IV IF proteins, appear to play a unique role in the control of axonal caliber. The temporal expression and distribution of peripherin, a class III IF protein expressed in neurons,17,32,82,105 suggest that its role is different from that of NF. In contrast to NF, peripherin is expressed at relatively high levels (in comparison to those in mature neurons) during early development94,111 and axon regeneration.77,112 These findings indicate that the radial growth of sensory axons during postnatal development correlates with a decline in the abundance of peripherin in large sensory neurons, and the somatofugal atrophy of regenerating axons is associated with increased expression of peripherin. Furthermore, unlike NF, which is expressed at highest levels in large sensory neurons, peripherin is expressed preferentially in small sensory neurons.13,111

Regulation of Neurofilament Gene Expression Although the factors that regulate NF gene expression are largely unknown, the available evidence suggests that interactions with target cells may play an important role in this process. Disconnection of neurons from their targets by axotomy results in downregulation of NF gene expression. Reconnection of regenerating axons with targets appears to be a prerequisite for the return of gene expression to preaxotomy levels.50,52 These observations have led to the suggestion that tropic molecules derived from target cells and retrogradely transported to the neuron cell body may play an important role in this process.50 Although the ability of exogenous nerve growth factor (NGF) to induce NF expression in sensory neurons with high-affinity NGF receptors is controversial,117,122 exogenous NGF does appear to prevent, at least in part, axotomy-induced reductions in the calibers of sensory axons.36 A role for target cells in the control of NF expression may also be inferred from the observation that the radial growth and subsequent myelination of sympathetic axons (which are

The Control of Axonal Caliber

not normally myelinated) can be stimulated by forcing these neurons to innervate an increased number of target cells.118 14.

CONCLUSIONS There is an ongoing dialogue between Schwann cells and the neurofilamentous cytoskeleton. The state of myelination of the axon by the Schwann cell influences NF spacing and numbers, and thus axonal caliber. Conversely, some evidence supports the concept that the radial growth of axons is at least one stimulus for myelin formation, and that axonal caliber may be one influence, along with internodal length, on the volume of myelin formed.

REFERENCES 1. Ackerley, S., Thornhill, P., Grierson, A. J., et al.: Neurofilament heavy chain side arm phosphorylation regulates axonal transport of neurofilaments. J. Cell Biol. 161:489, 2003. 2. Aguayo, A. J., Attiwell, M., Trecarten, J., et al.: Abnormal myelination in transplanted trembler mouse Schwann cells. Nature 265:73, 1977. 3. Al-Chalabi, A., and Miller, C. C. J.: Neurofilaments and neurological disease. Bioessays 24:346, 2003. 4. Angelides, K. J., Smith, K. E., and Takeda, M.: Assembly and exchange of intermediate filament proteins of neurons: neurofilaments are dynamic structures. J. Cell Biol. 108:1495, 1989. 5. Asbury, A. K., Gale, M. K., Cox, S. C., et al.: Giant axonal neuropathy: a unique case with segmental neurofilamentous masses. Acta Neuropathol. 20:237, 1972. 6. Berthold, C. H.: Morphometry of normal peripheral axons. In Waxman, S. G. (ed.): Physiology and Pathobiology of Axons. New York, Raven Press, p. 3, 1978. 7. Carpenter, S.: Proximal axonal enlargement in motor neuron disease. Neurology 18:841, 1968. 8. Chou, S.-M., and Hartman, H. A.: Axonal lesions and waltzing syndrome after IDPN administration in rats: with a concept—“axostasis.” Acta Neuropathol. 3:428, 1964. 9. Chou, S.-M., and Hartman, H. A.: Electron microscopy of focal neuroaxonal lesions produced by ␤,␤⬘-iminodipropionitrile (IDPN) in rats. Acta Neuropathol. 39:590, 1965. 10. Clark, A. W., Griffin, J. W., and Price, D. L.: The axonal pathology in chronic IDPN intoxication. J. Neuropathol. Exp. Neurol. 39:42, 1980. 11. deWaegh, S. M., Lee, V. M. Y., and Brady, S. T.: Local modulation of neurofilament phosphorylation, axonal caliber, and slow axonal transport by myelinating Schwann cells. Cell 68:451, 1992. 12. Ellisman, M. H., and Porter, K. R.: Microtrabecular structure of the axoplasmic matrix: visualization of cross-linking structures and their distribution. J. Cell Biol. 87:464, 1980. 13. Escurat, M., Djabali, K., Gumpel, M., et al.: Differential expression of two neuronal intermediate filament proteins, peripherin and the low-molecular-mass neurofilament

15. 16.

17.

18.

19.

20.

21.

22.

23. 24.

25.

26.

27.

28.

29.

30.

31.

443

protein (NF-L), during the development of rat. J. Neurosci. 10:764, 1990. Eyer, J., and Peterson, A.: Neurofilament-deficient axons and perikaryal aggregates in viable transgenic mice expressing a neurofilament–␤-galactosidase fusion protein. Neuron 12:389, 1994. Farel, P. B.: Reflex activity of regenerating frog spinal motorneurons. Brain Res. 158:331, 1978. Fath, K. R., and Lasek, R. J.: Two classes of actin microfilaments are associated with the inner cytoskeleton of axons. J. Cell Biol. 107:613, 1988. Franke, W. W.: Nuclear lamins and cytoskeletal intermediate filament proteins: a growing multigene family. Cell 48:3, 1987. Fried, K., Hildebrand, C., and Eldelyi, G.: Myelin thickness and internodal length of nerve fibers in developing feline inferior alveolar nerve. J. Neurol. Sci. 54:47, 1982. Friede, R., Meier, T., and Diem, M.: How is the exact length of an internode determined? J. Neurol. Sci. 50:217, 1981. Friede, R. L.: Control of myelin formation by axonal caliber with a model of the control mechanism. J. Comp. Neurol. 144:233, 1972. Friede, R. L.: Variance in relative internodal length (l/d) in the rat and its presumed significance for the safety factor and neuropathy. J. Neurol. Sci. 60:89, 1983. Friede, R. L., and Beuche, W.: A new approach toward analyzing peripheral nerve fiber populations. I. Variance in sheath thickness corresponds to different geometric proportions of the internodes. J. Neuropathol. Exp. Neurol. 44:60, 1985. Friede, R. L., and Bischhausen, R.: The precise geometry of large internodes. J. Neurol. Sci. 48:367, 1980. Friede, R. L., and Martinez, A. J.: Analysis of axon-sheath relations during early wallerian degeneration. Brain Res. 19:199, 1970. Friede, R. L., and Martinez, A. J.: Analysis of the process of sheath expansion in swollen nerve fibers. Brain Res. 19:165, 1970. Friede, R. L., and Miyaghishi, T.: Axon caliber, neurofilaments, microtubules, sheath thickness and cholesterol in cat optic nerve fiber. Anat. Rec. 108:365, 1971. Friede, R. L., and Samorajski, T.: Axonal caliber related to neurofilaments and microtubules in sciatic nerve fibers of rats and mice. Anat. Rec. 167:379, 1971. Garcia, M. L., Lobsiger, C. S., Shah, S. B., et al.: NF-M is an essential target for the myelin-directed “outside-in” signaling cascade that mediates radial axonal growth. J. Cell Biol. 163:1011, 2003. Gasser, H. S., and Grundfest, H.: Axon diameters in relation to the spike dimensions and the conduction velocity in mammalian A fibers. Am. J. Physiol. 127:393, 1939. Geisler, N., Kaufmann, E., Fischer, S., et al.: Neurofilament architecture combines structural principles of intermediate filaments with carboxy-terminal extensions increasing in size between triplet proteins. EMBO J. 2:1295, 1983. Geisler, N., and Weber, K.: Self-assembly in vitro of the 68,000 molecular weight component of the mammalian neurofilament triplet proteins into intermediate-size filaments. J. Mol. Biol. 151:565, 1981.

444

Neurobiology of the Peripheral Nervous System

32. Geisler, N., and Weber, K.: Structural aspects of intermediate filaments. In Shay, J. W. (ed.): Cell and Molecular Biology of the Cytoskeleton. New York, Plenum Publishing, p. 41, 1986. 33. Gilbert, M. R., Harding, B. L., Hoffman, P. N., et al.: Aluminum-induced neurofilamentous changes in cultured rat dorsal root ganglia explants. J. Neurosci. 12:1763, 1992. 34. Gill, S. R., Wong, P. C., Monteiro, M. J., et al.: Assembly properties of dominant and recessive mutations in the small mouse neurofilament (NF-L) subunit. J. Cell Biol. 111:2005, 1990. 35. Gillespie, M. J., and Stein, R. B.: The relationship between axon diameter, myelin thickness and conduction velocity during atrophy of mammalian peripheral nerves. Brain Res. 259:41, 1983. 36. Gold, B. G., Mobley, W. C., and Matheson, S. F.: Regulation of axonal caliber, neurofilament content, and nuclear localization in mature sensory neurons by nerve growth factor. J. Neurosci. 11:943, 1991. 37. Gombault, F. A. A.: Contribution a l’histoire anatomique de l’atrophie musculaire saturnine. Arch. Physiol. V:592, 1873. 38. Greenberg, S. G., and Lasek, R. J.: Neurofilament protein synthesis in DRG neurons decreases more after peripheral axotomy than after central axotomy. J. Neurosci. 8:1739, 1988. 39. Griffin, J. W., Cork, L. C., Hoffman, P. N., et al.: Experimental models of motor neuron degeneration. In Dyck, P. J., Thomas, P. K., Lambert, E. H., et al. (eds.): Peripheral Neuropathy, 3rd ed., Vol. I. Philadelphia, W. B. Saunders, p. 621, 1984. 40. Griffin, J. W., Gold, B. G., Cork, L. C., et al.: IDPN neuropathy in the cat: coexistence of proximal and distal axonal swellings. Neuropathol. Appl. Neurobiol. 8:351, 1982. 41. Griffin, J. W., Hoffman, P. N., Clark, A. W., et al.: Slow axonal transport of neurofilament proteins: impairment by ␤,␤⬘-iminodipropionitrile administration. Science 202:633, 1978. 42. Griffin, J. W., and Price, D. L.: Demyelination in experimental IDPN and hexacarbon neuropathies: evidence for an axonal influence. Lab. Invest. 45:130, 1981. 43. Griffin, J. W., Rosenfeld, J., Hoffman, P. N., et al.: The axonal cytoskeleton: influences on nerve fiber form and Schwann cell behavior. In Lasek, R. (ed.): Intrinsic Determinants of Neuronal Form and Function. New York, Alan R. Liss, p. 403, 1988. 44. Henneman, E., Somjen, G., and Carpenter, D. O.: Excitability and inhibitability of motoneurons of different sizes. J. Neurophysiol. 28:599, 1965. 45. Hirokawa, N.: Cross-linker system between neurofilaments, microtubules, and membranous organelles in frog axons revealed by the quick-freeze, deep-etching method. J. Cell Biol. 94:129, 1982. 46. Hirokawa, N., Glicksman, M. A., and Willard, M. B.: Organization of mammalian neurofilament polypeptides within the neuronal cytoskeleton. J. Cell Biol. 98:1523, 1984. 47. Hisanaga, S.-I., and Hirokawa, N.: Structure of the peripheral domains of neurofilaments revealed by low angle rotary shadowing. J. Mol. Biol. 202:297, 1988. 48. Hoffman, P. N.: Distinct roles of neurofilament and tubulin gene expression in axonal growth. Ciba Found. Symp. 138:192, 1988.

49. Hoffman, P. N.: Expression of GAP-43, a rapidly transported protein, and class II beta tubulin, a slowly transported protein, are coordinated in regenerating neurons. J. Neurosci. 9:893, 1989. 50. Hoffman, P. N., Cleveland, D. W., Griffin, J. W., et al.: Neurofilament gene expression: a major determinant of axonal caliber. Proc. Natl. Acad. Sci. U. S. A. 84:3472, 1987. 51. Hoffman, P. N., Griffin, J. W., Cold, B. G., et al.: Slowing of neurofilament transport and the radial growth of developing nerve fibers. J. Neurosci. 5:2920, 1985. 52. Hoffman, P. N., Griffin, J. W., and Price, D. L.: Control of axonal caliber by neurofilament transport. J. Cell Biol. 99:705, 1984. 53. Hoffman, P. N., Koo, E. H., Muma, N. A., et al.: Role of neurofilaments in the control of axonal caliber in myelinated nerve fibers. In Lasek, R. J., and Black, M. M. (eds.): Intrinsic Determinants of Neuronal Form and Function. New York, Alan R. Liss, p. 389, 1988. 54. Hoffman, P. N., and Lasek, R. J.: The slow component of axonal transport: identification of major structural polypeptides of the axon and their generality among mammalian neurons. J. Cell Biol. 66:351, 1975. 55. Hoffman, P. N., Thompson, G. W., Griffin, J. W., et al.: Changes in neurofilament transport coincide temporally with alterations in the caliber of axons in regenerating motor fibers. J. Cell Biol. 101:1332, 1985. 56. Hoke, A., Ho, T., Crawford, T. O., et al.: Glial cell linederived neurotrophic factor alters axon Schwann cell units and promotes myelination in unmyelinated nerve fibers. J. Neurosci. 23:561, 2003. 57. Hollenbeck, P. J., and Hollenbeck, P. J.: Organization and translation of mRNA in sympathetic axons. J. Cell Sci. 116:4467, 2003. 58. Hsieh, S.-T., Kidd, G. J., Crawford, T. O., et al.: Regional modulation of neurofilament organization by myelination in normal axons. J. Neurosci. 14:6392, 1994. 59. Julien, J.-P., Cote, F., Beaudet, L., et al.: Sequence and structure of the mouse gene coding for the largest neurofilament subunit. Gene 68:307, 1988. 60. Julien, J.-P., and Mushynski, W. E.: The distribution of phosphorylation sites among identified proteolytic fragments of mammalian neurofilaments. J. Biol. Chem. 258:4019, 1983. 61. Kumar, S., Yin, X., Trapp, B. D., et al.: Role of long-range repulsive forces in organizing axonal neurofilament distributions: evidence from mice deficient in myelin-associated glycoprotein. J. Neurosci. Res. 68:681, 2002. 62. Lasek, R. J.: Studying the intrinsic determinants of neuronal form and function. In Lasek, R. J. (ed.): Intrinsic Determinants of Neuronal Form and Function. New York, Alan R. Liss, 1988. 63. Lees, J. F., Shneidman, P. S., Skuntz, S. F., et al.: The structure and organization of the human heavy neurofilament subunit (NF-H) and the gene encoding it. EMBO J. 7:1947, 1988. 64. Letourneau, P.: Differences in the organization of actin in the growth cones compared with the neurites of cultured neurons from chick embryos. J. Cell Biol. 97:963, 1983. 65. Levy, E., Liem, R. K. H., D’Eustachio, P., et al.: Structure and evolutionary origin of the gene encoding mouse NF-M,

The Control of Axonal Caliber

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

the middle-molecular-mass neurofilament protein. Eur. J. Biochem. 166:71, 1987. Lewis, S. A., and Cowan, N. J.: Anomalous placement of introns in a member of the intermediate filament multigene family: an evolutionary conundrum. Mol. Cell Biol. 6:1529, 1986. Liem, R. K. H., Yen, S.-H., Salomon, G. D., et al.: Intermediate filaments in nervous tissue. J. Cell Biol. 79:637, 1978. Metuzals, J., and Tasaki, I.: Subaxolemmal filamentous network in the giant nerve fiber of the squid (Loligo pealei) and its possible role in excitability. J. Cell Biol. 78:597, 1978. Mitsuishi, K., Takahashi, A., Mizutani, M., et al.: ␤,␤⬘Iminodipropionitrile toxicity in normal and congenitally neurofilament-deficient Japanese quails. Acta Neuropathol. (Berl.) 86:578, 1993. Monaco, S., Autilio-Gambetti, L., Zabel, D., et al.: Giant axonal neuropathy: acceleration of neurofilament transport in optic axons. Proc. Natl. Acad. Sci. U. S. A. 82:920, 1985. Monaco, S., Jacob, J., Jenich, H., et al.: Axonal transport of neurofilament is accelerated in peripheral nerve during 2,5-hexanedione intoxication. Brain Res. 491:328, 1989. Monteiro, M. J., Hoffman, P. N., Gearhart, J. D., et al.: Expression of NF-1 in both neuronal and nonneuronal cells of transgenic mice: increased neurofilament density in axons without affecting caliber. J. Cell Biol. 111:1543, 1990. Muma, N. A., Slunt, H. H., and Hoffman, P. N.: Postnatal increase in neurofilament gene expression and the radial growth of axons. J. Neurocytol. 20:844, 1991. Myers, M. W., Lazzarini, R. A., Lee, V. M. Y., et al.: The human mid-size neurofilament subunit: a repeated protein sequence and the relationship of its gene to the intermediate filament gene family. EMBO J. 6:1617, 1987. Nixon, R. A., and Logvinenko, K. B.: Multiple fates of newly synthesized neurofilament proteins: evidence for a stationary neurofilament network distributed nonuniformly along axons of retinal ganglion cell neurons. J. Cell Biol. 102:647, 1986. Nukada, H., and Dyck, P. J.: Decreased axonal caliber and neurofilaments in hereditary motor and sensory neuropathy, type I. Ann. Neurol. 16:238, 1984. Oblinger, M. M., Wong, J., and Parysek, L. M.: Axotomyinduced changes in the expression of a type III neuronal intermediate filament. J. Neurosci. 9:3766, 1989. Ochs, S., Jersild, A. Jr., and Li, J.-M.: Slow transport of freely movable cytoskeletal components shown by beading partition of nerve fibers in the cat. Neuroscience 33:421, 1989. Owens, G. C., and Bunge, R. P.: Schwann cells infected with a recombinant retrovirus expressing myelin-associated glycoprotein antisense RNA do not form myelin. Neuron 7:565, 1991. Parhad, I. M., Clark, A. W., and Griffin, J. W.: The effect of changes in neurofilament content on caliber of axons: the ␤,␤⬘-iminodipropionitrile model. J. Neurosci. 7:2256, 1987. Parhad, I. M., Swedberg, E. A., Hoar, D. I., et al.: Neurofilament gene expression following ␤,␤⬘-iminodipropionitrile (IDPN) intoxication. Mol. Brain Res. 4:293, 1988. Parysek, L. M., and Goldman, R. D.: Distribution of a novel 57 kDa intermediate filament (IF) protein in the nervous system. J. Neurosci. 8:555, 1988.

445

83. Peters, A., and Vaughn, J. E.: Microtubules and filaments in the axons and astrocytes of early postnatal rat optic nerves. J. Cell Biol. 32:113, 1967. 84. Poteryaev, D., Titievsky, A., Sun, Y. F., et al.: GDNF triggers a novel ret-independent Src kinase family-coupled signaling via a GPI-linked GDNF receptor alpha1. FEBS Lett. 463:63, 1999. 85. Price, R. L., Paggi, P., Lasek, R. J., et al.: Neurofilaments are spaced randomly in the radial dimensions of axons. J. Neurocytol. 17:55, 1988. 86. Raine, C. S., Wisniewski, H., and Prineas, J.: An ultrastructural study of experimental demyelination and remyelination. Lab. Invest. 21:316, 1969. 87. Rao, M. V., Campbell, J., Yuan, A., et al.: The neurofilament middle molecular mass subunit carboxyl-terminal tail domain is essential for the radial growth and cytoskeletal architecture of axons but not for regulating neurofilament transport rate. J. Cell Biol. 163:1021, 2003. 88. Rao, M. V., Garcia, M. L., Miyazaki, Y., et al.: Gene replacement in mice reveals that the heavily phosphorylated tail of neurofilament heavy subunit does not affect axonal caliber or the transit of cargoes in slow axonal transport. J. Cell Biol. 158:681, 2002. 89. Roytta, M., Horwitz, S. B., and Raine, C. S.: Taxol-induced neuropathy: short-term effects of local injection. J. Neurocytol. 13:685, 1984. 90. Sanders, F. K.: The thickness of myelin sheaths of normal and regenerating peripheral nerve fibers. Proc. Natl. Acad. Sci. Lond. 135:323, 1948. 91. Schlaepfer, W. W., and Bruce, J.: Simultaneous up-regulation of neurofilament proteins during the postnatal development of the rat nervous system. J. Neurosci. 25:39, 1990. 92. Schlaepfer, W. W., and Freeman, L. A.: Neurofilament proteins of rat peripheral nerve and spinal cord. J. Cell Biol. 78:653, 1978. 93. Schroder, J. M.: Alteration between axonal diameter and myelin sheath thickness and regenerated fibers. Brain Res. 35:49, 1972. 94. Seymour, M. A., and Oblinger, M. M.: Differential regulation of type III and IV intermediate filament genes and tubulin genes in developing hamster brain. J. Cell Biol. 107:462a, 1988. 95. Sharma, P., Veeranna, Sharma, M., et al.: Phosphorylation of MEK1 by cdk5/p35 down-regulates the mitogen-activated protein kinase pathway. J. Biol. Chem. 277:528, 2002. 96. Sharp, G. A., Shaw, G., and Weber, K.: Immunoelectronmicroscopical localization of the three neurofilament triplet proteins along neurofilaments of cultured dorsal root ganglion cell neurones. Exp. Cell Res. 137:403, 1982. 97. Shaw, G., Osborn, M., and Weber, K.: An immunofluorescence microscopical study of neurofilament triplet proteins, vimentin, and glial fibrillary acidic protein within the adult rat brain. J. Cell Biol. 26:68, 1981. 98. Shea, T. B., Yabe, J. T., Ortiz, D., et al.: Cdk5 regulates axonal transport and phosphorylation of neurofilaments in cultured neurons. J. Cell Sci. 117:933, 2004. 99. Sheikh, K. A., Sun, J. L. Y., Kawai, H., et al.: Mice lacking complex gangliosides develop wallerian degeneration and myelination defects. Proc. Natl. Acad. Sci. U. S. A. 96:7532, 1999.

446

Neurobiology of the Peripheral Nervous System

100. Sheng, J. G., Zhu, S. G., Jones, R. A., et al.: Interleukin-1 promotes expression and phosphorylation of neurofilament and tau protein in vivo. Exp. Neurol. 163:388, 2000. 101. Smith, K. J., Blakemore, W. F., Murray, J. A., et al.: Internodal myelin volume and axon surface area: a relationship determining myelin thickness? J. Neurol. Sci. 55:231, 1982. 102. Spencer, P. S., Raine, C. S., and Wisniewski, H.: Axon diameter and myelin thickness: unusual relationships in dorsal root ganglia. Anat. Rec. 176:225, 1973. 103. Spencer, P. S., and Schaumburg, H. H.: Central-peripheral distal axonopathy—the pathogenesis of dying-back polyneuropathies. Prog. Neuropathol. 3:253, 1976. 104. Spencer, P. S., and Schaumburg, H. H.: Ultrastructural studies of the dying back process. III. The evolution of experimental peripheral giant axonal degeneration. J. Neuropathol. Exp. Neurol. 36:276, 1977. 105. Steinert, P. M., and Roop, D. R.: Molecular and cellular biology of intermediate filaments. Annu. Rev. Biochem. 57:593, 1988. 106. Sternberger, L. A., and Sternberger, N. H.: Monoclonal antibodies distinguish phosphorylated and nonphosphorylated forms of neurofilaments in situ. Proc. Natl. Acad. Sci. U. S. A. 80:6126, 1983. 107. Stevens, J. K., Trogadis, J., and Jacobs, J. R.: Development and control of axial neurite form: a serial electron microscopic analysis. In Lasek, R. J., and Black, M. M. (eds.): Intrinsic Determinants of Neuronal Form and Function. New York, Alan R. Liss, p. 115, 1988. 108. Tapscott, S. J., Bennett, G. S., Toyama, Y., et al.: Intermediate filament proteins in the developing chick spinal cord. Dev. Biol. 86:40, 1981. 109. Thomas, P. K., Barthold, C.-H., and Ochoa, J.: Microscopic anatomy of the peripheral nervous system. In Dyck, P. J., Thomas, P. K., Griffin, J. W., et al. (eds.): Peripheral Neuropathy, 3rd ed., Vol. 1. Philadelphia, W. B. Saunders, p. 28, 1993. 110. Troncoso, J. C., Price, D. L., Griffin, J. W., et al.: Neurofibrillary axonal pathology in aluminum intoxication. Ann. Neurol. 12:278, 1982. 111. Troy, C. M., Brown, K., Greene, L. A., et al.: Ontogeny of the neuronal intermediate filament protein, peripherin, in the mouse embryo. J. Neurosci. 36:217, 1990. 112. Troy, C. M., Muma, N. A., Greene, L. A., et al.: Contrasting regulation of peripherin and neurofilament expression in motor neuron regeneration. Brain Res. 529:232, 1990. 113. Trupp, M., Scott, R., Whittemore, S. R., et al.: Ret-dependent and -independent mechanisms of glial cell line-derived neurotrophic factor signaling in neuronal cells. J. Biol. Chem. 274:20885, 1999. 114. Tsukita, S., and Ishikawa, H.: The cytoskeleton in myelinated axons: serial section study. Biomed. Res. 2:424, 1981. 115. Tsukita, S., Tsukita, S., Kobayashi, T., and Matsumoto, G.: Subaxolemmal cytoskeleton in squid giant axon. II. Morphological identification of microtubule- and microfilament-associated domains of axolemma. J. Cell Biol. 102:1710, 1986.

116. Tsukita, S., Usukura, J., and Ishikawa, H.: The cytoskeleton in myelinated axons: a freeze-etch replica study. Neuroscience 7:2135, 1982. 117. Verge, V. M. K., Tetzlaff, W., Bisby, M. A., et al.: Influence of nerve growth factor on neurofilament gene expression in mature primary sensory neurons. J. Neurosci. 10:2018, 1990. 118. Voyvodic, J. T.: Target size regulates calibre and myelination of sympathetic axons. Nature 342:430, 1989. 119. Wang, K. C., Kim, J. A., Sivasankaran, R., et al.: p75 interacts with the Nogo receptor as a co-receptor for Nogo, MAG and OMgp. Nature 420:748, 2002. 120. Waxman, S. G.: Determinants of conduction velocity in myelinated nerve fiber. Muscle Nerve 3:141, 1980. 121. Windebank, A. J., Word, P., Bunge, R. P., et al.: Myelination determines the caliber of dorsal root ganglion neurons in culture. J. Neurosci. 6:1563, 1985. 122. Wong, J., and Oblinger, M. M.: NGF rescues substance P expression but not neurofilament or tubulin gene expression in axotomized sensory neurons. J. Neurosci. 11:543, 1991. 123. Wong, P. C., and Cleveland, D. W.: Characterization of dominant and recessive assembly-defective mutations in mouse neurofilament NF-M. J. Cell Biol. 111:1987, 1990. 124. Wong, S. T., Henley, J. R., Kanning, K. C., et al.: A p75(NTR) and Nogo receptor complex mediates repulsive signaling by myelin-associated glycoprotein. Nat. Neurosci. 5:1302, 2002. 125. Yagihashi, S., Kamijo, M., and Watanabe, K.: Reduced myelinated fiber size correlates with loss of axonal neurofilaments in peripheral nerve of chronically streptozotocin diabetic rats. Am. J. Pathol. 136:1365, 1990. 126. Yamasaki, H., Bennett, G. S., Itakura, C., et al.: Defective expression of neurofilament protein subunits in hereditary hypotrophic axonopathy of quail. Lab. Invest. 66:734, 1992. 127. Yamasaki, H., Itakura, C., and Mizutani, M.: Hereditary hypotrophic axonopathy with neurofilament deficiency in a mutant strain of Japanese quail. Acta Neuropathol. (Berl.) 82:427, 1991. 128. Yamashita, T., Higuchi, H., and Tohyama, M.: The p75 receptor transduces the signal from myelin-associated glycoprotein to Rho. J. Cell Biol. 157:565, 2002. 129. Yang, L. J. S., Zeller, C. B., Shaper, N. L., et al.: Gangliosides are neuronal ligands for myelin-associated glycoprotein. Proc. Natl. Acad. Sci. U. S. A. 93:814, 1996. 130. Yen, S.-H., and Fields, K. L.: Antibodies to neurofilament, glial filament and fibroblast intermediate filament proteins bind to different cell types in the nervous system. J. Cell Biol. 88:115, 1981. 131. Yeung, K. B., Thomas, P. K., King, R. H. M., et al.: The clinical spectrum of peripheral neuropathies associated with benign monoclonal IgM, IgG and IgA paraproteinemia: comparative clinical, immunological and nerve biopsy findings. J. Neurol. 238:383, 1991. 132. Yin, X., Crawford, T. O., Griffin, J. W., et al.: Myelinassociated glycoprotein is a myelin signal that modulates the caliber of myelinated axons. J. Neurosci. 18:1953, 1998.

21 Guidance of Axons to Targets in Development and in Disease HARALD WITTE AND FRANK BRADKE

Growth Cones and Cytoskeletal Dynamics Neurite Outgrowth and Steering Actin Dynamics Axon Guidance in Development Ephrin/Eph

Slit/Robo Netrin/DCC/UNC-5 Semaphorin/Plexin Axon Guidance in Disease Epilepsy

Neurons must connect with the proper partners to establish functional circuits. How do neuronal axons grow to their specific targets? Ramon y Cajal and Speidel proposed one of the first theories to explain the amazing wiring capacity of the nervous system,294,351 suggesting that axonal growth was directional and that axons were specifically growing on their routes to their specific target. Sperry’s later experiments provided support for this hypothesis, showing that regenerating retinal ganglion cells correctly target their axons to the optic tectum.352 These experiments led Sperry to formulate his “chemoaffinity theory” postulating that specific ligands and receptors would allow the axons to find and connect to the appropriate target. The molecular identity of these signals, however, remained elusive until recently. With the identification of four main classes of axon guidance ligand-receptor systems (ephrin/Eph, Slit/Robo, netrin/DCC/UNC-5, and semaphorin/plexin), the last 15 years have seen explosive progress in elucidating the mechanisms of wiring within the nervous system. Here we describe these systems in detail, giving special attention to the signaling cascades and molecular downstream targets associated with them. For a better understanding of the signaling events involved, we start the chapter with a general description of the underlying cell biologic mechanisms regulating axonal growth. Recent studies have also revealed that molecules involved in axon guidance—either at the level of ligand-receptor interactions or in the downstream signaling—account for

Schizophrenia CRASH Syndrome/MASA Syndrome/ L1 Syndrome Kallmann’s Syndrome Outlook

various diseases. Progress in unveiling the mechanisms behind these syndromes is discussed in the last part of the chapter.

GROWTH CONES AND CYTOSKELETAL DYNAMICS In this section we introduce the reader to the growth cone, its steering mechanisms, and the molecules involved in axonal growth or halt. Many of the molecular mechanisms underlying cellular locomotion have been elucidated using the migrating fibroblast system. Migrating fibroblasts have also been studied through the years as a more accessible alternative to neuronal growth cones for understanding axonal growth and locomotion. Despite the many differences between these two systems, it has been astonishing to find many similarities on a molecular level. Hence, to allow a better overall view, we include current research development on actin dynamics in migrating fibroblasts in this section.

Neurite Outgrowth and Steering The Growth Cone: The Steering Apparatus of the Axon A unique structure at the leading edge of growing axons is the growth cone, first described by Ramon y Cajal in 1890.294 Growth cones are equipped with a large set of receptors to react to a multitude of guidance signals. These guidance cues are integrated in growth cones, allowing the 447

448

Neurobiology of the Peripheral Nervous System

shaft into the central domain. The peripheral domain contains very few microtubules but instead has a dense network of cross-linked, dynamic actin filaments. The actin filaments of the periphery form filopodia, thin membrane extensions emerging from the leading edge of the growth cone, and lamellipodia, fan-shaped structures. Filopodia and lamellipodia explore the extracellular environment (see Fig. 21–1A) with dynamic advance and retraction movements, and mediate guidance responses to external cues.76 Upon depolymerization of the actin filaments in filopodia and lamellipodia, axon growth is disoriented.24,61 Apart from their role in guiding the axons, actin organization and dynamics appear to play a major role in restraining microtubules from entering the peripheral zone. The dense actin meshwork excludes particles with a diameter of 24 nm from the peripheral zone, including organelles and macromolecules.229 Thus microtubules may be

growing axon to correctly navigate to its ultimate target over long distances. Growth cones contain the machinery required for both translation and protein degradation48 and are able to adjust the receptor composition on their surface according to the local environment. Guidance cues can be attractive or repulsive; that is, an axon will grow toward or away from the source of the guidance signal, respectively. Guidance cues act either by direct contact (e.g., between the axonal growth cone and a repelling environment) or as diffusible factors that are secreted by the target region. Thus they act as short-range or long-range cues. Growth Cone Cytoskeleton Dynamics A growth cone consists of three distinct domains: the central domain, the peripheral domain, and the transition zone in between these two108,178 (Fig. 21–1A). Bundles of microtubules and organelles emanate from the axonal B

A

C

FIGURE 21–1 Growth cone structure and function. A, The central domain of the growth cone is rich in organelles and stable microtubules, while the peripheral domain is characterized by a dense network of cross-linked, highly mobile actin filaments and only few, individual microtubules (for clarity, only a few actin filaments are shown). Actin filaments and microtubules are subject to disassembly and turnover in the transition zone. Filopodia and lamellipodia emanating from the peripheral domain probe the environment of the growth cone and mediate the response to guidance cues. B, Actin dynamics cause highly dynamic advance and retraction of filopodia. G-actin molecules add to the distally oriented plus (⫹) ends (barbed ends) of preexisting actin fibers. These fibers undergo myosin-dependent retrograde transport. Attractive guidance cues stimulate F-actin assembly at the tips of filopodia and reduce the rate of retrograde F-actin flow, while repulsive guidance cues act conversely. C, In the transition zone, actin filaments are depolymerized. At the minus (⫺) end (pointed end), actin monomers dissociate from the filaments. The released actin monomers repolymerize at the tips of filopodia.

Guidance of Axons to Targets in Development and in Disease

restrained from entering the peripheral zone by steric hindrance.108 Indeed, depletion of actin filaments using actindepolymerizing drugs allows microtubules to protrude into the peripheral zone of the growth cone in neurons of the sea slug Aplysia.108 Moreover, destabilization of actin filaments allows hippocampal neurons in cell culture to form supernumerary axons.35,36 Microtubule protrusion into the peripheral domain may—alternatively or additionally—be prevented by signaling between actin and microtubulebinding proteins.390,418 Despite the dense peripheral actin network, individual microtubules explore the peripheral domain along actin filaments326 (see Fig. 21–1A) in a process of alternation of polymerization and depolymerization phases, a state referred to as dynamic instability. Actin filaments also undergo treadmilling: actin monomers are added to the plus (barbed) ends of existing filaments at the leading edge (Fig. 21–1B), and they dissociate from the minus (pointed) ends (Fig. 21–1C). Retrograde F-actin flow driven by myosin motors mediates the transport of actin filaments

449

centripetally away from the leading edge (peripheral domain) (see Fig. 21–1B) toward the transition zone where actin bundles are depolymerized108,227 (see Fig. 21–1C). Subsequently, the actin monomers diffuse to the leading edge, where they re-form actin filaments by polymerization. The growth of microtubules toward the periphery of the growth cone is opposed by this retrograde actin flow. Along with actin filaments, microtubules are transported toward the transition zone, where microtubule turnover occurs.326 Cytoskeletal dynamics are subject to tight regulation during axon growth and guidance. As an example, contact between beads coated with Aplysia cell adhesion molecules and cultured Aplysia neuronal growth cones induces growth cone turning toward the site to which the bead is attached.361 On the level of the cytoskeleton, actin polymerizes at the site of bead attachment and retrograde actin filament flow is locally attenuated. Part of the peripheral domain adjacent to the bead thus becomes devoid of actin filaments. Protruding microtubules from the central domain invade this area, resulting in a net turning response (Fig. 21–2).

FIGURE 21–2 Remodeling of the cytoskeleton during axon guidance (all images refer to the same growth cone). A, Upon contact between a bead coated with the anti–Aplysia cell adhesion molecule (apCAM) antibody (4E8) and the peripheral domain of an Aplysia bag cell growth cone, the central domain of the growth cone extends towards the bead (arrowhead marks initial boundary of the central domain) (B). C and D, Rhodamine-phalloidin labeling of F-actin (C) and ␤-tubulin immunostaining (D) at the 5-minute time point. (Arrowhead marks the position of the bead, visible as a ring because the protein A coating of the bead also binds the secondary antibody). Microtubules (visualized by video-enhanced contrast–differential-interference contrast microscopy) protrude into the peripheral domain as the F-actin network is destabilized in the direction of the interaction axis, that is, toward the guidance signal. E, Pseudocolor overlay of F-actin (red) and tubulin (green) stainings. Note that the actin network in the remaining peripheral domain almost completely excludes microtubules. (From Suter, D. M., Errante, L. D., Belotserkovsky, V., and Forscher, P.: The Ig superfamily cell adhesion molecule, apCAM, mediates growth cone steering by substrate-cytoskeletal coupling. J. Cell Biol. 141:227, 1998 by copyright permission of The Rockefeller University Press and Paul Forscher, Yale University.) See Color Plate

450

Neurobiology of the Peripheral Nervous System

Summary Guidance cues regulate cytoskeletal polymerization, depolymerization, and the retrograde F-actin flow to modulate the turning behavior of a growing axon (see Fig. 21–1B). A major challenge in recent years was to decipher the signaling events that lead to downstream cytoskeletal changes upon ligand receptor binding. Because changes in the actin dynamics seem to be one of the major forces to orchestrate axonal growth and guidance, we next discuss the molecules that are involved in changing actin dynamics by modulating the polymerizing, depolymerizing, or treadmilling rate of actin filaments.

Actin Dynamics Rho GTPases: Key Modulators of Actin Dynamics The role of members of the Rho family of small GTPases (Rho GTPases) in actin dynamics has been studied extensively in non-neuronal cells, in which they control various actin-dependent cellular processes, including migration, phagocytosis, secretion, cell division, and polarity phenomena (for review, see Etienne-Manneville and Hall98 and Raftopoulou and Hall291). Importantly, there is also broad evidence for their involvement in the regulation of cytoskeletal dynamics in growth cones (for review, see Dickson85 and Luo230). Rho GTPases cycle between a GDP-bound, inactive state and a GTP-bound, active form, which functions by interacting with a broad variety of downstream effectors such as kinases or scaffold proteins28 (Fig. 21–3). The activity of Rho GTPases is controlled by three groups of regulatory proteins (see Fig. 21–3): guanine nucleotide exchange factors (GEFs),60,328 guanine nucleotide dissociation inhibitors,276 and GTPase-activating proteins (GAPs).283 Some of these regulators show specificity for one or two Rho GTPases, whereas others are less specific (Fig. 21–4). As is discussed later, axon guidance receptors act, either directly or indirectly, as regulators of the Rho GTPases. The Rho GTPases most comprehensively analyzed are Cdc42, Rac, and RhoA. In cells from the Swiss 3T3 fibroblast cell line, microinjection of constitutively active Cdc42 triggers the formation of filopodia,202,272 constitutively active Rac promotes the formation of lamellipodia,300 and constitutively active RhoA induces the formation of contractile actin/myosin filaments, the stress fibers.299 Similarly, inactivation of Rho induces loss of actin filaments and enhances neurite outgrowth in PC12 cells (derived from adrenal gland tumors), N1E-115 neuroblastoma cells, and cerebellar and hippocampal neurons, whereas presumed activation of Rho by lysophosphatidic acid causes growth cone collapse and neurite retraction.29,74,175,176,269 In contrast, Rac and Cdc42 appear to be positive regulators of axonal growth.42,185,231,232 Interestingly,

FIGURE 21–3 Regulation of Rho GTPase activity. Rho GTPases cycle between an inactive GDP-bound and an active GTP-bound state. Activity of Rho GTPases is controlled by GTPase-activating proteins (GAPs), guanine nucleotide dissociation inhibitors (GDIs), and guanine nucleotide exchange factor (GEFs), which are themselves subject to regulation by guidance signals.

Rac and Cdc42 negatively regulate Rho in cells from the NIH 3T3 fibroblast cell line.319 If a similar correlation exists in neurons, one could assume that the migrating growth cone of a growing axon with active filopodia and lamellipodia contains more active Cdc42 and Rac and less active Rho, reflected by high Cdc42-GTP/Cdc42-GDP and RacGTP/Rac-GDP ratios as well as a low Rho-GTP/Rho-GDP ratio. An axon that has stopped growing should show the opposite. The Rho-GTPases act on the actin cytoskeleton via various downstream targets. These actin-associated molecules are discussed next. Actin Polymerization and Nucleation Actin-modulating proteins change polymerization rate, depolymerization rate, and motility of actin filaments. We first discuss molecules involved in actin polymerization and nucleation.

Guidance of Axons to Targets in Development and in Disease

451

FIGURE 21–4 Downstream effectors of Rho GTPases. Rho GTPases modulate cytoskeletal dynamics by multiple downstream effectors (for details see text). Actin polymerization and depolymerization and retrograde flow of F-actin are the main parameters influenced by Rho GTPases to regulate growth cone dynamics. Generally, effectors of Cdc42 and Rac mediate growth cone attraction and neurite growth, whereas RhoA targets cause growth cone repulsion and collapse (for details see text).

Profilin, Thymosin, and ROCK. The actin-binding proteins profilin and thymosin ␤4 are key regulators of actin polymerization because they control the addition of actin monomers to growing (plus) ends of actin filaments.51,53,183 Actin monomers are sequestered by the actin monomer–binding protein thymosin ␤4 to prevent their spontaneous polymerization.53,313 Like thymosin ␤4, profilin also associates with monomeric actin. It has been suggested that profilin releases actin monomers from thymosin ␤4 and allows their assembly onto the plus ends of actin filaments, thereby enhancing actin polymerization.183 Regulation of profilin activity is a means for Rho GTPases to influence actin assembly at the plus ends of actin filaments. Rho kinase (ROCK), a downstream effector kinase of activated Rho, complexes with and phosphorylates the brain-specific profilin II (see Fig. 21–4). Activation of this pathway increases actin stability and inhibits neuritogenesis74,401 (compare with

discussion in Rho GTPases: Key Modulators of Actin Dynamics earlier). Nucleation of Actin Filaments by Members of the Formin Family. In contrast to the polymerization of actin filaments, its initial step, the nucleation of the filaments de novo, is not well characterized yet. Only recently has the identification of formin-1 as a nucleator of actin filaments in yeast shed some light on the process of actin nucleation.99,315 Formins recruit actin monomers via their proline-rich, profilin-bound, G-actin–binding domain.196,316 Mammalian diaphanous (mDia), a member of the formin protein family that also interacts with the insulin receptor tyrosine kinase substrate protein (IRSp53),115 provides a link between profilin and the GTPase Rho in fibroblasts.204,389 mDia as part of a Rho/mDia complex might have a role similar to that of formin-1 in Rho-mediated actin polymerization (see Fig. 21–4). Indeed, mDia is necessary

452

Neurobiology of the Peripheral Nervous System

for stromal cell–derived factor-1␣–mediated axon formation in cerebellar granule cells6 (compare with Fig. 21–5 later in the chapter). Activation of Arp2/3 by WASP/Scar Family Proteins. The Arp2/3 complex nucleates branching of actin filaments upon activation by nucleation-promoting factors of the Wiskott-Aldrich syndrome protein (WASP)/Scar family, including neuronal WASP (N-WASP) and Scar/WASP-family verprolin homologous protein (WAVE)31,391 (see Fig. 21–4). N-WASP and Scar are effectors of the Rho-GTPases Cdc42 and Rac, respectively. N-WASP and Cdc42 associate directly via the GTPase binding domain/Cdc42/Rac interactive binding region (CRIB) of N-WASP,1,311 and IRSp53 mediates binding of Scar/WAVE to Rac.248,249 In conjunction with Cdc42 and Rac, respectively, N-WASP and Scar/WAVE act upon profilin and the Arp2/3 complex, which function synergistically to facilitate nucleation and branching of actin filaments.235,236,247,302,357 Promotion of Actin Polymerization by Mena Anticapping Function. Proteins of the Ena/VASP family, including enabled (Ena) of the fruit fly Drosophila,124 vasodilator-stimulated phosphoprotein (VASP),140,141 and mammalian Ena (Mena),125 are important modulators of actin polymerization in processes such as axon guidance, cellular migration, T-cell activation, and phagocytosis (for review, see Krause and colleagues203). Capping proteins associate with the plus ends of actin filaments and terminate actin polymerization.170 Ena/VASP proteins also interact with plus ends and prevent the binding of capping proteins, thereby stimulating actin polymerization by their anti-capping function.22 The Rho GTPase Cdc42 interacts with IRSp53 to trigger the formation of an IRSp53/Mena complex.205 Mena in turn promotes the release of capping proteins from the plus ends of actin filaments, leading to a Cdc42-mediated increase of actin polymerization in fibroblasts22,205 (see Fig. 21–4). Mena is localized at the leading edge of neuronal growth cones, suggesting a similar Mena/VASP-dependent mechanism in neurons.216 Actin Depolymerization In addition to actin polymerization, Rho GTPases also control growth cone dynamics by modulating the depolymerization rate of actin filaments. Members of the actin depolymerizing factor (ADF)/cofilin family, which are key regulators of actin dynamics, have the ability to sever actin filaments and increase filament turnover by enhancing the dissociation rate of actin subunits from the minus end of actin filaments.15,206,321 The F-actin severing protein cofilin,268 which makes up the majority of ADF/cofilin proteins in the human central nervous system (CNS), predominantly localizes to neuronal somata and growth cones.14,252 Cofilin activity is subject to phosphoregulation by the cytosolic serine/threonine kinase LIM kinase (LIMK) and

the phosphatase slingshot7,96,271 (see Fig. 21–4). LIMK and slingshot decrease and increase cofilin activity, respectively. Although upstream regulators of slingshot phosphatase have not been identified yet, modulators of LIMK activity are well described. p21-Activated kinase (PAK) and ROCK both activate LIMK by phosphorylation, thereby providing a link to Rho GTPases.7,92,237,273,358,411 PAK is a downstream effector kinase of Cdc42 and Rac,239 and ROCK is activated by Rho174,221 (see Fig. 21–4). Activation of Rho GTPases leads to phosphorylation and thereby activation of LIMK via PAK or ROCK; activated LIMK in turn promotes an increase in F-actin levels by inactivating the depolymerization activity of cofilin. Multiple lines of experiments suggest that cofilin has an important role in axonal growth and guidance. Overexpression of ADF in cortical neurons causes both increased actin-driven growth cone dynamics and enhanced axonal growth.243 Similarly, overexpression of an active, nonphosphorylatable form of cofilin in dorsal root ganglion (DRG) neurons increases axonal growth rates.96 Moreover, semaphorin 3A (Sema3A)–induced growth cone collapse results in a change in phosphorylation levels of cofilin.3 Together, cell culture studies showed that cofilin and its regulators appear to be central modulators of actin dynamics in axonal growth. It will be interesting now to see whether the respective knockout mice, assuming nonlethality, show changes in axonal growth and guidance. Retrograde F-Actin Flow Retrograde flow of actin filaments driven by nonmuscle myosins is another crucial process in growth cone steering in addition to actin assembly and disassembly.227,361 In this respect, control of myosin activity and actin/myosinmediated contractility provides a means for Rho GTPases to regulate growth cone progression or retraction. The activity of myosin is dependent on the phosphorylation state of the myosin light chain (MLC).182 Phosphorylated myosin shows higher ATPase activity resulting in higher contractility. MLC phosphorylation is controlled by ROCK, MLC kinase (MLCK) and MLC phosphatase (MLCP), which are themselves subject to regulation by Rho GTPases (see Fig. 21–4). Rho interacts with MLCP and triggers its inactivation via phosphorylation by ROCK.191 Inactivation of MLCP increases the fraction of phosphorylated, active myosin and leads to an increase in contractility, thus promoting growth cone retraction or collapse. In addition to inactivating MLCP, ROCK directly phosphorylates MLC, also contributing to myosin activation.4 Cdc42 and Rac signaling has the opposite effect on myosin activity. PAK, their shared downstream effector kinase, phosphorylates MLCK, which in turn leads to its inactivation320 (see Fig. 21–4). Consequently, myosin phosphorylation drops and the fraction of dephosphorylated, inactive myosin rises; therefore, Cdc42 and Rac inhibit growth cone retraction.

Guidance of Axons to Targets in Development and in Disease

Summary Actin dynamics control axonal growth and guidance events. We have discussed the key molecular players orchestrating actin dynamics, both those within the growth cone and migrating fibroblasts (including Rho-GTPases), the factors involved in actin nucleation and polymerization (formins/mDia, Arp2/3, Ena/VASP family members, profilin), proteins that depolymerize actin filaments (including members of the ADF/cofilin family), and myosins that generate the molecular force creating retrograde flow of actin filaments within the growth cone. In the next section we present the different ligand-receptor systems involved in axon guidance. The discussion reveals that, despite many differences in the structure and action of these systems, they merge onto similar downstream molecular targets that regulate the actin cytoskeleton. Thus the described actin modulators will reoccur as downstream targets of the receptor signaling.

AXON GUIDANCE IN DEVELOPMENT In this section we present the four main classes of ligandreceptor systems that are currently well described in axon guidance. The vast expansion of this field does not allow coverage of all the receptor and ligand subtypes in a single chapter. We therefore refer the reader to recent reviews for further information. We concentrate here on major and welldescribed signaling events leading to axonal guidance and growth. Other ligand-receptor systems that show changes in pathologic conditions, including the Wnt signaling pathway and cell adhesion receptors such as L1 and N-CAM, are discussed in Axon Guidance in Disease later in this chapter.

Ephrin/Eph The ephrin/Eph family is largely known for its key role in setting up the correct connections in the retinal tectal system, which has been fascinating scientists for over half a century. Axons of retinal ganglion neurons project from the retina to the optical tectum. Depending on the location of the neurons within the retina, their axons arborize at different locations within the optical tectum. Axons from the nasal and temporal retina project to the posterior and anterior tectum, respectively, and axons from the dorsal and ventral retina project to the lateral and medial tectum, respectively. Characterization of the distribution of Eph receptors and ephrin ligands in retinal ganglion cells and the optical tectum showed that ephrin ligands and Eph receptors are present in a gradient and that the concentrations of ligands and receptors give positional information to retinal ganglion axons in the tectum (for review, see Drescher and co-workers,88 Flanagan and Vanderhaeghen,105 and Knoll and Drescher194). Recently, ephrin/Eph signaling was also shown to be involved in the formation of other topographic maps,

453

including the vomeronasal olfactory system, which is necessary for the detection of pheromones.195 Members of the ephrin/Eph family mediate axon pathfinding in other contexts as well, including the guidance of forebrain commissural axons and commissural neurons in the spinal cord.154,208,209,211 There they act as a midline repulsion signal. Receptor Complexes The Eph Receptor Family. With 13 members, the family of Eph tyrosine kinases, receptors for the ephrins, constitutes the largest family of receptor tyrosine kinases in the mammalian genome (reviewed by Kullander and Klein210 and Wilkinson395). Depending on their ligand specificity and sequence similarities, the Eph receptors are divided into two subclasses, EphA and EphB receptors. EphA receptors (EphA1 through EphA8) bind glycosylphosphatidylinositol (GPI)-linked ephrin-As (ephrin-A1 through ephrin-A5), and EphB receptors (EphB1 through EphB4 and EphB6) bind transmembrane ephrin-Bs (ephrin-B1 through ephrin-B3). One exception is EphA4, which binds both A-type and most B-type ephrins. Structure of Eph Receptors. The extracellular part of the Eph receptors consists of a ligand-binding globular domain, a cysteine-rich region, and two fibronectin type III (FNIII) repeats. The cytoplasmic part contains a juxtamembrane domain (with two conserved tyrosine residues), a protein tyrosine kinase domain, a sterile ␣-motif (SAM) domain, and a C-terminal PDZ-binding motif (protein-protein interaction domain mediating interaction with C-terminal polypeptides or other postsynaptic density protein 95, Drosophila discs large tumor suppressor, zonula occludens-1 [PDZ] domains). In conjunction with their ligands, the ephrins, Eph receptors dimerize and form higher order clusters possibly enhanced by their SAM domain.23,155 With both Eph receptors and ephrin ligands being membrane attached, intercellular contact is required for ephrin signaling. In this respect binding by clustered, membrane-bound ephrins, which may facilitate the formation of ephrin-Eph clusters, was shown to be necessary for Eph receptor activity.78 Ephrin Forward Signaling One interesting feature of Eph/ephrin signaling is its bidirectionality. That is, in addition to regular downstream or “forward” signaling (into the Eph receptor–expressing cell), there is also “reverse” signaling (into the ephrin ligand–expressing cell)44,159 (Fig. 21–5). In Eph receptors the nonphosphorylated juxtamembrane domain interacts with the kinase domain and prevents it from adopting its catalytically active conformation.407 Upon ligand binding, autophosphorylation of the receptor dimer leads to phosphorylation of intracellular tyrosine residues within the juxtamembrane domain and the kinase domain activation segment, enabling the Eph receptors to exhibit their biologic and catalytic activity.27,180

454

Neurobiology of the Peripheral Nervous System

FIGURE 21–5 Bidirectional ephrin/Eph signaling. Both ephrin ligands and Eph receptors are linked to downstream effectors that modulate actin dynamics and cellular adhesion. The Eph receptor tyrosine kinases mediate forward signaling and promote axon repulsion. GPI-anchored ephrin-As and transmembrane ephrin-Bs are mediators of reverse signaling, for which ephrin-As require other transmembrane proteins. Ephrin reverse signaling can promote both attractive and repulsive responses.

In line with the broad variety of biologic functions they control, the activated Eph receptors associate with a large number of proteins (reviewed by Kullander and Klein210 and Wilkinson395) (see Fig. 21–5, forward signaling). We focus here on those relevant for axon guidance. Regulation of Rho GTPases by Eph-Associated GEF Ephexin. In a yeast two-hybrid screen for interactors of the cytoplasmic domain of EphA receptors, a novel GEF termed ephexin (Eph-interacting exchange protein) was identified.336 The Dbl-homology/pleckstrin-homology region of ephexin interacts with the C-terminal lobe of the EphA4 kinase domain, and the activity of ephexin is regulated by ephrin-A signaling. Ephrin-A–induced activation of ephexin leads to growth cone collapse in rat

retinal ganglion cells.336 The identification of ephexin provides an important link between EphA receptors and the cytoskeleton in both functional and mechanistic respects. Without ephrin stimulation, the GEF activity of ephexin activates RhoA, Cdc42, and Rac.336 Upon stimulation with ephrin-A, however, the EphA receptor inhibits ephexin activation of Cdc42 and Rac.336 This results in a shift of ephexin activity toward RhoA, which modulates actin dynamics in conjunction with its downstream kinase ROCK to cause growth cone collapse336,385 (see Fig. 21–5; compare with Fig. 21–4). Abl and Arg Kinase. Another link to the actin cytoskeleton is provided by the interaction of EphB2 and EphA4 receptors with Abelson kinase (Abl) (also see Phosphorylated

Guidance of Axons to Targets in Development and in Disease

Ephrin-B: Association with Grb4 later) and Abl-related gene kinase (Arg), two nonreceptor tyrosine kinases of the Abl kinase family. Eph receptors associate directly with Abl and Arg in both a phosphorylation-dependent and phosphorylation-independent manner as well as indirectly via a so far unidentified bridging protein.414 Abl and Arg both possess F-actin binding domains,242,414 and Abl localizes to growth cones in the CNS.70 In addition, Abl tyrosinephosphorylates Mena,367 which in turn stimulates actin polymerization by releasing actin-capping proteins22 (see Promotion of Actin Polymerization by Mena Anti-capping Function, earlier, and Fig. 21–4). FAK and Cas. Another downstream target of Eph receptors that acts on the actin cytoskeleton is focal adhesion kinase (FAK), a regulator of integrin-mediated cell adhesion (see Fig. 21–5, forward signaling). In fibroblasts, stimulation with ephrin-A1 induces cytoskeletal reorganization, which results in enhanced spreading and adhesion.52 Both FAK and Crk-associated tyrosine kinase substrate (Cas) are required for this process. Cas is also implicated in actin filament assembly and integrin-dependent processes such as migration, cell cycle control, and transformation.161,277,324,362 In the absence of ephrin, EphA2 receptors are associated with FAK.245,280 Upon stimulation with ephrin-A1, the protein tyrosine phosphatase Shp2, a regulator of FAK phosphorylation, is recruited to the EphA2 receptor with delayed kinetics (compare with discussion in Tyrosine Phosphatase Shp-2 Binding of UNC-5 later). Subsequently, FAK is rapidly dephosphorylated and thereby inactivated and dissociates from the EphA2 receptor, which may result in the termination of FAK-mediated adhesion. Promotion of Neurite Retraction by Inactivation of the MAPK Pathway. Ephrin-mediated neurite retraction can also be triggered by inactivation of the mitogen-activated protein kinase (MAPK) pathway95,246 (compare with discussion in Promotion of Local Protein Synthesis and Degradation by Netrin Signaling later). As mentioned earlier, activation of the EphB2 receptor leads to autophosphorylation of tyrosine residues within the juxtamembrane domain. Subsequent downregulation of Ras activity may be promoted by Src homology 2 (SH2) domain–mediated binding of p120-RasGAP to EphB2. A functional link between Ras activity and neurite retraction might be the cross talk between the Ras and Rho signaling pathways. In fibroblasts, MAPK signaling activated by Ras mediates inhibition of the Rho target ROCK, leading to increased cell motility.317 Ras inactivated by ephrin signaling no longer activates the MAPK pathway; therefore, ROCK inhibition is released, which will contribute to neurite retraction upon stimulation with ephrins (see Fig. 21–5). In the context of EphB1-dependent cell migration, however, Eph/ephrin signaling activates MAPKs.382 Activated EphB1 recruits the adaptor proteins Grb2 and p52Shc.

455

Subsequently, the tyrosine kinase c-Src as well as the extracellular signal-regulated kinase (ERK) MAPKs are activated, which promotes chemotaxis in human renal microvascular endothelial cells. Ephrin Reverse Signaling In contrast to other signaling pathways, in the case of ephrin signaling not only the receptors (Ephs) but also the ligands (ephrins) have both phosphorylation-dependent and phosphorylation-independent signaling capacity. For example, ephrin-B reverse signaling causes forebrain commissural axons to turn away from EphB-expressing regions.154 Somewhat unexpectedly, in addition to the transmembrane ephrin-Bs, the GPI-anchored ephrin-As also possess this reverse signaling capacity, indicating that ephrin-A reverse signaling functions via so far unidentified coreceptors.79 Recent work on the vomeronasal system suggests that vomeronasal axons receive guidance signals through their ephrin-A “ligands.”195 Phosphorylation of Ephrins by Tyrosine Kinases. Three conserved tyrosine residues within the cytoplasmic domain of ephrin-B ligands can be phosphorylated by Src family tyrosine kinases (SFKs), fibroblast growth factor (FGF) receptor tyrosine kinases, or platelet-derived growth factor receptor tyrosine kinases.63,180,279 Upon Eph/ephrin engagement, the protein-tyrosine phosphatase PTP-BL is recruited to ephrins in a PDZ domain–dependent manner with delayed kinetics279 (see Fig. 21–5, reverse signaling; compare with discussion in Phosphorylation-Independent, PDZ Domain–Mediated Interactions of B-Type Ephrins later). PTP-BL subsequently dephosphorylates ephrin-Bs and thereby antagonizes the phosphorylation by tyrosine kinases. Phosphorylated Ephrin-B: Association with Grb4. The SH2–Src homology 3 (SH3) adaptor protein Grb4, an adaptor protein for multiple regulators of actin dynamics, binds the phosphorylated tyrosine residues of ephrin-Bs by virtue of its SH2 domain, and it also interacts via its SH3 domains with other proteins implicated in cytoskeletal dynamics.72 These Grb4 effectors include Cbl-associated protein (CAP)/ponsin, a regulatory protein in the formation of actin stress fibers and focal adhesions298; the Abl-interacting protein-1 (Abi-1), an interaction partner of Abl tyrosine kinase70,338; dynamin, a GTPase implicated in endocytosis and actin dynamics73,318; PAK1, the effector kinase of Cdc42 and Rac239 (see Fig. 21–4); heterogeneous ribonucleoprotein K, an interactor of the Rac GEF Vav158,165; and axin, a scaffold protein in the Wnt signaling pathway, a pathway recently found to be involved in guidance of commissural axons.143,234,417 What is more, Grb4 can also interact with Abl71 (see Fig. 21–5, reverse signaling; compare with discussion in Abl and Arg Kinase earlier), thereby providing another

456

Neurobiology of the Peripheral Nervous System

link to actin dynamics. Abi-1 promotes the Abl-mediated tyrosine phosphorylation of Mena,367 which in turn modulates actin polymerization by release of actin filament capping proteins22 (see Promotion of Actin Polymerization by Mena Anti-capping Function, earlier, and Fig. 21–4). Phosphorylation-Independent, PDZ Domain–Mediated Interactions of B-Type Ephrins. In addition to phosphorylation-dependent interactions via their SH2 motif, B-type ephrins can also undergo phosphorylationindependent interactions mediated by their PDZ domain (see Fig. 21–5, reverse signaling). Such interactors include glutamate receptor–interacting proteins GRIP1 and GRIP2,43 syntenin,135 the tyrosine phosphatase PTP-BL,153 protein kinase C–interacting protein-1 (Pick1),374 and PDZRGS3.228 PTP-BL releases the phosphorylation of tyrosine residues within the cytoplasmic domain of ephrin-Bs,279 antagonizing the tyrosine phosphorylation by SFKs and thereby the phosphorylation-dependent binding of Grb4 and its interactors. GPI-Linked A-Type Ephrins: Capacity for Reverse Signaling. Although reverse signaling by transmembrane ephrin-B ligands is more evident, GPI-linked ephrin-A ligands also have the capacity for reverse signaling. However, the mechanism by which they execute this effect is not yet understood. GPI-anchored proteins are clustered in membrane microdomains (rafts) via their GPI anchors.5,40 In contrast to EphA2 receptor signaling, which leads to FAK dephosphorylation and reduced integrin-mediated adhesion245 (see FAK and Cas earlier), compartmentalized ephrin-A5 signaling increases phosphorylation of FAK and cellular adhesion in fibroblasts and neuroblastoma cells. In addition, an unidentified 120-kDa membrane protein was also phosphorylated upon ephrin-A5 signaling (see Fig. 21–5, reverse signaling). These ephrin-mediated effects depend on ␤1 integrin and activation of lipid raft–associated Fyn tyrosine kinase, a member of the Src kinase family that is one of the kinases upstream of FAK.79,80,134,168 Modulation of Ephrin/Eph Signaling Three mechanisms have been identified so far that enable targets to modulate their responsiveness to ephrin/Eph signaling. The ephrin-A2 ligand forms a complex with the metalloprotease Kuzbanian, which specifically and locally cleaves ephrin-A2 upon Eph receptor binding. A mutation in ephrin-A2 that inhibits cleavage delays axon withdrawal. Proteolytic cleavage is a common theme in the regulation of response to guidance cues (compare with discussion in Receptor Degradation or Proteolytic Cleavage of Robo and in Proteolytic Cleavage of the DCC Receptor later). It therefore provides a mechanism for efficient control and possibly termination of ephrin-A2 signaling.148 Local protein synthesis provides another means to adjust the surface receptor composition and thereby the

responsiveness of the growth cone to its environment. EphA2 receptor messenger RNA (mRNA) is locally translated in distal segments of axons after they have reached their intermediate targets in the midline. The resulting increase in Eph receptors on the growth cone surface changes the responsiveness to ephrin signals.37 Endocytosis constitutes another possibility to clear ligandreceptor complexes from the growth cone surface. In fibroblasts and endothelial cells, the internalization of EphB4/ephrin-B2 complexes and the subsequent cell retraction require Rac-dependent actin polymerization within the receptor-expressing cells.240 Upon binding of ephrin-B1 to the EphB2 receptor, growth cones of cultured telencephalic or hippocampal neurons detach and withdraw. Subsequently, they internalize the entire ligandreceptor complex.420 Like ephrin signaling, endocytosis is bidirectional, that is, both ligand- and receptor-expressing cell take up the ligand-receptor complex. In vivo this process is likely to be regulated by additional signals, because the C-terminal part of ephrin and the kinase domain of the Eph receptor are necessary for the uptake of the complex into the ephrin- and Eph-expressing cell, respectively. Bidirectional endocytosis of ephrin/Eph complexes represents an efficient mechanism to terminate cellular adhesion and allow growth cone retraction.

Slit/Robo Receptor Complexes Roundabout (Robo), a single-pass transmembrane protein of the immunoglobulin (Ig) superfamily, is the cell surface receptor mediating the repulsive effects of the guidance cue molecule Slit.39,190,223 Initially, Robo was discovered in a genetic screen for mutations that affect the development of CNS axon pathways in Drosophila.332 Axons that are prevented from crossing the midline in wild-type flies can cross and recross in Robo mutants, suggesting that Robo is a component of a repulsive signaling system at the midline, whereas Slit is the repulsive cue located at the midline. Robo Structure and Differential Expression Patterns. The extracellular part of Robo consists of five Ig domains and three FNIII repeats. The cytoplasmic region of Robo contains four conserved motifs (CC0, CC1, CC2, and CC3), and each of these is crucial for certain aspects of Robo signaling.18,354 In Drosophila, the pattern of overlapping expression of its three Robo receptors (Robo-1, -2, and -3, with both Robo-2 and -3 lacking the CC2 and CC3 motifs) in the CNS defines a pattern of differential Slit sensitivity, the Robo code.292,341 These expression domains make up medial, intermediate, and lateral zones within the longitudinal pathways that allow a well-defined medial-to-lateral distribution of longitudinal axon bundles in conjunction with cell adhesion molecules such as fasciclin and neural cadherin (N-cadherin).94,292,297,341,344

Guidance of Axons to Targets in Development and in Disease

Robo receptors seem to have differential expression patterns in vertebrates as well, as recently shown for Robo-1 and -2 in the developing mouse nervous system.360 Robo-1/2 expression in the mouse spinal cord corresponds with the Robo functions proposed for Drosophila; however, the more complex patterns in murine brain suggest additional roles for Robo in vertebrate neural development. Slit Proteins: Ligands of the Robo Receptor. Mutations in the slit genes were first known to cause midline defects in the Drosophila CNS308; later on, Slit proteins were identified as repulsive guidance cue molecules signaling through the Robo receptor.39,190,223 Slit proteins constitute a family of guidance molecules structurally conserved between species as distant as insects, nematodes, and mammals. In mammals, three Slit proteins (Slit-1, -2, and -3) have been identified to date. Similar to Drosophila, mammalian Slit is expressed in the midline.39 Mice deficient for both Slit-1 and Slit-2 have defects in the formation of the optic chiasm.287 However, in contrast to Drosophila, these mice show normal midline guidance in the spinal cord. This different effect in midline guidance may be due to the redundancy of Slit-3 present at the midline of the spinal cord but not in the optic chiasm during development.287 Similarly, astray/robo2 mutant zebra fish show axon guidance defects in retinal axons at the midline.112 Moreover, Slit proteins act as axon guidance cues for corticofugal, callosal, thalamocortical, and olfactory neurons,13,265 and they regulate branching of axons and dendrites.386,394 Slit Structure and Proteolytic Cleavage. Members of the Slit family share a typical conserved domain structure. They contain an N-terminal signal peptide targeting them for secretion, four leucine-rich repeats mediating receptor interaction, seven (Drosophila) to nine (vertebrates) glutamic acid–glycine-phenylalanine (EGF) repeats, a lamin G domain, and a C-terminal cysteine knot.21,58,307–309 Slit proteins are subject to proteolytic cleavage (in the case of human Slit-2, between EGF repeats 5 and 6); a so far unidentified protease generates two fragments originating from the N- and C-terminus (Slit-N and Slit-C, respectively).39,386 The function of both Slit-C and the C-terminal part of uncleaved Slit is not well defined yet. Binding of Slit to Robo and Robo/Slit-mediated axon guidance depend on heparan sulfate proteoglycans,166,353 and Slit-C has higher affinity to heparan sulfate proteoglycans than Slit-N.226,305 There is also indirect evidence for a role of the C-terminal part of Slit in the regulation of Slit diffusion58; however, it is dispensable for receptor interaction and signaling.21,58,263 Uncleaved Slit and Slit-N act as a repulsive cue for axons21,58,263; however, depending on the type of neuron, there are differences in the response. For example, both

457

Slit-N and uncleaved Slit promote growth cone collapse in retinal ganglion cells, but only Slit-N leads to growth cone collapse in olfactory bulb neurons,263 and uncleaved Slit antagonizes the branch-promoting activity of Slit-N in DRG neurons.263,386 Robo/Slit Signaling Abl and Mena: Implementation of Axon Repulsion Mediated by Robo/Slit Signaling. Studies in Drosophila have shed some light on the mechanisms by which Robo/Slit signaling mediates axon repulsion. A similar mechanism likely exists in vertebrates. Both Abl (compare with discussion in Abl and Arg Kinase earlier) and Ena (see Promotion of Actin Polymerization by Mena Anti-capping Function earlier), one of its downstream targets123 involved in actin dynamics, associate with Robo (Fig. 21–6). Abl interacts with the proline-rich CC3 motif of Robo via its SH3 domain and phosphorylates Robo in the conserved CC1 motif.19 Tyrosine phosphorylation serves as a negative regulatory switch for both Ena and Robo activity. The Robo CC2 motif contains a consensus binding site for the EnaVASP homology (EVH1) domain of Ena, and Ena indeed binds to Robo, via interactions with its CC1 and CC2 motifs.19 Axon repulsion in the nematode Caenorhabditis elegans involves signaling through SLT/SAX-3, the homologues of Slit/Robo. This signaling pathway depends on the Ena homologue UNC-34,415 which modulates actin dynamics in the growth cone (compare with discussion in Promotion of Actin Polymerization by Mena Anti-capping Function, earlier, and Fig. 21–4). Abl Target Capulet and Robo/Slit Signaling. Yet another Abl target, Capulet (Capt), a homologue of yeast adenylyl cyclase–associated protein (CAP), seems to be involved in Robo/Slit signaling (see Fig. 21–6). Abl and Capt interact and function in midline axon guidance.398 CAP can interact directly with actin,111 and its function is also linked to profilin.139,383 Capt lacking the adenylyl cyclase binding domain still rescues capt mutations; however, this rescue is only partial in comparison to fulllength Capt.398 Adenylyl cyclase and modulation of intracellular cyclic AMP (cAMP) levels may play a role in Robo/Slit signaling (see Modulation of Robo/Slit Signaling by Intracellular cGMP Levels later). Protein Tyrosine Phosphatases: Role in Antagonizing Abl Tyrosine Kinase. In contrast to Ena and Capt, the targets of the tyrosine kinase Abl, it is intriguing that Ptp10D and Ptp69D, two Drosophila PTPs, are positive regulators of Robo/Slit repulsive signaling.359 Dlar (Drosophila leukocyte antigen–related-like), another Drosophila PTP, is involved in motor axon guidance in conjunction with Abl and Ena397; it is therefore likely that the role of Ptp10D and Ptp69D in Robo/Slit signaling is to antagonize Abl kinase (see Fig. 21–6).

458

Neurobiology of the Peripheral Nervous System

FIGURE 21–6 Slit/Robo signaling. The tyrosine kinase Abl and its targets are important players in mediating the repulsive effects of Slit/Robo signaling. Their function is likely to be antagonized by protein tyrosine phosphatases. The Robo receptor associates with members of a subfamily of GAPs, the Slit/Robo GAPs (srGAPs). srGAPs decrease Cdc42 activity, thus also promoting Slit-mediated axon repulsion. Robo is able to form heteromultimeric DCC/Robo receptor complexes, which interfere with netrin-mediated attraction.

Although extensively studied in Drosophila, the roles of Abl, Ena/Mena, and Capt in vertebrate Robo/Slit signaling are not yet defined. Slit/Robo GAPs: Linking Robo/Slit Signaling to Rho GTPases and Actin Dynamics. The activity of Rho GTPases involved in Robo/Slit signaling is regulated by a subfamily of Rho GAPs, the Slit/Robo GAPs (srGAPs), which bind directly to the proline-rich CC3 motif of Robo via their SH3 domain405 (see Fig. 21–6). As discussed, GAPs stimulate the conversion of active, GTP-bound Rho GTPases to their inactive, GDP-bound form (see Rho GTPases: Key Modulators of Actin Dynamics, earlier, and Fig. 21–3). srGAP1 interacts with Cdc42, RhoA, and Robo; the interaction with Robo is SH3 dependent. Stimulation with Slit increases the binding of srGAP1 to Robo and Cdc42 and subsequently decreases Cdc42 activity (see Fig. 21–6). Conversely, Slit reduces srGAP-RhoA binding, yet the effect (a slight increase of RhoA-activity) is modest compared to Cdc42. A constitutively active Cdc42 overrides the repulsive effect of Slit on migrating cells from

the anterior subventricular zone,405 further supporting a role of srGAPs and Cdc42 in Robo/Slit signaling. Involvement of the Rho GTPase Rac in Robo/Slit Signaling. Another study in Drosophila implicates Rac1 in Robo/Slit signaling.102 Mutations in Rac, the SH2-SH3 adaptor protein dreadlocks (Dock)/Nck, or the serine/threonine kinase Pak, the Cdc42/Rac downstream effector (see Fig. 21–4), interfered with Robo/Slit signaling. Stimulation with Slit recruits a complex of Dock and Pak to the Robo receptor and mediates an increase in Rac1 activity (see Fig. 21–6). Rac functions in lamellipodia formation and attractive guidance; its precise role in repulsion is not yet understood. Rac may be able to trigger both attraction and repulsion. Modulation of Robo/Slit Signaling Receptor Degradation or Proteolytic Cleavage of Robo. Robo/Slit signaling is tightly regulated both to restrain ipsilaterally projecting axons from crossing the midline and to prevent contralaterally projecting axons from recrossing.

Guidance of Axons to Targets in Development and in Disease

In Drosophila, surface levels of Robo are controlled by commissureless (Comm),369 a sorting receptor for Robo. Via its conserved sorting signal (LPSY), Comm sorts Robo directly from the trans-Golgi network into vesicles targeted to late endosomes and lysosomes. This leads to degradation of both Robo and Comm.187 This sorting mechanism also involves active Nedd4 ubiquitin ligase. Nedd4 conjugates Comm to ubiquitin, thereby marking it for protein degradation by the ubiquitin-proteasome system.258 An alternative mechanism such as direct proteolytic cleavage, as observed for ephrin/Eph signaling (see Modulation of Ephrin/Eph Signaling earlier), may also contribute to control Robo levels. Usually Robo is not present on neurons in commissural tracts; however, Drosophila commissural axons expressing a dominant negative form of the metalloprotease Kuzbanian fail to clear Robo from the cell surface and show midline crossing defects.327 Modulation of Robo/Slit Signaling by Intracellular cGMP Levels. Intracellular levels of cyclic nucleotides are known to modulate turning responses to several guidance cues 348 (e.g., semaphorin 347 or netrin 254 ) (see also Modulation of Responses to Semaphorin by Intracellular cGMP Levels, and Modulation of Response to Netrin Signaling by Intracellular cAMP Levels, later). The possible implication of capt, the homologue of yeast CAP, in Robo/Slit signaling (see Abl Target Capulet and Robo/Slit Signaling earlier) may provide a link to regulation by cyclic nucleotide levels. The level of intracellular cGMP modulates the response to Slit in DRG neurons,264 and for netrin signaling the ratio of cAMP to cGMP proved to be important in the turning response.270 Capt may associate with adenylyl cyclase to alter intracellular cAMP levels in response to Slit stimulation (see Fig. 21–6); however, an actual involvement of adenylyl cyclase in Robo/Slit signaling remains to be shown. Cross Talk between Robo/Slit and DCC/Netrin Signaling. Studies in C. elegans415 and embryonic Xenopus spinal neurons354 suggest cross talk between Robo/Slit and deleted in colorectal cancer (DCC)/netrin signaling (compare with discussion in Silencing of NetrinMediated Attraction by Formation of Robo/DCC Complexes later). In Xenopus, Robo and the netrin receptor DCC associate via interactions of their respective CC1 and P3 cytoplasmic domains to form a heteromultimeric receptor complex. In this complex, activation of the Robo receptor by Slit silences netrinmediated attraction; however, the growth-stimulatory effect of netrin is not impaired354 (see Fig. 21–6). By this means, growing axons can be attracted to the midline by netrin but are not trapped there because upregulation of Robo activity at the midline silences netrin signaling. Subsequently, Slit-mediated repulsion allows axons to leave the midline again.187,339,354

459

Unlike in Xenopus, Robo/Slit-mediated repulsion in C. elegans may not be opposing signaling by netrin receptor UNC-40/DCC. Instead a netrin-independent function of UNC-40/DCC potentiates repulsion from Slit, possibly mediated by an UNC-40/Robo heteromeric receptor complex.415

Netrin/DCC/UNC-5 The netrins have been independently discovered by biochemical purification of a diffusible outgrowth-promoting activity for axons in the spinal cord335,371 and by genetic analysis of the C. elegans mutant UNC-6 showing abnormal axonal growth toward the midline151,172 (for review, see Kennedy188 and Merz and Culotti244). The netrin family consists of phylogenetically conserved secreted proteins sharing similarity with members of the laminin family. Netrins can function in both an attracting and repelling manner; netrin-1, for example, attracts spinal cord commissural axons but repels trochlear motor axons.65,189,334 Receptor Complexes Receptors for Netrins: DCC, UNC-5, and A2b Receptor. Netrin function is mediated by two different classes of transmembrane receptors, homologues to C. elegans UNC-40 and UNC-5. Vertebrate DCC and neogenin104,186,379 are homologous to C. elegans UNC40151,152 and Drosophila frazzled.200 Vertebrate UNC-5H-1, -2, -3, and -42,97,220 are homologous to C. elegans UNC-5.222 DCC appears to be involved in attraction, whereas UNC-5 appears to mediate repulsion in both vertebrates and invertebrates.56,82,104,144,186,222 In addition to DCC and UNC5, other coreceptors may be involved in netrin signaling, for example, the adenosine A2b receptor, a G protein–coupled receptor inducing cAMP accumulation upon adenosine binding.293 Before the direct netrin-DCC interaction had been established, one study found an interaction of the A2b receptor with both netrin and DCC in a yeast two-hybrid screen and in 293T cells, suggesting that the A2b receptor was a netrin receptor that mediates cAMP elevation upon netrin binding.69 However, a conflicting report later showed the direct netrin-DCC interaction without the need for the A2b receptor in netrin-induced axon outgrowth and attraction.355 The precise role of the A2b receptor in netrin signaling therefore awaits further clarification. Structure of DCC and UNC-5. The extracellular part of DCC receptors includes several Ig domains and FNIII repeats, and three motifs (P1, P2, and P3) conserved from insects to nematodes and mammals200 characterize the intracellular moiety of DCC. Upon ligand stimulation DCC multimerizes via its P3 domains, and this multimerization is necessary for netrin-mediated attraction.355 The organization of UNC-5 diverges substantially from DCC receptors. The extracellular part contains some

460

Neurobiology of the Peripheral Nervous System

Ig and thrombospondin domains, and the cytosolic part includes several other conserved motifs. First it contains a ZU-5 domain, also present in zonula occludens-1, a component of gap junctions. It further contains a death domain, a protein-protein interaction module present in many proteins involved in apoptosis, and a DCC-binding domain necessary for interaction with DCC. Association of DCC and UNC-5 via their cytoplasmic domains converts netrin/DCC-mediated attraction to netrin/DCC/UNC5–driven repulsion in Xenopus spinal axons.162 Netrin Signaling How do netrin receptors function to regulate axon growth and guidance? Several studies suggest that they modulate the actin cytoskeleton via Rho GTPases in response to netrin. Again, although many signaling events are very similar across species, there are also distinct differences. In C. elegans, several mutations suppress a gain-of-function mutation of the netrin receptor UNC-40/DCC, including CED-10, a Rac homologue; UNC-115, a protein homologous to the vertebrate actin-binding protein AbLIM; and

UNC-34, a homologue of Mena (see Promotion of Actin Polymerization by Mena Anti-capping Function earlier), which is an important modulator of actin polymerization.129 Coupling of the Rho GTPase Rac and Its Effector Kinase PAK to Netrin Signaling by the Adaptor Protein Nck. CED-10/Rac and UNC-115/AbLIM act in one pathway dependent on the conserved P2 domain of UNC40/DCC.129 It is interesting to note that the adaptor protein Nck associates with DCC in rat commissural neurons and that Nck is required for the activation of Rac by netrin signaling224 (Fig. 21–7). In Drosophila, the Nck Drosophila homologue Dock as well as the GEF trio are necessary for photoreceptor and motor axon guidance.20,119,262 Both Rac and Cdc42 activity are required for netrin-induced neurite outgrowth in a neuroblastoma cell line, and Rac activity is increased by netrin stimulation.225 Taking into account its homology with Drosophila Dock, vertebrate Nck is likely to associate with a GEF to activate Rac upon netrin stimulation.224 Nck also binds the Rac effector kinase PAK32,117 (see Fig. 21–4), which could be

FIGURE 21–7 Netrin/DCC/UNC-5. Depending on the receptor complex, netrin can trigger both axon attraction and repulsion. Upon ligand stimulation, the DCC receptor multimerizes (for clarity, only two DCC receptors are shown) and promotes netrin-mediated axon attraction and outgrowth via modulation of actin dynamics, transcription, and local protein synthesis. In contrast, a heteromeric DCC/UNC-5 receptor complex triggers netrin-mediated repulsion. The DCC/Robo receptor complex, which interferes with netrin-mediated attraction, is shown in Figure 21–6.

Guidance of Axons to Targets in Development and in Disease

locally activated by Rac to mediate netrin-induced neurite outgrowth by activation of cofilin (see Actin Depolymerization earlier) and downregulation of retrograde F-actin flow (see Retrograde F-Actin Flow earlier). Function of UNC-34/Mena in Netrin Signaling. Caenorhabditis elegans UNC-34, the aforementioned Mena homologue (see Promotion of Actin Polymerization by Mena Anti-capping Function, earlier, and Fig. 21–4), suppresses a gain-of-function mutation of the netrin receptor UNC40/DCC, suggesting a role for UNC-34/Mena in netrin signaling.129 Being homologous to members of the Ena/VASP family, UNC-34 probably functions in netrin signaling via modulating actin polymerization (compare with discussion in Promotion of Actin Polymerization by Mena Anti-capping Function, earlier, and Fig. 21–4). UNC-34 was initially identified as a suppressor of UNC-5–mediated growth cone steering.66 UNC-34 function depends on the conserved P1 motif of UNC-40/DCC (see Fig. 21–7). The UNC-34/DCC interaction is likely to be indirect because the P1 motif of UNC-40/DCC does not contain any known Ena-binding motifs.129 Involvement of MAX-1 in a DCC-Independent Netrin Signaling Pathway. Further genetic analysis in C. elegans mutants suggests a UNC-5–dependent pathway for netrin-mediated axon repulsion independent of UNC40/DCC.169 MAX-1 (required for motor neuron axon guidance), a cytosolic protein containing two pleckstrin homology domains, a myosin tail homology 4 domain, and a Band4.1/ezrin/radixin/moesin domain, genetically interact with UNC-5 and UNC-6/netrin, but not UNC40/DCC. Because of its characteristic domain structure, MAX-1 may be membrane localized and involved in phosphatidylinositol phospholipid signaling. Such a role would be in line with the observation that phospholipase C-␥ and phosphoinositide 3-kinase are required for the netrin turning response.253 The precise role of MAX-1 in netrin signaling, however, needs to be further scrutinized. Tyrosine Phosphatase Shp-2 Binding of UNC-5. Netrin stimulation increases UNC-5 tyrosine phosphorylation, possibly via Src kinase. The PTP Shp-2 subsequently binds to the phosphorylated ZU-5 motif of UNC-5373 (see Fig. 21–7). Because Shp2 regulates RhoA activity,329 migration, and cellular adhesion,413 this interaction provides a possible explanation for netrin-controlled regulation of adhesion and actin dynamics (compare with discussion in FAK and Cas earlier). Netrin Signaling at the Transcriptional Level NFAT-Mediated Transcription. Nuclear factor of activated T-cells (NFAT) plays a role in netrin-mediated embryonic axon outgrowth.133 The serine/threonine phosphatase calcineurin, which is activated by an increase

461

of intracellular Ca 2⫹ levels, dephosphorylates the cytoplasmic NFATc subunits and promotes their import into the nucleus, where they assemble with nuclear NFAT subunits to the active transcription factor.64,106,192 Netrins, which also modulate Ca2⫹ influx in growth cones,163 stimulate this calcineurin-dependent nuclear localization of NFAT subunits, thereby activating NFAT-mediated gene transcription (see Fig. 21–7). In contrast to netrinmediated axon outgrowth, netrin-induced axon turning is NFAT independent. Promotion of Local Protein Synthesis and Degradation by Netrin Signaling. Netrin stimulates both local protein synthesis and protein degradation in growth cones, the latter via the ubiquitin-proteasome system. These events are required for both attraction and repulsion in response to netrin.48,49 In this respect, the netrin response diverges from the response to semaphorin, for which protein synthesis but not degradation is necessary to mediate its chemotropic effect48 (see Requirement of Local Protein Synthesis for Sema3A Signaling later). The MAPK pathway, which is also essential for netrin signaling107 (see Fig. 21–7) is one of the major pathways controlling protein synthesis in response to growth factors.346 Upon netrin binding, DCC recruits ERK, which mediates activation of MAPK signaling and phosphorylation of the translation initiation factor eIF-4E via MAP kinase-interacting kinase.372,388 Phosphorylation of eIF-4E, observed in growth cones upon netrin stimulation, increases eIF-4E affinity for the mRNA cap, thereby leading to stimulation of protein synthesis127 and facilitating netrin-mediated growth cone steering.48 Modulation of Netrin Signaling Modulation of Response to Netrin Signaling by Intracellular cAMP Levels. Decreased cAMP levels convert netrin-mediated attraction to repulsion in Xenopus spinal neurons, suggesting a role for cyclic nucleotide levels in the modulation of netrin signaling.254 As discussed before, the regulation of intracellular levels of cyclic nucleotides is one means to adjust the turning response to guidance cues (compare with discussion in Modulation of Robo/Slit Signaling by Intracellular cGMP Levels earlier and Modulation of Responses to Semaphorin by Intracellular cGMP Levels later). The ratio of cAMP to cGMP turned out to be important in the turning response, with high ratios favoring attraction and low ratios favoring repulsion.270 Netrin itself may influence intracellular cAMP levels because it associates with both DCC and the adenosine A2b receptor, which activates cAMP production69 (see Receptors for Netrins: DCC, UNC-5, and A2b Receptor earlier). Reflecting intracellular cAMP levels, differences in cAMP-dependent protein kinase A activity also result in a modulation of the response to netrin and other guidance cues.348

462

Neurobiology of the Peripheral Nervous System

Sequestration of Netrin by Netrin Receptors. In some contexts netrin receptors themselves may influence the response to netrin. During the guidance of Drosophila photoreceptor axons for retinal projections, the DCC orthologue frazzled is not required in growing axons.130 Instead, frazzled is necessary in the lamina target, where it captures netrin and mediates its active relocation.156 This netrin rearrangement by frazzled/DCC creates positional information and allows the presentation of netrin to other growth cones and possibly its recognition by other receptors. Silencing of Netrin-Mediated Attraction by Formation of Robo/DCC Complexes. Netrin-mediated attraction can also be silenced by other signaling pathways. As discussed for Robo/Slit signaling (see Cross Talk between Robo/Slit and DCC/Netrin Signaling, earlier, and Fig. 21–6), DCC and the Slit-receptor Robo form a heteromultimeric receptor complex. Activation of this complex by Slit silences netrin-mediated attraction without affecting the growthstimulatory effect of netrin.354 Proteolytic Cleavage of the DCC Receptor. Similar to the cleavage of ephrin-A2 in ephrin/Eph signaling (compare with discussion in Modulation of Ephrin/Eph Signaling earlier), the number of functional DCC receptors present on the cell surface is regulated by metalloproteases, which promote shedding of the extracellular domain.118 Specific metalloprotease inhibitors enhanced netrin-mediated axon outgrowth and increased the DCC protein levels in rat spinal cord explants. Proteolytic cleavage thereby provides another means to modulate the activity of netrin signaling.

Semaphorin/Plexin Similar to the netrins, the semaphorins also were discovered in a parallel search for axon guidance molecules in both vertebrates and invertebrates. The first semaphorin identified was grasshopper semaphorin 1a (fasciclin IV), shown to be expressed in stripes in the epidermis of the leg. Upon antibody treatment, pioneer sensory axons grow misdirected.199 The first vertebrate semaphorin was discovered by biochemical purification and testing of the fractions for their ability to collapse growth cones.233 Semaphorins are a large family of phylogenetically conserved axon guidance molecules. They are categorized into eight subclasses on the basis of structure, sequence similarities, and species of origin333 (reviewed by Pasterkamp and Kolodkin281 and Raper295). The semaphorin family contains both secreted and membrane-associated proteins, with subclasses 1 and 2 found in invertebrates, subclasses 3 through 7 found in vertebrates, and subclass V being of viral origin. A hallmark all semaphorins share is their conserved Sema domain at the N-terminus.

Many semaphorins are potent chemorepellents,57,233,408 but, as discussed later, there is also evidence that some semaphorins function in a chemoattractive manner.81,282,288 Receptor Complexes The semaphorin receptor complexes for the different semaphorin subclasses, or even for different members of the subclasses, vary substantially in their composition. In summary, most semaphorin holoreceptor complexes include members of the plexin or neuropilin receptor families and receptors such as Otk (off-track), L1, Met, CD72, or Tim-2 (T-cell immunoglobulin domain and mucin domain). An exception is Sema7A, which associates with integrins in vivo282 (see Promotion of Axon Outgrowth by Sema7A via Integrin Signaling later). Semaphorin Receptors: Plexins and Other Components. Plexins, components of most semaphorin receptor complexes (Fig. 21–8A, B, and D), constitute a family of transmembrane receptors with at least nine members in four subfamilies (plexin-A, -B, -C, and -D) identified so far in the mammalian genome (reviewed by Fujisawa and Kitsukawa114 and Tamagone and Comoglio366). Like semaphorins, plexins contain conserved Sema domains. A direct semaphorin-plexin interaction has been shown for invertebrate semaphorins, vertebrate membrane-associated semaphorins, and viral semaphorins.67,365,399 Plexins are also part of the receptor complex for secreted class 3 semaphorins (see Fig. 21–8A). In this complex, neuropilins are additional receptor components149,198 that mediate Sema3 binding, and plexins constitute the signaling component303 (see Fig. 21–8A). In addition, L1, a cell adhesion molecule of the Ig superfamily, represents a part of an L1/neuropilin/plexin receptor complex mediating Sema3A (but not Sema3B, 3C, or 3E) signaling54,55 (see Fig. 21–8A; also see L1, the Semaphorin Receptor, and Modulation of Semaphorin Signaling later). Like neuropilins, Otk and Met also form receptor complexes with plexins. The receptor tyrosine kinase Met (scatter factor 1/hepatocyte growth factor receptor) transduces Sema4D signaling in association with plexin-B1 (see Fig. 21–8B) and promotes invasive growth of epithelial cells.128 Drosophila transmembrane protein Otk mediates Sema1A repulsive functions required for correct projections of motor neurons to their muscle targets. Otk, which forms a receptor complex with plexin-A (see Fig. 21–8D), resembles receptor tyrosine kinases but has a catalytically inactive kinase domain.400 Other Semaphorin Receptors. Cellular processes distinct from axon guidance may require their own semaphorin receptors. In lymphoid tissues CD72, a type 2 transmembrane protein belonging to the calciumdependent C-type lectin superfamily, functions as a receptor for the class 4 semaphorin CD100 to regulate

FIGURE 21–8 Semaphorin/plexin signaling. A, Secreted class 3 semaphorins are potent chemorepellents. Sema3A dimerizes in order to induce growth cone repulsion or collapse.193 Neuropilin, A-type plexins, and the cell adhesion molecule L1 form the receptor complex for Sema3A. This complex recruits a broad variety of downstream effectors to modulate actin and tubulin dynamics. B, Class 4 semaphorins mediate growth cone turning and repulsion by sequestration of Rac and activation of RhoA. C, GPI-anchored class 7 semaphorins function via their arginine-glycine–aspartic acid (RGD)-mediated interaction with ␤-integrins. Activation of focal adhesion kinase (FAK) and mitogen-activated protein kinase (MAPK) promotes axonal outgrowth. D, In Drosophila, the association of class 1 semaphorins with MICAL (molecule associated with CasL) causes axon repulsion. MICAL is a putative flavoprotein monooxygenase, pointing to a possible role of redox signaling in semaphorin-mediated axon repulsion. A similar mechanism may exist in vertebrates as well.

464

Neurobiology of the Peripheral Nervous System

B-cell responsiveness.213 Also implicated in Sema4A signaling, Tim-2, a member of the family of T-cell immunoglobulin domain and mucin domain (Tim) proteins, enhances the generation of antigen-specific T-cells in response to Sema4D/CD100.212 Semaphorin Signaling Rho GTPases and Semaphorin Signaling. Like other axon guidance signaling pathways, semaphorin signaling exerts its effects on axon guidance via modulation of cytoskeletal dynamics in the growth cone upon receptor activation.100,101,113 There is evidence for direct and indirect interactions between plexins and Rho GTPases, the key regulators of the actin cytoskeleton. The cytoplasmic domain of plexins shares significant sequence similarity with the GAP domain of Ras GAPs.304 GAP activity of plexins might allow a direct regulation of Rho GTPases (compare with discussion in Rho GTPases: Key Modulators of Actin Dynamics, earlier, and Fig. 21–3); however, such activity and putative targets remain to be shown. In vitro studies with chick DRG neurons implicated Rac1 in semaphorin signaling. DRG neurons transfected with dominant negative or constitutively active Rac1 decreased or enhanced Sema3A-induced growth cone collapse, respectively.177 Rac is implicated in growth cone advance (see Actin Depolymerization, and Retrograde F-Actin Flow, earlier, and Fig. 21–4), so its role in Sema3A-mediated growth cone collapse is not self-evident. There is also direct and specific interaction between GTP-bound active Rac and plexinB1.90,381 Plexin-B1 clustering in fibroblasts does not induce lamellipodia formation as one would expect in the case when Rac is activated. Instead, actin-myosin fibers are formed and the cell retracts, suggesting that Rho might be activated.90 Plexin-B1 sequesters active Rac from its downstream target PAK, thereby inhibiting Rac-induced PAK activation380 (see Fig. 21–8B). Rac mediates the inactivation of the actindepolymerizing factor cofilin, so if Rac is sequestered, cofilin remains active, which promotes semaphorin-induced growth cone collapse (see Actin Depolymerization, earlier, and Fig. 21-4). In Drosophila, a direct interaction between plexin-B and RhoA was found that leads to RhoA activation.167 In vertebrates, the plexin-B–RhoA interaction is indirect: Two Rho-specific GEFs, PDZ-Rho-GEF and LARG, interact with the C-terminal PDZ recognition sequence of plexin-B via their PDZ domains. These GEFs activate RhoA, which in turn mediates growth cone collapse10,62,157,286,363 (see Fig. 21–4). Unlike plexin-B, other plexin subfamilies do not contain the CRIB motif or the C-terminal PDZ domain, which suggests a different mode of regulation of cytoskeletal dynamics. Rnd1 and RhoD, members of the Rho GTPase family implicated in spine formation173 and motility of early endo-

somes,121 respectively, bind the cytoplasmic domain of plexin-A1 and compete for the same binding site416 (see Fig. 21–8A). Rnd1 increases the GAP activity of a Rhospecific GAP, p190GAP,392 so the antagonism of Rnd1 and RhoD binding might regulate plexin-A1 activity and thereby Sema3A signaling. However, another group arrived at a different result, showing that Rnd1 interacts directly with plexin-B1 but not plexin-A1, suggesting that Rnd1 might activate RhoA via PDZ-RhoGEF.274 Kinases Downstream of Semaphorin Signaling. Fes tyrosine kinase binds to the intracellular domain of plexinA1 and recruits the collapsin response mediator protein (CRMP)/CRMP-associated molecule (CRAM) complex132,255 (see Fig. 21–8A). In the absence of Sema3A, neuropilin-1 (Npn-1) blocks the association of plexin-A and Fes. Upon stimulation with Sema3A, which uses Npn-1 as receptor, Fes binds plexin-A and phosphorylates plexin-A1, CRMP, and CRAM.255 CRMP binds tubulin heterodimers directly.116 CRMP is also a target of ROCK8 and alters the responses to Rho and Rac signaling.142 Taken together, these results implicate CRMP/CRAM in the regulation of actin and microtubule dynamics in semaphorin signaling. Sema3A-induced growth cone collapse can be prevented by kinase inhibitors specific for tyrosine kinases or cyclindependent kinases; it was therefore suggested that plexin-A signaling depends on both the Src family tyrosine kinase Fyn and the serine/threonine kinase cyclin-dependent kinase 5 (Cdk5).323 Fyn kinase associates with plexin-A1 and -A2 and phosphorylates their cytoplasmic domains (see Fig. 21–8A). Additionally, plexin-A2 associates with Cdk5 via an interaction with active Fyn, in which Fyn phosphorylates and thereby activates Cdk5. The microtubule-associated protein (MAP) tau is one of the targets of Cdk5. MAPs are the phosphoregulated key modulators of microtubule stability and dynamics (for review, see Drewes and colleagues89), so tau phosphorylation by Cdk5 upon Sema3A-stimulation might provide a link to microtubule dynamics.59 Cdk5 and its neuron-specific regulator p35 associate with GTP-bound Rac and its effector kinase PAK (see Fig. 21–8A). In this complex, hyperphosphorylation of PAK by Cdk5/p35 downregulates PAK activity,267 thereby modulating dynamics of the actin cytoskeleton to allow repulsion and growth cone collapse (compare Fig. 21–4). An inactive, phosphorylated pool of glycogen synthase kinase-3 (GSK-3), another kinase able to phosphorylate tau,145 co-localizes with F-actin in filopodia and at the leading edge of lamellipodia in growth cones of DRG neurons. Stimulation with Sema3A activates this GSK-3 pool at the leading edge of neuronal growth cones93 (see Fig. 21–8A). There is also evidence for cross talk between GSK-3 and Rho signaling. G␣13, an ␣-subunit of heterotrimeric G proteins, interacts with the Rho-specific GEF p115RhoGEF,

465

Guidance of Axons to Targets in Development and in Disease

and GSK-3 activation by G␣13 is Rho mediated.201,325 However, further studies will be necessary to elucidate the involvement of GSK-3 in Sema3A signaling.

Sema3A as a result of asymmetrical localization of soluble guanylate cyclase,288 so cGMP levels may be an endogenous regulator of Sema3A signaling.

Promotion of Axon Outgrowth by Sema7A via Integrin Signaling. In contrast to many other semaphorins, Sema7A promotes axon outgrowth. In vitro studies suggested plexin-C1 to be the binding partner of Sema7A.365 The growth-promoting effect of Sema7A, however, is independent of plexin-C1.282 In vivo, Sema7A associates with integrins by virtue of its arginine-glycine– aspartic acid motif. Integrins activate FAK and subsequently the MAPKs ERK 1 and 2, thereby mediating the growth-promoting effect of Sema7A signaling in neurons282 (see Fig. 21–8C).

Requirement of Local Protein Synthesis for Sema3A Signaling. Inhibition of protein synthesis prevented Sema3A-induced growth cone turning or collapse. Upon Sema3A stimulation, local protein synthesis in growth cones was rapidly increased by inactivation of the translation repressor eIF-4EBP1 and subsequent activation of the translation initiation factor eIF-4E. Although not shown directly, local protein synthesis may alter the receptor composition on the growth cone surface, a mechanism similarly found in ephrin signaling (see Modulation of Ephrin/Eph Signaling earlier). In contrast, inhibition of the ubiquitin-proteasome system did not alter the chemotropic response to Sema3A.48 Therefore, local protein synthesis but not degradation seems to be required for Sema3A signaling.

A Putative Role for Redox Signaling in SemaphorinMediated Axon Repulsion. Drosophila MICAL (molecule associated with CasL), a multidomain cytosolic protein expressed in axons that also contains a flavoprotein monooxygenase domain, was found to interact with plexinA.370 MICAL turned out to be necessary for repulsive axon guidance of motor neurons mediated by Sema1A/plexin-A (see Fig. 21–8D). Vertebrate Sema6A is homologous to insect class 1 semaphorins.419 Vertebrate MICALs show specific neuronal and non-neuronal expression patterns during development, and human MICAL-1 and mouse MICAL-2 seem to specifically interact with human plexinA3 and mouse plexin-A4, respectively.370 In cultured DRG neurons, inhibition of 12/15-lipoxygenase blocks Sema3Ainduced growth cone collapse, which can be induced by exogenously applied product of 12/15-lipoxygenase.250 These results point to an involvement of redox signaling in semaphorin-mediated axon repulsion; however, its exact role remains to be determined. MICAL interactions may provide a link to cytoskeletal dynamics because MICAL associates with intermediate filaments and the adaptor protein CasL (compare with discussion in FAK and Cas earlier), a member of the Cas family, which is implicated in actin filament assembly and integrin-dependent processes such as migration, cell-cycle control, and transformation.161,277,324,362 Modulation of Semaphorin Signaling Modulation of Responses to Semaphorin by Intracellular cGMP Levels. Differences in intracellular cAMP levels are known to result in opposite turning of growth cones in response to the same guidance cue.348 As in netrin and Slit signaling (see Modulation of Response to Netrin Signaling by Intracellular cAMP Levels, and Modulation of Robo/Slit Signaling by Intracellular cGMP Levels earlier), the response to semaphorin is also subject to modulation by cyclic nucleotides. The collapse of cultured rat sensory growth cones upon Sema3A stimulation can be inhibited by activation of the cGMP pathway.347 Axons and dendrites differently respond to

AXON GUIDANCE IN DISEASE Surprisingly, the rapid progress in identifying and characterizing the large number of key molecules involved in axon guidance during development has revealed the involvement of these same molecules in only a small number of diseases. However, considering the complex process of brain development, one explanation may be the devastating consequences of mutations in main guidance systems on neural development. A fetus carrying a severe mutation is not likely to reach peri- or postnatal stages. Another explanation may be a certain functional redundancy between guidance molecules, which may lead to compensation for less severe mutations. The potential role of axon guidance molecules as inhibitors of the axonal regrowth after CNS traumatic lesions is discussed in recent reviews.83,197 We concentrate here on the role of axon guidance molecules in diseases, specifically epilepsy, schizophrenia, CRASH syndrome, and Kallmann’s syndrome.

Epilepsy Background Epilepsy is a neurologic disorder characterized by a tendency to recurrent seizures that can cause temporal impairment or loss of consciousness, abnormal motor phenomena such as convulsions, psychic or sensory disturbances, and perturbation of the autonomic nervous system (reviewed by Riviello301 and Shneker and Fountain340). Approximately 1% to 3% of the population worldwide is affected by epilepsy. Seizures are the result of excessive electrical discharges in the brain, which usually occur briefly and suddenly. If a seizure arises from one or more localized areas of the brain

466

Neurobiology of the Peripheral Nervous System

and does not cause the loss of consciousness, it is called a partial or focal seizure. A generalized seizure, in contrast, brings about a loss of consciousness and involves the entire brain. A state referred to as “status epilepticus” is characterized by frequent seizures without recovery of consciousness between the episodes. Mossy Fiber Sprouting One of the pathophysiologic hallmarks of temporal lobe epilepsy is the hyperexcitability of the hippocampal formation, which may be caused by a phenomenon called mossy fiber sprouting (MFS).11,343 Mossy fibers, the axons of dentate granule cells, normally innervate CA3 pyramidal cells and hilar cells. Upon loss of their proper targets, however, mossy fibers start to sprout into the inner molecular layer of the dentate gyrus, where they form functional synapses with dendrites of granule cells in animal models.275,368,393 This aberrant innervation may contribute to the formation of recurrent excitatory circuits, which may in turn facilitate the spontaneous recurrence of partial seizures and propagate the epileptic state12 (for review, see Nadler259). Loss of hippocampal tissue may be due to an initial precipitating injury, which may cause hippocampal sclerosis and trigger MFS.241 Glycosaminoglycans, Proteoglycans, and N-Cadherin Axon targeting during formation of the hippocampus is regulated by multiple guidance signals. In the adult brain, inhibitory guidance cues stabilize the complex, balanced neuronal network, which is very sensitive to further axon outgrowth and wiring. The levels of several guidance molecules are altered and may thereby facilitate MFS in the epileptic brain. Glycosaminoglycans (e.g., hyaluronic acid and chondroitin sulfate) and proteoglycans (proteins covalently linked to glycosaminoglycans) are elements of the extracellular matrix. They function as coreceptors for growth factors in the nervous system and regulate a broad variety of physiologic processes, including cell migration, neurite outgrowth, axonal pathfinding, synaptogenesis, and structural plasticity (for review, see Bandtlow and Zimmerman16 and Bovolenta and Fernaud-Espinosa34). Alterations in the levels of the glycosaminoglycans chondroitin sulfate and hyaluronic acid as well as receptor PTP ␨/␤, a chondroitin sulfate proteoglycan implicated in axonal sprouting, suggest a perturbance of guidance signals in the epileptic brain.260,285,345 Similarly, N-cadherin is upregulated in the inner molecular layer of the dentate gyrus, corresponding to MFS in this hippocampal region.337 N-cadherin is involved in synapse formation during development103,375; in line with this, N-cadherin localizes to newly formed synapses between mossy fibers and granule cell dendrites.337 Semaphorin Signaling Semaphorin signaling also seems to play a role in MFS. Mossy fibers express plexin-A and neuropilin,257 which

form the receptor complex for the chemorepellent Sema3A (see Semaphorin Receptors: Plexins and Other Components, and Fig. 21–8A earlier). Sema3A, which is secreted by cells of the entorhinal cortex,126 diffuses into the adjacent dentate gyrus, thereby repelling mossy fibers harboring the Sema3A receptor. In a rat model, Sema3A mRNA was transiently downregulated in the entorhinal cortex after induction of status epilepticus by electrical stimulation.160 This study could not show a reduction of Sema3A at the protein level; therefore, further inquiries are necessary to support the relevance of this finding. A possible cutback of Sema3A-mediated repulsion (see Semaphorin/Plexin earlier) plausibly explains sprouting of mossy fibers into the inner molecular layer of the dentate gyrus, which no longer represents a repelling environment. Further support for a role of semaphorin signaling in epilepsy comes from the phenotype of plexin-A3 knockout mice, which are susceptible to kainate-induced seizures (H. J. Cheng, M. Tessier-Lavigne, and S. Pleasure, personal communication, 2004). Plexin-A3 is the receptor for Sema3A; therefore, impaired Sema3A signaling (as during Sema3A downregulation) may contribute to MFS and the formation of additional neuronal circuits, which would account for the hyperexcitability of the hippocampal formation in patients with epilepsy. Summary Disturbances of the levels of guidance cues contribute to aberrant axon sprouting and synapse formation in the epileptic brain. This may facilitate the formation of recurrent excitatory circuits that in turn promote seizures and propagate the epileptic state.

Schizophrenia Background Schizophrenia is a severe mental disorder that is generally characterized by fundamental and characteristic distortions of thinking and perception, such as delusions and hallucinations, and by inappropriate or blunted affect.406 The lifetime prevalence of schizophrenia worldwide has been estimated to be about 1%, with the average onset occurring from adolescence through early adulthood. Several anatomic abnormalities, including ventricular enlargement and decreased cerebral (cortical and hippocampal) volume, are hallmarks of schizophrenia.146 The causes of schizophrenia are still largely obscure; however, it is now accepted that genetic factors play a role in forming a predisposition to developing the illness,50 and external factors influence the timing of its possible onset. Multiple neuropathologic studies suggest an abnormal early neurodevelopment of patients with schizophrenia, especially a disturbance in neuronal migration, synaptogenesis, and process formation.9,68,146,384

Guidance of Axons to Targets in Development and in Disease

Wnt Signaling The Wnt signaling pathway is crucial for a large number of developmental processes, including embryonic patterning, cell fate determination, and the regulation of cell proliferation, polarity, and lymphocyte development.41,376,402 Additionally, Wnt signaling is implicated in anterior-posterior guidance of commissural axons234 and organization of the patterning during cortical development.136 Frizzled proteins constitute a large family of seven-pass transmembrane proteins, some of which serve as receptors for Wnt ligands, a family of secreted glycoproteins.25,412 One member of the frizzled family, frizzled-3 (FZD3), functions in development of major fiber tracts in the rostral CNS387 and formation of the neural crest in Xenopus.84 Several studies point to an association of the FZD3 locus and the susceptibility to schizophrenia.184,410 The molecular basis for this correlation is not yet elucidated; however, mutations in FZD3 might contribute to defects in axon guidance and neural connection, thus elevating susceptibility to schizophrenia. DISC-1/NUDEL/LIS-1 Genetic studies using family analysis were able to link the locus for DISC-1 (disrupted in schizophrenia-1) to susceptibility to schizophrenia.30,251 The mutant form of DISC-1 occurring in these families is C-terminally truncated and fails to bind NudE-like (NUDEL),278 a DISC-1 interaction partner involved in nucleokinesis and neuronal migration.147,215 NUDEL also forms a complex with LIS-1,266,322 a factor implicated in lissencephaly. Lissencephaly is a human brain malformation exhibiting a smooth cerebral cortex and abnormal neuronal migration.296 The NUDEL/LIS-1 complex participates in organizing the microtubule cytoskeleton during nuclear translocation and neurite outgrowth.306 In line with these results, another study showed a direct implication of DISC-1 in neurite extension.256 Upon overexpression of DISC-1, PC12 cells show longer neurites and DISC-1 localizes in growth cones. In rats, DISC-1 is highly expressed in the cerebral cortex and the hippocampus in late development278; these are two regions shown to be affected in schizophrenia. Mutations in DISC-1 may prevent an adequate response of migrating neurons or growing axons to guidance cues in development, thereby raising the susceptibility to schizophrenia. Semaphorin Signaling Sema3A, a potent chemorepellant in axon guidance (see Semaphorin/Plexin earlier), was recently shown to be upregulated in the cerebellum of adult schizophrenia patients.91 Along with the Sema3A upregulation, a decrease of reelin expression was found,91 for both cerebellum and prefrontal cortex.137 The reelin signaling pathway is involved in neuronal migration, and mutations in reelin cause an autosomal recessive form of lissencephaly.164 Moreover, reelin mutant mice (reeler mice) show

467

decreased axon branching of entorhinal axons targeting the hippocampus, altered hippocampal synaptogenesis, and misrouted fibers.33 Alteration in Sema3A and reelin levels in patients with schizophrenia might impair normal connectivity and synaptogenesis during development, thus contributing to predisposition for schizophrenia. NCAM and L1: Immunoglobulin Superfamily Members Molecules of the Ig superfamily such as neural cell adhesion molecule (NCAM) and L1 are involved in axon guidance, synapse stability, cell migration, and maintenance and function of neuronal networks.310,342 In the cerebrospinal fluid of patients with schizophrenia, levels of L1 and a cleaved isoform of NCAM are decreased and elevated, respectively. Additionally, the cleaved form of NCAM is increased in the hippocampus and the prefrontal cortex.289,377 Augmented cleavage of Ig proteins may be due to abnormal proteolysis in schizophrenia, either by enhanced proteolytic activity or decreased resistance to cleavage. Removal of polysialic acid from NCAM by neuraminidase may account for decreased protease resistance,377 and polysialylated NCAM is decreased in hippocampi of patients with schizophrenia.17 Genetic and twin studies indicate that changes of Ig protein levels are not likely to be caused by a genetic predisposition resulting from structural alterations of the genes,290,378 but might rather be a consequence of schizophrenia onset. The relevance of altered Ig protein levels for onset and progression of schizophrenia therefore awaits further clarification.

CRASH Syndrome/MASA Syndrome/L1 Syndrome Background The NCAM L1 (compare with discussion in Semaphorin Receptors: Plexins and Other Components, earlier, and Fig. 21–8A), a single-pass transmembrane glycoprotein of the Ig superfamily, is involved in the regulation of neurite growth, axon guidance, cell migration, the activation of second messenger signaling cascades, and possibly also cytoskeletal rearrangements.45,47,77,181 Mutations in the X-chromosomal gene for L1 cause several X-linked mental retardation syndromes, which were first considered different diseases, including X-linked hydrocephalus (hydrocephalus resulting from stenosis of the aqueduct of Sylvius), MASA syndrome26 (an acronym for mental retardation, aphasia, shuffling gait, and adducted thumbs), X-linked spastic paraplegia/paraparesis, and X-linked agenesis of the corpus callosum. The clinical syndrome of L1 mutation is now sometimes also referred to as CRASH syndrome, owing to its main clinical features (corpus callosum agenesis, mental retardation, adducted thumbs, spastic paraplegia, and hydrocephalus).110 Given the widespread expression of L1,

468

Neurobiology of the Peripheral Nervous System

alterations in brain morphology of patients with CRASH syndrome are surprisingly modest. The most prominent consequences of mutations in L1 are reduction or lack of two long axonal tracts, the aforementioned corpus callosum, which connects the cerebral hemispheres, and the corticospinal tract, which is involved in controlling voluntary motor functions. Reduction of the corpus callosum often coincides with mental retardation, and malformation in the corticospinal tract can account for the spasticity syndromes. A large number of mutations giving rise to defects in L1 function has been described by now, causing a broad range of clinical symptoms (for a list of L1 mutations and further links, see www.uia.ac.be/dnalab/l1/). These L1 mutations can be classified into three categories: class 1 includes mutations in the cytoplasmic domain, class 2 exhibits point mutations in the extracellular domain, and class 3 is characterized by a premature stop codon in the extracellular domain. Because of the many physiologic processes L1 is involved in, severity of the symptoms can vary widely and clearly correlates with the site of mutation, depending on the function influenced. Patients with class 1 mutations are least affected, and class 3 mutations cause the most severe symptoms.409 Fibroblast Growth Factor Receptor and Src Kinase L1 itself does not possess any catalytic domains; however, it engages several binding partners to activate signaling pathways involved in neurite growth and guidance. Surfacebound L1 constitutes a substrate that promotes neurite outgrowth in vitro. This provides an assay to test L1-function. Neurons from L1 knockout mice lose the ability to extend neurites on L1 substrate, suggesting that a homophilic interaction between cellular and extracellular L1 is necessary to promote neurite outgrowth.75,219 This effect is not simply due to cellular adhesion, because a soluble form of L1 can also trigger neurite extension.87 A putative receptor for L1 is fibroblast growth factor receptor (FGFR). FGFR signaling not only promotes survival and proliferation but also is involved in guiding axons of retinal ganglion cells to the optic fissure.38 Activation of FGFR is necessary for L1-mediated axon outgrowth,396 and antibody-mediated inactivation of FGFR396 or expression of dominant negative FGFR314 abolishes neurite extension triggered by L1. The tyrosine kinase Src may be part of this signaling pathway, because neurons deficient for Src show a diminished rate of neurite extension on L1 substrate.171 Axonin-1 and Associated Kinases Heterophilic association of L1 with axon-associated cell adhesion molecule axonin-1 effectively promotes neurite outgrowth of embryonic DRG neurons.207,356 Axonin-1 interacts with the tyrosine kinase Fyn,214 which provides a link for L1 signaling to cellular adhesion and cytoskeleton dynamics (compare with discussion in Kinases

Downstream of Semaphorin Signaling, and GPI-Linked A-Type Ephrins: Capacity for Reverse Signaling earlier). Serine/threonine casein kinase II,403 serine/threonine ribosomal S6 kinase (rsk),404 and the Eph receptor tyrosine kinase homologue chicken embryo kinase 5421 phosphorylate the cytoplasmic domain of L1 and are therefore likely to modulate L1 function. Indeed, cell-cell contacts induce L1–axonin-1 clustering, which modulates the activity of axonin-associated Fyn kinase and L1-associated casein kinase II- and rsk-related activity.214 Cytoskeleton Association via Ankyrin The highly conserved cytoplasmic domain of L1 binds to ankyrin, thereby associating with the spectrin-based membrane cytoskeleton.77 Tyrosine phosphorylation of the intracellular L1 domain abolishes ankyrin binding and increases the lateral mobility of L1. Phosphorylation by Src, Fyn, or an Eph receptor kinase might therefore regulate adhesive properties of L1-expressing cells during neuronal migration or axon outgrowth.120 L1, the Semaphorin Receptor, and Modulation of Semaphorin Signaling Cortical axons from L1-deficient mice fail to respond to a repulsive signal from the spinal cord, which could be identified as the secreted chemorepellent Sema3A. L1 and Npn-1 form a Sema3A-binding complex via direct interaction of their extracellular domains.54 Together with plexinA, this complex constitutes the receptor that triggers Sema3A signaling364,365 (see Semaphorin Receptors: Plexins and Other Components, earlier, and Fig. 21–8A), implicating L1 in semaphorin-mediated long-distance axon guidance. A soluble form of L1 interacting with Npn-1 converts Sema3A-induced chemorepulsion to attraction by activation of neuronal nitric oxide (NO) synthase. Subsequent NO synthesis stimulates cGMP guanylyl cyclase, which increases internal cGMP levels, thereby mediating the switch from repulsion to attraction54 (compare with discussion in Modulation of Robo/Slit Signaling by Intracellular cGMP Levels, Modulation of Response to Netrin Signaling by Intracellular cAMP Levels, and Modulation of Responses to Semaphorin by Intracellular cGMP Levels earlier). In vivo, L1 is subject to proteolytic cleavage by the serine protease plasmin261 and the metalloprotease ADAM 10,138 a vertebrate orthologue to Drosophila Kuzbanian. Proteolytic cleavage might modulate the response to Sema3A by producing a soluble L1 cleavage product or by disrupting the L1/Npn-1/plexin-A receptor complex.

Kallmann’s Syndrome Background The main characteristics of the congenital Kallmann’s syndrome (KS) are hypogonadotropic hypogonadism, that is, defective gonadal development or function resulting from

Guidance of Axons to Targets in Development and in Disease

gonadotropic hormone deficiency, and anosmia or hyposmia (the absence or reduction of the sense of smell). These symptoms were already described in 1856 by Maestre de San Juan,238 and further characterized as a clinical entity by Kallmann in 1944.179 Different forms of KS are inherited in an X-chromosomal or autosomal manner, with varying associated anomalies, including upper body mirror movements, renal agenesis, and midline/craniofacial abnormalities such as harelip or cleft palate. The prevalence of KS is about five times higher in males (~1:10,000) than in females. The anosmia observed in KS is due to reduction or absence of the olfactory bulb. Early innervation of the olfactory bulb primordium by axons from the olfactory epithelium is necessary for induction of bulb development. Because migration and targeting of axons from the olfactory epithelium is impaired in KS, the olfactory bulb does not form properly.131 Hypogonadotropic hypogonadism in KS is secondary to a lack of gonadotropin-releasing hormone (GnRH). Neurons secreting GnRH use axons of olfactory receptor neurons as guiding tracks on their way from the olfactory placode to the brain.331 Deprived of their guides, GnRHsynthesizing neurons fail to reach their targets, which results in hypogonadism in KS.330 Anosmin-1, Gene Product of Kal-1, and Its Dual Branch-Promoting and Guidance Activity Kal-1 was the first gene to be discovered that is responsible for the X-linked form of KS.109,217 Kal-1 codes for anosmin-1, an extracellular matrix protein containing four FNIII domains and a whey acidic protein domain, often found in proteinase inhibitors. Anosmin-1, which is secreted by the olfactory cortex, has a dual branchpromoting and guidance activity. First, it promotes axonal branching of cultured rat olfactory bulb neurons, resulting in collateral branches.349 Second, it functions as an axonal guidance cue, which allows the olfactory cortex to attract these collaterals. Anosmin-1, Heparan Sulfate Proteoglycans, and FGFR Activity of anosmin-1 is regulated in embryonic development, and seems to depend on an additional factor.349 Heparan sulfate proteoglycans (see Glycosaminoglycans, Proteoglycans, and N-Cadherin earlier) are good cofactor candidates because they regulate processes such as cell migration or axonal pathfinding16,34 and are likely to tether anosmin-1 to the cellular membrane.350 In line with this, mutations in heparan-6-O-sulfotransferase were able to suppress the axon-branching and misrouting phenotype of Kal-1 mutations in C. elegans.46 Anosmin-1 functions in C. elegans and vertebrates are comparable because they show high functional conservation: Similarly to Kal-1 in vertebrates, C. elegans Kal-1 seems to control axon branching and guidance, because overexpression of Kal-1 in C. elegans

469

promotes enhanced axon branching and faulty axon guidance.46 The high conservation is also stressed by the fact that human Kal-1 can rescue C. elegans Kal-1 mutations.312 In addition to their interaction with anosmin-1, heparan sulfate proteoglycans likely promote the dimerization of the binary complex of FGF and its receptor (FGFR)284 (also see Fibroblast Growth Factor Receptor and Src Kinase earlier). Interestingly, an autosomal dominant variant of KS is caused by loss-of-function mutations in FGFR,86 and FGFR is required for the formation of the olfactory bulb.150 These observations suggest a functional link between anosmin-1 and FGFR, because mutations in both impair olfactory bulb and gonad development. Anosmin-1 may be implicated in FGF signaling, resulting in the symptoms of KS.

OUTLOOK Although recent years have seen the rapid identification and characterization of the four major axon guidance systems, ephrin/Eph, Slit/Robo, Netrin/DCC/UNC-5, and semaphorin/plexin, we are still far from understanding how the brain is wired. First, although the identified axon guidance ligands and receptors give basic insights into how axons find their appropriate target, they only account for a minute fraction of guidance events that occur within the nervous system. Thus other axon guidance cues wait to be discovered. Genetic approaches in which a gene is silenced or mutated together with expression of a reporter within the affected neurons should lead to such discoveries.218 Second, although numerous signaling components downstream of the receptors are now known, our knowledge is still fragmentary. Improved detection of protein-protein interaction and mass spectroscopic analysis of whole signaling complexes122 will help to reveal the mechanisms underlying guidance phenomena. Finally, better understanding of the axon guidance cues and signaling will allow us to correlate more diseases and defective axon guidance molecules, which may in some cases lead to improved treatment and perspectives for affected patients.

ACKNOWLEDGMENTS We are very grateful to Christel Bauereiss and Robert Schorner for help with the figure preparation. We would like to thank Drs. Boyan Garvalov, Rüdiger Klein, Edgar Meinl, Brian Storrie, and Gaia Tavosanis for reading the manuscript. We are also indebted to Drs. Hwai-Jong Cheng, Samuel J. Pleasure, and Marc Tessier-Lavigne for sharing unpublished results with us. We are very grateful to Dr. Paul Forscher (Yale University) for granting permission to reproduce Figure 21–2. Frank Bradke is a recipient of a Career Development Award from the Human Frontier Science Program.

470

Neurobiology of the Peripheral Nervous System

REFERENCES 1. Abdul-Manan, N., Aghazadeh, B., Liu, G. A., et al.: Structure of Cdc42 in complex with the GTPase-binding domain of the ‘Wiskott-Aldrich syndrome’ protein. Nature 399:379, 1999. 2. Ackerman, S. L., Kozak, L. P., Przyborski, S. A., et al.: The mouse rostral cerebellar malformation gene encodes an UNC-5-like protein. Nature 386:838, 1997. 3. Aizawa, H., Wakatsuki, S., Ishii, A., et al.: Phosphorylation of cofilin by LIM-kinase is necessary for semaphorin 3A-induced growth cone collapse. Nat. Neurosci. 4:367, 2001. 4. Amano, M., Ito, M., Kimura, K., et al.: Phosphorylation and activation of myosin by Rho-associated kinase (Rho-kinase). J. Biol. Chem. 271:20246, 1996. 5. Anderson, R. G.: Plasmalemmal caveolae and GPI-anchored membrane proteins. Curr. Opin. Cell Biol. 5:647, 1993. 6. Arakawa, Y., Bito, H., Furuyashiki, T., et al.: Control of axon elongation via an SDF-1alpha/Rho/mDia pathway in cultured cerebellar granule neurons. J. Cell Biol. 161:381, 2003. 7. Arber, S., Barbayannis, F. A., Hanser, H., et al.: Regulation of actin dynamics through phosphorylation of cofilin by LIM-kinase. Nature 393:805, 1998. 8. Arimura, N., Inagaki, N., Chihara, K., et al.: Phosphorylation of collapsin response mediator protein-2 by Rho-kinase: evidence for two separate signaling pathways for growth cone collapse. J. Biol. Chem. 275:23973, 2000. 9. Arnold, S. E.: Neurodevelopmental abnormalities in schizophrenia: insights from neuropathology. Dev. Psychopathol. 11:439, 1999. 10. Aurandt, J., Vikis, H. G., Gutkind, J. S., et al.: The semaphorin receptor plexin-B1 signals through a direct interaction with the Rho-specific nucleotide exchange factor, LARG. Proc. Natl. Acad. Sci. U. S. A. 99:12085, 2002. 11. Babb, T. L., Lieb, J. P., Brown, W. J., et al.: Distribution of pyramidal cell density and hyperexcitability in the epileptic human hippocampal formation. Epilepsia 25:721, 1984. 12. Babb, T. L., Pretorius, J. K., Mello, L. E., et al.: Synaptic reorganizations in epileptic human and rat kainate hippocampus may contribute to feedback and feedforward excitation. Epilepsy Res. Suppl. 9:193, 1992. 13. Bagri, A., Marin, O., Plump, A. S., et al.: Slit proteins prevent midline crossing and determine the dorsoventral position of major axonal pathways in the mammalian forebrain. Neuron 33:233, 2002. 14. Bamburg, J. R., and Bray, D.: Distribution and cellular localization of actin depolymerizing factor. J. Cell Biol. 105:2817, 1987. 15. Bamburg, J. R., McGough, A., and Ono, S.: Putting a new twist on actin: ADF/cofilins modulate actin dynamics. Trends Cell Biol. 9:364, 1999. 16. Bandtlow, C. E., and Zimmermann, D. R.: Proteoglycans in the developing brain: new conceptual insights for old proteins. Physiol. Rev. 80:1267, 2000. 17. Barbeau, D., Liang, J. J., Robitalille, Y., et al.: Decreased expression of the embryonic form of the neural cell adhesion molecule in schizophrenic brains. Proc. Natl. Acad. Sci. U. S. A. 92:2785, 1995. 18. Bashaw, G. J., and Goodman, C. S.: Chimeric axon guidance receptors: the cytoplasmic domains of slit and netrin receptors specify attraction versus repulsion. Cell 97:917, 1999.

19. Bashaw, G. J., Kidd, T., Murray, D., et al.: Repulsive axon guidance: Abelson and Enabled play opposing roles downstream of the roundabout receptor. Cell 101:703, 2000. 20. Bateman, J., Shu, H., and Van Vactor, D.: The guanine nucleotide exchange factor trio mediates axonal development in the Drosophila embryo. Neuron 26:93, 2000. 21. Battye, R., Stevens, A., Perry, R. L., and Jacobs, J. R.: Repellent signaling by Slit requires the leucine-rich repeats. J. Neurosci. 21:4290, 2001. 22. Bear, J. E., Svitkina, T. M., Krause, M., et al.: Antagonism between Ena/VASP proteins and actin filament capping regulates fibroblast motility. Cell 109:509, 2002. 23. Behlke, J., Labudde, D., and Ristau, O.: Self-association studies on the EphB2 receptor SAM domain using analytical ultracentrifugation. Eur. Biophys. J. 30:411, 2001. 24. Bentley, D., and Toroian-Raymond, A.: Disoriented pathfinding by pioneer neurone growth cones deprived of filopodia by cytochalasin treatment. Nature 323:712, 1986. 25. Bhanot, P., Brink, M., Samos, C. H., et al.: A new member of the frizzled family from Drosophila functions as a Wingless receptor. Nature 382:225, 1996. 26. Bianchine, J. W., and Lewis, R. C. Jr.: The MASA syndrome: a new heritable mental retardation syndrome. Clin. Genet. 5:298, 1974. 27. Binns, K. L., Taylor, P. P., Sicheri, F., et al.: Phosphorylation of tyrosine residues in the kinase domain and juxtamembrane region regulates the biological and catalytic activities of Eph receptors. Mol. Cell. Biol. 20:4791, 2000. 28. Bishop, A. L., and Hall, A.: Rho GTPases and their effector proteins. Biochem. J. 348(Pt. 2):241, 2000. 29. Bito, H., Furuyashiki, T., Ishihara, H., et al.: A critical role for a Rho-associated kinase, p160ROCK, in determining axon outgrowth in mammalian CNS neurons. Neuron 26:431, 2000. 30. Blackwood, D. H., Fordyce, A., Walker, M. T., et al.: Schizophrenia and affective disorders—cosegregation with a translocation at chromosome 1q42 that directly disrupts brain-expressed genes: clinical and P300 findings in a family. Am. J. Hum. Genet. 69:428, 2001. 31. Blanchoin, L., Amann, K. J., Higgs, H. N., et al.: Direct observation of dendritic actin filament networks nucleated by Arp2/3 complex and WASP/Scar proteins. Nature 404:1007, 2000. 32. Bokoch, G. M., Wang, Y., Bohl, B. P., et al.: Interaction of the Nck adapter protein with p21-activated kinase (PAK1). J. Biol. Chem. 271:25746, 1996. 33. Borrell, V., Del Rio, J. A., Alcantara, S., et al.: Reelin regulates the development and synaptogenesis of the layer-specific entorhino-hippocampal connections. J. Neurosci. 19:1345, 1999. 34. Bovolenta, P., and Fernaud-Espinosa, I.: Nervous system proteoglycans as modulators of neurite outgrowth. Prog. Neurobiol. 61:113, 2000. 35. Bradke, F., and Dotti, C. G.: The role of local actin instability in axon formation. Science 283:1931, 1999. 36. Bradke, F., and Dotti, C. G.: Differentiated neurons retain the capacity to generate axons from dendrites. Curr. Biol. 10:1467, 2000.

Guidance of Axons to Targets in Development and in Disease 37. Brittis, P. A., Lu, Q., and Flanagan, J. G.: Axonal protein synthesis provides a mechanism for localized regulation at an intermediate target. Cell 110:223, 2002. 38. Brittis, P. A., Silver, J., Walsh, F. S., and Doherty, P.: Fibroblast growth factor receptor function is required for the orderly projection of ganglion cell axons in the developing mammalian retina. Mol. Cell. Neurosci. 8:120, 1996. 39. Brose, K., Bland, K. S., Wang, K. H., et al.: Slit proteins bind Robo receptors and have an evolutionarily conserved role in repulsive axon guidance. Cell 96:795, 1999. 40. Brown, D. A., and Rose, J. K.: Sorting of GPI-anchored proteins to glycolipid-enriched membrane subdomains during transport to the apical cell surface. Cell 68:533, 1992. 41. Brown, J. D., and Moon, R. T.: Wnt signaling: why is everything so negative? Curr. Opin. Cell Biol. 10:182, 1998. 42. Brown, M. D., Cornejo, B. J., Kuhn, T. B., and Bamburg, J. R.: Cdc42 stimulates neurite outgrowth and formation of growth cone filopodia and lamellipodia. J. Neurobiol. 43:352, 2000. 43. Bruckner, K., Pablo Labrador, J., Scheiffele, P., et al.: EphrinB ligands recruit GRIP family PDZ adaptor proteins into raft membrane microdomains. Neuron 22:511, 1999. 44. Bruckner, K., Pasquale, E. B., and Klein, R.: Tyrosine phosphorylation of transmembrane ligands for Eph receptors. Science 275:1640, 1997. 45. Brummendorf, T., and Rathjen, F. G.: Structure/function relationships of axon-associated adhesion receptors of the immunoglobulin superfamily. Curr. Opin. Neurobiol. 6:584, 1996. 46. Bulow, H. E., Berry, K. L., Topper, L. H., et al.: Heparan sulfate proteoglycan-dependent induction of axon branching and axon misrouting by the Kallmann syndrome gene kal-1. Proc. Natl. Acad. Sci. U. S. A. 99:6346, 2002. 47. Burden-Gulley, S. M., Pendergast, M., and Lemmon, V.: The role of cell adhesion molecule L1 in axonal extension, growth cone motility, and signal transduction. Cell Tissue Res. 290:415, 1997. 48. Campbell, D. S., and Holt, C. E.: Chemotropic responses of retinal growth cones mediated by rapid local protein synthesis and degradation. Neuron 32:1013, 2001. 49. Campbell, D. S., and Holt, C. E.: Apoptotic pathway and MAPKs differentially regulate chemotropic responses of retinal growth cones. Neuron 37:939, 2003. 50. Cannon, T. D., Kaprio, J., Lonnqvist, J., et al.: The genetic epidemiology of schizophrenia in a Finnish twin cohort: a population-based modeling study. Arch. Gen. Psychiatry 55:67, 1998. 51. Carlsson, L., Nystrom, L. E., Sundkvist, I., et al.: Actin polymerizability is influenced by profilin, a low molecular weight protein in non-muscle cells. J. Mol. Biol. 115:465, 1977. 52. Carter, N., Nakamoto, T., Hirai, H., and Hunter, T.: EphrinA1-induced cytoskeletal re-organization requires FAK and p130(cas). Nat. Cell Biol. 4:565, 2002. 53. Cassimeris, L., Safer, D., Nachmias, V. T., and Zigmond, S. H.: Thymosin beta 4 sequesters the majority of G-actin in resting human polymorphonuclear leukocytes. J. Cell Biol. 119:1261, 1992. 54. Castellani, V., Chedotal, A., Schachner, M., et al.: Analysis of the L1-deficient mouse phenotype reveals cross-talk

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

471

between Sema3A and L1 signaling pathways in axonal guidance. Neuron 27:237, 2000. Castellani, V., De Angelis, E., Kenwrick, S., and Rougon, G.: Cis and trans interactions of L1 with neuropilin-1 control axonal responses to semaphorin 3A. EMBO J. 21:6348, 2002. Chan, S. S., Zheng, H., Su, M. W., et al.: UNC-40, a C.elegans homolog of DCC (Deleted in Colorectal Cancer), is required in motile cells responding to UNC-6 netrin cues. Cell 87:187, 1996. Chedotal, A., Del Rio, J. A., Ruiz, M., et al.: Semaphorins III and IV repel hippocampal axons via two distinct receptors. Development 125:4313, 1998. Chen, J. H., Wen, L., Dupuis, S., et al.: The N-terminal leucine-rich regions in Slit are sufficient to repel olfactory bulb axons and subventricular zone neurons. J. Neurosci. 21:1548, 2001. Cheng, Q., Sasaki, Y., Shoji, M., et al.: Cdk5/p35 and Rhokinase mediate ephrin-A5-induced signaling in retinal ganglion cells. Mol. Cell. Neurosci. 24:632, 2003. Cherfils, J., and Chardin, P.: GEFs: structural basis for their activation of small GTP-binding proteins. Trends Biochem. Sci. 24:306, 1999. Chien, C. B., Rosenthal, D. E., Harris, W. A., and Holt, C. E.: Navigational errors made by growth cones without filopodia in the embryonic Xenopus brain. Neuron 11:237, 1993. Chikumi, H., Fukuhara, S., and Gutkind, J. S.: Regulation of G protein-linked guanine nucleotide exchange factors for Rho, PDZ-RhoGEF, and LARG by tyrosine phosphorylation: evidence of a role for focal adhesion kinase. J. Biol. Chem. 277:12463, 2002. Chong, L. D., Park, E. K., Latimer, E., et al.: Fibroblast growth factor receptor-mediated rescue of x-ephrin B1induced cell dissociation in Xenopus embryos. Mol. Cell. Biol. 20:724, 2000. Clipstone, N. A., and Crabtree, G. R.: Identification of calcineurin as a key signalling enzyme in T-lymphocyte activation. Nature 357:695, 1992. Colamarino, S. A., and Tessier-Lavigne, M.: The axonal chemoattractant netrin-1 is also a chemorepellent for trochlear motor axons. Cell 81:621, 1995. Colavita, A., and Culotti, J. G.: Suppressors of ectopic UNC-5 growth cone steering identify eight genes involved in axon guidance in Caenorhabditis elegans. Dev. Biol. 194:72, 1998. Comeau, M. R., Johnson, R., DuBose, R. F., et al.: A poxvirus-encoded semaphorin induces cytokine production from monocytes and binds to a novel cellular semaphorin receptor, VESPR. Immunity 8:473, 1998. Conrad, A. J., and Scheibel, A. B.: Schizophrenia and the hippocampus: the embryological hypothesis extended. Schizophr. Bull. 13:577, 1987. Corset, V., Nguyen-Ba-Charvet, K. T., Forcet, C., et al.: Netrin-1-mediated axon outgrowth and cAMP production requires interaction with adenosine A2b receptor. Nature 407:747, 2000. Courtney, K. D., Grove, M., Vandongen, H., et al.: Localization and phosphorylation of Abl-interactor proteins, Abi-1 and Abi-2, in the developing nervous system. Mol. Cell. Neurosci. 16:244, 2000.

472

Neurobiology of the Peripheral Nervous System

71. Coutinho, S., Jahn, T., Lewitzky, M., et al.: Characterization of Grb4, an adapter protein interacting with Bcr-Abl. Blood 96:618, 2000. 72. Cowan, C. A., and Henkemeyer, M.: The SH2/SH3 adaptor Grb4 transduces B-ephrin reverse signals. Nature 413:174, 2001. 73. da Costa, S. R., Okamoto, C. T., and Hamm-Alvarez, S. F.: Actin microfilaments et al.—the many components, effectors and regulators of epithelial cell endocytosis. Adv. Drug Deliv. Rev. 55:1359, 2003. 74. Da Silva, J. S., Medina, M., Zuliani, C., et al.: RhoA/ROCK regulation of neuritogenesis via profilin IIa-mediated control of actin stability. J. Cell Biol. 162:1267, 2003. 75. Dahme, M., Bartsch, U., Martini, R., et al.: Disruption of the mouse L1 gene leads to malformations of the nervous system. Nat. Genet. 17:346, 1997. 76. Davenport, R. W., Dou, P., Rehder, V., and Kater, S. B.: A sensory role for neuronal growth cone filopodia. Nature 361:721, 1993. 77. Davis, J. Q., and Bennett, V.: Ankyrin binding activity shared by the neurofascin/L1/NrCAM family of nervous system cell adhesion molecules. J. Biol. Chem. 269:27163, 1994. 78. Davis, S., Gale, N. W., Aldrich, T. H., et al.: Ligands for EPH-related receptor tyrosine kinases that require membrane attachment or clustering for activity. Science 266:816, 1994. 79. Davy, A., Gale, N. W., Murray, E. W., et al.: Compartmentalized signaling by GPI-anchored ephrin-A5 requires the Fyn tyrosine kinase to regulate cellular adhesion. Genes Dev. 13:3125, 1999. 80. Davy, A., and Robbins, S. M.: Ephrin-A5 modulates cell adhesion and morphology in an integrin-dependent manner. EMBO J. 19:5396, 2000. 81. de Castro, F., Hu, L., Drabkin, H., et al.: Chemoattraction and chemorepulsion of olfactory bulb axons by different secreted semaphorins. J. Neurosci. 19:4428, 1999. 82. de la Torre, J. R., Hopker, V. H., Ming, G. L., et al.: Turning of retinal growth cones in a netrin-1 gradient mediated by the netrin receptor DCC. Neuron 19:1211, 1997. 83. de Wit, J., and Verhaagen, J.: Role of semaphorins in the adult nervous system. Prog. Neurobiol. 71:249, 2003. 84. Deardorff, M. A., Tan, C., Saint-Jeannet, J. P., and Klein, P. S.: A role for frizzled 3 in neural crest development. Development 128:3655, 2001. 85. Dickson, B. J.: Rho GTPases in growth cone guidance. Curr. Opin. Neurobiol. 11:103, 2001. 86. Dode, C., Levilliers, J., Dupont, J. M., et al.: Loss-of-function mutations in FGFR1 cause autosomal dominant Kallmann syndrome. Nat. Genet. 33:463, 2003. 87. Doherty, P., Williams, E., and Walsh, F. S.: A soluble chimeric form of the L1 glycoprotein stimulates neurite outgrowth. Neuron 14:57, 1995. 88. Drescher, U., Bonhoeffer, F., and Muller, B. K.: The Eph family in retinal axon guidance. Curr. Opin. Neurobiol. 7:75, 1997. 89. Drewes, G., Ebneth, A., and Mandelkow, E. M.: MAPs, MARKs and microtubule dynamics. Trends Biochem. Sci. 23:307, 1998. 90. Driessens, M. H., Hu, H., Nobes, C. D., et al.: Plexin-B semaphorin receptors interact directly with active Rac and

91.

92.

93.

94.

95.

96.

97.

98. 99.

100.

101.

102.

103.

104.

105.

106.

107.

regulate the actin cytoskeleton by activating Rho. Curr. Biol. 11:339, 2001. Eastwood, S. L., Law, A. J., Everall, I. P., and Harrison, P. J.: The axonal chemorepellant semaphorin 3A is increased in the cerebellum in schizophrenia and may contribute to its synaptic pathology. Mol. Psychiatry 8:148, 2003. Edwards, D. C., Sanders, L. C., Bokoch, G. M., and Gill, G. N.: Activation of LIM-kinase by Pak1 couples Rac/Cdc42 GTPase signalling to actin cytoskeletal dynamics. Nat. Cell Biol. 1:253, 1999. Eickholt, B. J., Walsh, F. S., and Doherty, P.: An inactive pool of GSK-3 at the leading edge of growth cones is implicated in Semaphorin 3A signaling. J. Cell Biol. 157:211, 2002. Elkins, T., Hortsch, M., Bieber, A. J., et al.: Drosophila fasciclin I is a novel homophilic adhesion molecule that along with fasciclin III can mediate cell sorting. J. Cell Biol. 110:1825, 1990. Elowe, S., Holland, S. J., Kulkarni, S., and Pawson, T.: Downregulation of the Ras-mitogen-activated protein kinase pathway by the EphB2 receptor tyrosine kinase is required for ephrin-induced neurite retraction. Mol. Cell. Biol. 21:7429, 2001. Endo, M., Ohashi, K., Sasaki, Y., et al.: Control of growth cone motility and morphology by LIM kinase and Slingshot via phosphorylation and dephosphorylation of cofilin. J. Neurosci. 23:2527, 2003. Engelkamp, D.: Cloning of three mouse unc5 genes and their expression patterns at mid-gestation. Mech. Dev. 118:191, 2002. Etienne-Manneville, S., and Hall, A.: Rho GTPases in cell biology. Nature 420:629, 2002. Evangelista, M., Pruyne, D., Amberg, D. C., et al.: Formins direct Arp2/3-independent actin filament assembly to polarize cell growth in yeast. Nat. Cell Biol. 4:260, 2002. Fan, J., Mansfield, S. G., Redmond, T., et al.: The organization of F-actin and microtubules in growth cones exposed to a brain-derived collapsing factor. J. Cell Biol. 121:867, 1993. Fan, J., and Raper, J. A.: Localized collapsing cues can steer growth cones without inducing their full collapse. Neuron 14:263, 1995. Fan, X., Labrador, J. P., Hing, H., and Bashaw, G. J.: Slit stimulation recruits Dock and Pak to the roundabout receptor and increases Rac activity to regulate axon repulsion at the CNS midline. Neuron 40:113, 2003. Fannon, A. M., and Colman, D. R.: A model for central synaptic junctional complex formation based on the differential adhesive specificities of the cadherins. Neuron 17:423, 1996. Fazeli, A., Dickinson, S. L., Hermiston, M. L., et al.: Phenotype of mice lacking functional Deleted in colorectal cancer (Dcc) gene. Nature 386:796, 1997. Flanagan, J. G., and Vanderhaeghen, P.: The ephrins and Eph receptors in neural development. Annu. Rev. Neurosci. 21:309, 1998. Flanagan, W. M., Corthesy, B., Bram, R. J., and Crabtree, G. R.: Nuclear association of a T-cell transcription factor blocked by FK-506 and cyclosporin A. Nature 352:803, 1991. Forcet, C., Stein, E., Pays, L., et al.: Netrin-1-mediated axon outgrowth requires deleted in colorectal cancer-dependent MAPK activation. Nature 417:443, 2002.

Guidance of Axons to Targets in Development and in Disease 108. Forscher, P., and Smith, S. J.: Actions of cytochalasins on the organization of actin filaments and microtubules in a neuronal growth cone. J. Cell Biol. 107:1505, 1988. 109. Franco, B., Guioli, S., Pragliola, A., et al.: A gene deleted in Kallmann’s syndrome shares homology with neural cell adhesion and axonal path-finding molecules. Nature 353:529, 1991. 110. Fransen, E., Lemmon, V., Van Camp, G., et al.: CRASH syndrome: clinical spectrum of corpus callosum hypoplasia, retardation, adducted thumbs, spastic paraparesis and hydrocephalus due to mutations in one single gene, L1. Eur. J. Hum. Genet. 3:273, 1995. 111. Freeman, N. L., Chen, Z., Horenstein, J., et al.: An actin monomer binding activity localizes to the carboxyl-terminal half of the Saccharomyces cerevisiae cyclase-associated protein. J. Biol. Chem. 270:5680, 1995. 112. Fricke, C., Lee, J. S., Geiger-Rudolph, S., et al.: Astray, a zebrafish roundabout homolog required for retinal axon guidance. Science 292:507, 2001. 113. Fritsche, J., Reber, B. F., Schindelholz, B., and Bandtlow, C. E.: Differential cytoskeletal changes during growth cone collapse in response to hSema III and thrombin. Mol. Cell. Neurosci. 14:398, 1999. 114. Fujisawa, H., and Kitsukawa, T.: Receptors for collapsin/semaphorins. Curr. Opin. Neurobiol. 8:587, 1998. 115. Fujiwara, T., Mammoto, A., Kim, Y., and Takai, Y.: Rho small G-protein-dependent binding of mDia to an Src homology 3 domain-containing IRSp53/BAIAP2. Biochem. Biophys. Res. Commun. 271:626, 2000. 116. Fukata, Y., Itoh, T. J., Kimura, T., et al.: CRMP-2 binds to tubulin heterodimers to promote microtubule assembly. Nat. Cell Biol. 4:583, 2002. 117. Galisteo, M. L., Chernoff, J., Su, Y. C., et al.: The adaptor protein Nck links receptor tyrosine kinases with the serinethreonine kinase Pak1. J. Biol. Chem. 271:20997, 1996. 118. Galko, M. J., and Tessier-Lavigne, M.: Function of an axonal chemoattractant modulated by metalloprotease activity. Science 289:1365, 2000. 119. Garrity, P. A., Rao, Y., Salecker, I., et al.: Drosophila photoreceptor axon guidance and targeting requires the dreadlocks SH2/SH3 adapter protein. Cell 85:639, 1996. 120. Garver, T. D., Ren, Q., Tuvia, S., and Bennett, V.: Tyrosine phosphorylation at a site highly conserved in the L1 family of cell adhesion molecules abolishes ankyrin binding and increases lateral mobility of neurofascin. J. Cell Biol. 137:703, 1997. 121. Gasman, S., Kalaidzidis, Y., and Zerial, M.: RhoD regulates endosome dynamics through Diaphanous-related Formin and Src tyrosine kinase. Nat. Cell Biol. 5:195, 2003. 122. Gavin, A. C., Bosche, M., Krause, R., et al.: Functional organization of the yeast proteome by systematic analysis of protein complexes. Nature 415:141, 2002. 123. Gertler, F. B., Comer, A. R., Juang, J. L., et al.: Enabled, a dosage-sensitive suppressor of mutations in the Drosophila Abl tyrosine kinase, encodes an Abl substrate with SH3 domain-binding properties. Genes Dev. 9:521, 1995. 124. Gertler, F. B., Doctor, J. S., and Hoffmann, F. M.: Genetic suppression of mutations in the Drosophila abl protooncogene homolog. Science 248:857, 1990.

473

125. Gertler, F. B., Niebuhr, K., Reinhard, M., et al.: Mena, a relative of VASP and Drosophila Enabled, is implicated in the control of microfilament dynamics. Cell 87:227, 1996. 126. Giger, R. J., Pasterkamp, R. J., Heijnen, S., et al.: Anatomical distribution of the chemorepellent semaphorin III/collapsin-1 in the adult rat and human brain: predominant expression in structures of the olfactory-hippocampal pathway and the motor system. J. Neurosci. Res 52:27, 1998. 127. Gingras, A. C., Raught, B., and Sonenberg, N.: eIF4 initiation factors: effectors of mRNA recruitment to ribosomes and regulators of translation. Annu. Rev. Biochem. 68:913, 1999. 128. Giordano, S., Corso, S., Conrotto, P., et al.: The semaphorin 4D receptor controls invasive growth by coupling with Met. Nat. Cell Biol. 4:720, 2002. 129. Gitai, Z., Yu, T. W., Lundquist, E. A., et al.: The netrin receptor UNC-40/DCC stimulates axon attraction and outgrowth through enabled and, in parallel, Rac and UNC115/AbLIM. Neuron 37:53, 2003. 130. Gong, Q., Rangarajan, R., Seeger, M., and Gaul, U.: The netrin receptor frazzled is required in the target for establishment of retinal projections in the Drosophila visual system. Development 126:1451, 1999. 131. Gong, Q., and Shipley, M. T.: Evidence that pioneer olfactory axons regulate telencephalon cell cycle kinetics to induce the formation of the olfactory bulb. Neuron 14:91, 1995. 132. Goshima, Y., Nakamura, F., Strittmatter, P., and Strittmatter, S. M.: Collapsin-induced growth cone collapse mediated by an intracellular protein related to UNC-33. Nature 376:509, 1995. 133. Graef, I. A., Wang, F., Charron, F., et al.: Neurotrophins and netrins require calcineurin/NFAT signaling to stimulate outgrowth of embryonic axons. Cell 113:657, 2003. 134. Grant, S. G., Karl, K. A., Kiebler, M. A., and Kandel, E. R.: Focal adhesion kinase in the brain: novel subcellular localization and specific regulation by Fyn tyrosine kinase in mutant mice. Genes Dev. 9:1909, 1995. 135. Grootjans, J. J., Zimmermann, P., Reekmans, G., et al.: Syntenin, a PDZ protein that binds syndecan cytoplasmic domains. Proc. Natl. Acad. Sci. U. S. A. 94:13683, 1997. 136. Grove, E. A., Tole, S., Limon, J., et al.: The hem of the embryonic cerebral cortex is defined by the expression of multiple Wnt genes and is compromised in Gli3-deficient mice. Development 125:2315, 1998. 137. Guidotti, A., Auta, J., Davis, J. M., et al.: Decrease in reelin and glutamic acid decarboxylase67 (GAD67) expression in schizophrenia and bipolar disorder: a postmortem brain study. Arch. Gen. Psychiatry 57:1061, 2000. 138. Gutwein, P., Oleszewski, M., Mechtersheimer, S., et al.: Role of Src kinases in the ADAM-mediated release of L1 adhesion molecule from human tumor cells. J. Biol. Chem. 275:15490, 2000. 139. Haarer, B. K., Petzold, A. S., and Brown, S. S.: Mutational analysis of yeast profilin. Mol. Cell. Biol. 13:7864, 1993. 140. Haffner, C., Jarchau, T., Reinhard, M., et al.: Molecular cloning, structural analysis and functional expression of the proline-rich focal adhesion and microfilament-associated protein VASP. EMBO J. 14:19, 1995.

474

Neurobiology of the Peripheral Nervous System

141. Halbrugge, M., and Walter, U.: Purification of a vasodilatorregulated phosphoprotein from human platelets. Eur. J. Biochem. 185:41, 1989. 142. Hall, C., Brown, M., Jacobs, T., et al.: Collapsin response mediator protein switches RhoA and Rac1 morphology in N1E-115 neuroblastoma cells and is regulated by Rho kinase. J. Biol. Chem. 276:43482, 2001. 143. Hamada, F., Tomoyasu, Y., Takatsu, Y., et al.: Negative regulation of Wingless signaling by D-axin, a Drosophila homolog of axin. Science 283:1739, 1999. 144. Hamelin, M., Zhou, Y., Su, M. W., et al.: Expression of the UNC-5 guidance receptor in the touch neurons of C. elegans steers their axons dorsally. Nature 364:327, 1993. 145. Hanger, D. P., Hughes, K., Woodgett, J. R., et al.: Glycogen synthase kinase-3 induces Alzheimer’s disease-like phosphorylation of tau: generation of paired helical filament epitopes and neuronal localisation of the kinase. Neurosci. Lett. 147:58, 1992. 146. Harrison, P. J.: The neuropathology of schizophrenia: a critical review of the data and their interpretation. Brain 122(Pt. 4):593, 1999. 147. Hatten, M. E.: New directions in neuronal migration. Science 297:1660, 2002. 148. Hattori, M., Osterfield, M., and Flanagan, J. G.: Regulated cleavage of a contact-mediated axon repellent. Science 289:1360, 2000. 149. He, Z., and Tessier-Lavigne, M.: Neuropilin is a receptor for the axonal chemorepellent Semaphorin III. Cell 90:739, 1997. 150. Hebert, J. M., Lin, M., Partanen, J., et al.: FGF signaling through FGFR1 is required for olfactory bulb morphogenesis. Development 130:1101, 2003. 151. Hedgecock, E. M., Culotti, J. G., and Hall, D. H.: The unc-5, unc-6, and unc-40 genes guide circumferential migrations of pioneer axons and mesodermal cells on the epidermis in C. elegans. Neuron 4:61, 1990. 152. Hedgecock, E. M., Culotti, J. G., Hall, D. H., and Stern, B. D.: Genetics of cell and axon migrations in Caenorhabditis elegans. Development 100:365, 1987. 153. Hendriks, W., Schepens, J., Bachner, D., et al.: Molecular cloning of a mouse epithelial protein-tyrosine phosphatase with similarities to submembranous proteins. J. Cell. Biochem. 59:418, 1995. 154. Henkemeyer, M., Orioli, D., Henderson, J. T., et al.: Nuk controls pathfinding of commissural axons in the mammalian central nervous system. Cell 86:35, 1996. 155. Himanen, J. P., Rajashankar, K. R., Lackmann, M., et al.: Crystal structure of an Eph receptor-ephrin complex. Nature 414:933, 2001. 156. Hiramoto, M., Hiromi, Y., Giniger, E., and Hotta, Y.: The Drosophila Netrin receptor Frazzled guides axons by controlling Netrin distribution. Nature 406:886, 2000. 157. Hirotani, M., Ohoka, Y., Yamamoto, T., et al.: Interaction of plexin-B1 with PDZ domain-containing Rho guanine nucleotide exchange factors. Biochem. Biophys. Res. Commun. 297:32, 2002. 158. Hobert, O., Jallal, B., Schlessinger, J., and Ullrich, A.: Novel signaling pathway suggested by SH3 domain-mediated p95vav/heterogeneous ribonucleoprotein K interaction. J. Biol. Chem. 269:20225, 1994.

159. Holland, S. J., Gale, N. W., Mbamalu, G., et al.: Bidirectional signalling through the EPH-family receptor Nuk and its transmembrane ligands. Nature 383:722, 1996. 160. Holtmaat, A. J., Gorter, J. A., De Wit, J., et al.: Transient downregulation of Sema3A mRNA in a rat model for temporal lobe epilepsy: a novel molecular event potentially contributing to mossy fiber sprouting. Exp. Neurol. 182:142, 2003. 161. Honda, H., Oda, H., Nakamoto, T., et al.: Cardiovascular anomaly, impaired actin bundling and resistance to Srcinduced transformation in mice lacking p130Cas. Nat. Genet. 19:361, 1998. 162. Hong, K., Hinck, L., Nishiyama, M., et al.: A ligand-gated association between cytoplasmic domains of UNC5 and DCC family receptors converts netrin-induced growth cone attraction to repulsion. Cell 97:927, 1999. 163. Hong, K., Nishiyama, M., Henley, J., et al.: Calcium signalling in the guidance of nerve growth by netrin-1. Nature 403:93, 2000. 164. Hong, S. E., Shugart, Y. Y., Huang, D. T., et al.: Autosomal recessive lissencephaly with cerebellar hypoplasia is associated with human RELN mutations. Nat. Genet. 26:93, 2000. 165. Hornstein, I., Alcover, A., and Katzav, S.: Vav proteins, masters of the world of cytoskeleton organization. Cell. Signal. 16:1, 2004. 166. Hu, H.: Cell-surface heparan sulfate is involved in the repulsive guidance activities of Slit2 protein. Nat. Neurosci. 4:695, 2001. 167. Hu, H., Marton, T. F., and Goodman, C. S.: Plexin B mediates axon guidance in Drosophila by simultaneously inhibiting active Rac and enhancing RhoA signaling. Neuron 32:39, 2001. 168. Huai, J., and Drescher, U.: An ephrin-A-dependent signaling pathway controls integrin function and is linked to the tyrosine phosphorylation of a 120-kDa protein. J. Biol. Chem. 276:6689, 2001. 169. Huang, X., Cheng, H. J., Tessier-Lavigne, M., and Jin, Y.: MAX-1, a novel PH/MyTH4/FERM domain cytoplasmic protein implicated in netrin-mediated axon repulsion. Neuron 34:563, 2002. 170. Hug, C., Jay, P. Y., Reddy, I., et al.: Capping protein levels influence actin assembly and cell motility in Dictyostelium. Cell 81:591, 1995. 171. Ignelzi, M. A. Jr., Miller, D. R., Soriano, P., and Maness, P. F.: Impaired neurite outgrowth of src-minus cerebellar neurons on the cell adhesion molecule L1. Neuron 12:873, 1994. 172. Ishii, N., Wadsworth, W. G., Stern, B. D., et al.: UNC-6, a laminin-related protein, guides cell and pioneer axon migrations in C. elegans. Neuron 9:873, 1992. 173. Ishikawa, Y., Katoh, H., and Negishi, M.: A role of Rnd1 GTPase in dendritic spine formation in hippocampal neurons. J. Neurosci. 23:11065, 2003. 174. Ishizaki, T., Maekawa, M., Fujisawa, K., et al.: The small GTP-binding protein Rho binds to and activates a 160 kDa Ser/Thr protein kinase homologous to myotonic dystrophy kinase. EMBO J. 15:1885, 1996. 175. Jalink, K., Eichholtz, T., Postma, F. R., et al.: Lysophosphatidic acid induces neuronal shape changes via a novel, receptor-mediated signaling pathway: similarity to thrombin action. Cell Growth Differ. 4:247, 1993.

Guidance of Axons to Targets in Development and in Disease 176. Jalink, K., van Corven, E. J., Hengeveld, T., et al.: Inhibition of lysophosphatidate- and thrombin-induced neurite retraction and neuronal cell rounding by ADP ribosylation of the small GTP-binding protein Rho. J. Cell Biol. 126:801, 1994. 177. Jin, Z., and Strittmatter, S. M.: Rac1 mediates collapsin-1induced growth cone collapse. J. Neurosci. 17:6256, 1997. 178. Kabir, N., Schaefer, A. W., Nakhost, A., et al.: Protein kinase C activation promotes microtubule advance in neuronal growth cones by increasing average microtubule growth lifetimes. J. Cell Biol. 152:1033, 2001. 179. Kallmann, F. J., Schoenfeld, W. A., and Barrera, S. E.: The genetic aspects of primary eunuchoidism. Am. J. Ment. Defic. XLVIII:203, 1944. 180. Kalo, M. S., and Pasquale, E. B.: Multiple in vivo tyrosine phosphorylation sites in EphB receptors. Biochemistry 38:14396, 1999. 181. Kamiguchi, H., Hlavin, M. L., Yamasaki, M., and Lemmon, V.: Adhesion molecules and inherited diseases of the human nervous system. Annu. Rev. Neurosci. 21:97, 1998. 182. Kamm, K. E., and Stull, J. T.: Myosin phosphorylation, force, and maximal shortening velocity in neurally stimulated tracheal smooth muscle. Am. J. Physiol. 249:C238, 1985. 183. Kang, F., Purich, D. L., and Southwick, F. S.: Profilin promotes barbed-end actin filament assembly without lowering the critical concentration. J. Biol. Chem. 274:36963, 1999. 184. Katsu, T., Ujike, H., Nakano, T., et al.: The human frizzled-3 (FZD3) gene on chromosome 8p21, a receptor gene for Wnt ligands, is associated with the susceptibility to schizophrenia. Neurosci. Lett. 353:53, 2003. 185. Kaufmann, N., Wills, Z. P., and Van Vactor, D.: Drosophila Rac1 controls motor axon guidance. Development 125:453, 1998. 186. Keino-Masu, K., Masu, M., Hinck, L., et al.: Deleted in Colorectal Cancer (DCC) encodes a netrin receptor. Cell 87:175, 1996. 187. Keleman, K., Rajagopalan, S., Cleppien, D., et al.: Comm sorts robo to control axon guidance at the Drosophila midline. Cell 110:415, 2002. 188. Kennedy, T. E.: Cellular mechanisms of netrin function: long-range and short-range actions. Biochem. Cell Biol. 78:569, 2000. 189. Kennedy, T. E., Serafini, T., de la Torre, J. R., and TessierLavigne, M.: Netrins are diffusible chemotropic factors for commissural axons in the embryonic spinal cord. Cell 78:425, 1994. 190. Kidd, T., Bland, K. S., and Goodman, C. S.: Slit is the midline repellent for the robo receptor in Drosophila. Cell 96:785, 1999. 191. Kimura, K., Ito, M., Amano, M., et al.: Regulation of myosin phosphatase by Rho and Rho-associated kinase (Rho-kinase). Science 273:245, 1996. 192. Klee, C. B., Crouch, T. H., and Krinks, M. H.: Calcineurin: a calcium- and calmodulin-binding protein of the nervous system. Proc. Natl. Acad. Sci. U. S. A. 76:6270, 1979. 193. Klostermann, A., Lohrum, M., Adams, R. H., and Puschel, A. W.: The chemorepulsive activity of the axonal guidance signal semaphorin D requires dimerization. J. Biol. Chem. 273:7326, 1998. 194. Knoll, B., and Drescher, U.: Ephrin-As as receptors in topographic projections. Trends Neurosci. 25:145, 2002.

475

195. Knoll, B., Zarbalis, K., Wurst, W., and Drescher, U.: A role for the EphA family in the topographic targeting of vomeronasal axons. Development 128:895, 2001. 196. Kobielak, A., Pasolli, H. A., and Fuchs, E.: Mammalian formin-1 participates in adherens junctions and polymerization of linear actin cables. Nat. Cell Biol. 6:21, 2004. 197. Koeberle, P. D., and Bahr, M.: Growth and guidance cues for regenerating axons: where have they gone? J. Neurobiol. 59:162, 2004. 198. Kolodkin, A. L., Levengood, D. V., Rowe, E. G., et al.: Neuropilin is a semaphorin III receptor. Cell 90:753, 1997. 199. Kolodkin, A. L., Matthes, D. J., O’Connor, T. P., et al.: Fasciclin IV: sequence, expression, and function during growth cone guidance in the grasshopper embryo. Neuron 9:831, 1992. 200. Kolodziej, P. A., Timpe, L. C., Mitchell, K. J., et al.: Frazzled encodes a Drosophila member of the DCC immunoglobulin subfamily and is required for CNS and motor axon guidance. Cell 87:197, 1996. 201. Kozasa, T., Jiang, X., Hart, M. J., et al.: p115 RhoGEF, a GTPase activating protein for Galpha12 and Galpha13. Science 280:2109, 1998. 202. Kozma, R., Ahmed, S., Best, A., and Lim, L.: The Ras-related protein Cdc42Hs and bradykinin promote formation of peripheral actin microspikes and filopodia in Swiss 3T3 fibroblasts. Mol. Cell. Biol. 15:1942, 1995. 203. Krause, M., Dent, E. W., Bear, J. E., et al.: Ena/VASP proteins: regulators of the actin cytoskeleton and cell migration. Annu. Rev. Cell Dev. Biol. 19:541, 2003. 204. Krebs, A., Rothkegel, M., Klar, M., and Jockusch, B. M.: Characterization of functional domains of mDia1, a link between the small GTPase Rho and the actin cytoskeleton. J. Cell Sci. 114:3663, 2001. 205. Krugmann, S., Jordens, I., Gevaert, K., et al.: Cdc42 induces filopodia by promoting the formation of an IRSp53:Mena complex. Curr. Biol. 11:1645, 2001. 206. Kuhn, T. B., Meberg, P. J., Brown, M. D., et al.: Regulating actin dynamics in neuronal growth cones by ADF/cofilin and rho family GTPases. J. Neurobiol. 44:126, 2000. 207. Kuhn, T. B., Stoeckli, E. T., Condrau, M. A., et al.: Neurite outgrowth on immobilized axonin-1 is mediated by a heterophilic interaction with L1(G4). J. Cell Biol. 115:1113, 1991. 208. Kullander, K., Butt, S. J., Lebret, J. M., et al.: Role of EphA4 and EphrinB3 in local neuronal circuits that control walking. Science 299:1889, 2003. 209. Kullander, K., Croll, S. D., Zimmer, M., et al.: Ephrin-B3 is the midline barrier that prevents corticospinal tract axons from recrossing, allowing for unilateral motor control. Genes Dev. 15:877, 2001. 210. Kullander, K., and Klein, R.: Mechanisms and functions of Eph and ephrin signalling. Nat. Rev. Mol. Cell. Biol. 3:475, 2002. 211. Kullander, K., Mather, N. K., Diella, F., et al.: Kinasedependent and kinase-independent functions of EphA4 receptors in major axon tract formation in vivo. Neuron 29:73, 2001. 212. Kumanogoh, A., Marukawa, S., Suzuki, K., et al.: Class IV semaphorin Sema4A enhances T-cell activation and interacts with Tim-2. Nature 419:629, 2002.

476

Neurobiology of the Peripheral Nervous System

213. Kumanogoh, A., Watanabe, C., Lee, I., et al.: Identification of CD72 as a lymphocyte receptor for the class IV semaphorin CD100: a novel mechanism for regulating B cell signaling. Immunity 13:621, 2000. 214. Kunz, S., Ziegler, U., Kunz, B., and Sonderegger, P.: Intracellular signaling is changed after clustering of the neural cell adhesion molecules axonin-1 and NgCAM during neurite fasciculation. J. Cell Biol. 135:253, 1996. 215. Lambert de Rouvroit, C., and Goffinet, A. M.: Neuronal migration. Mech. Dev. 105:47, 2001. 216. Lanier, L. M., Gates, M. A., Witke, W., et al.: Mena is required for neurulation and commissure formation. Neuron 22:313, 1999. 217. Legouis, R., Hardelin, J. P., Levilliers, J., et al.: The candidate gene for the X-linked Kallmann syndrome encodes a protein related to adhesion molecules. Cell 67:423, 1991. 218. Leighton, P. A., Mitchell, K. J., Goodrich, L. V., et al.: Defining brain wiring patterns and mechanisms through gene trapping in mice. Nature 410:174, 2001. 219. Lemmon, V., Farr, K. L., and Lagenaur, C.: L1-mediated axon outgrowth occurs via a homophilic binding mechanism. Neuron 2:1597, 1989. 220. Leonardo, E. D., Hinck, L., Masu, M., et al.: Vertebrate homologues of C. elegans UNC-5 are candidate netrin receptors. Nature 386:833, 1997. 221. Leung, T., Chen, X. Q., Manser, E., and Lim, L.: The p160 RhoA-binding kinase ROK alpha is a member of a kinase family and is involved in the reorganization of the cytoskeleton. Mol. Cell. Biol. 16:5313, 1996. 222. Leung-Hagesteijn, C., Spence, A. M., Stern, B. D., et al.: UNC-5, a transmembrane protein with immunoglobulin and thrombospondin type 1 domains, guides cell and pioneer axon migrations in C. elegans. Cell 71:289, 1992. 223. Li, H. S., Chen, J. H., Wu, W., et al.: Vertebrate slit, a secreted ligand for the transmembrane protein roundabout, is a repellent for olfactory bulb axons. Cell 96:807, 1999. 224. Li, X., Meriane, M., Triki, I., et al.: The adaptor protein Nck1 couples the netrin-1 receptor DCC (deleted in colorectal cancer) to the activation of the small GTPase Rac1 through an atypical mechanism. J. Biol. Chem. 277:37788, 2002. 225. Li, X., Saint-Cyr-Proulx, E., Aktories, K., and LamarcheVane, N.: Rac1 and Cdc42 but not RhoA or Rho kinase activities are required for neurite outgrowth induced by the Netrin-1 receptor DCC (deleted in colorectal cancer) in N1E-115 neuroblastoma cells. J. Biol. Chem. 277:15207, 2002. 226. Liang, Y., Annan, R. S., Carr, S. A., et al.: Mammalian homologues of the Drosophila slit protein are ligands of the heparan sulfate proteoglycan glypican-1 in brain. J. Biol. Chem. 274:17885, 1999. 227. Lin, C. H., Espreafico, E. M., Mooseker, M. S., and Forscher, P.: Myosin drives retrograde F-actin flow in neuronal growth cones. Neuron 16:769, 1996. 228. Lu, Q., Sun, E. E., Klein, R. S., and Flanagan, J. G.: Ephrin-B reverse signaling is mediated by a novel PDZRGS protein and selectively inhibits G protein-coupled chemoattraction. Cell 105:69, 2001. 229. Luby-Phelps, K., and Taylor, D. L.: Subcellular compartmentalization by local differentiation of cytoplasmic structure. Cell Motil. Cytoskeleton 10:28, 1988.

230. Luo, L.: Actin cytoskeleton regulation in neuronal morphogenesis and structural plasticity. Annu. Rev. Cell Dev. Biol. 18:601, 2002. 231. Luo, L., Hensch, T. K., Ackerman, L., et al.: Differential effects of the Rac GTPase on Purkinje cell axons and dendritic trunks and spines. Nature 379:837, 1996. 232. Luo, L., Liao, Y. J., Jan, L. Y., and Jan, Y. N.: Distinct morphogenetic functions of similar small GTPases: Drosophila Drac1 is involved in axonal outgrowth and myoblast fusion. Genes Dev. 8:1787, 1994. 233. Luo, Y., Raible, D., and Raper, J. A.: Collapsin: a protein in brain that induces the collapse and paralysis of neuronal growth cones. Cell 75:217, 1993. 234. Lyuksyutova, A. I., Lu, C. C., Milanesio, N., et al.: Anteriorposterior guidance of commissural axons by Wnt-frizzled signaling. Science 302:1984, 2003. 235. Machesky, L. M., and Insall, R. H.: Scar1 and the related Wiskott-Aldrich syndrome protein, WASP, regulate the actin cytoskeleton through the Arp2/3 complex. Curr. Biol. 8:1347, 1998. 236. Machesky, L. M., Mullins, R. D., Higgs, H. N., et al.: Scar, a WASp-related protein, activates nucleation of actin filaments by the Arp2/3 complex. Proc. Natl. Acad. Sci. U. S. A. 96:3739, 1999. 237. Maekawa, M., Ishizaki, T., Boku, S., et al.: Signaling from Rho to the actin cytoskeleton through protein kinases ROCK and LIM-kinase. Science 285:895, 1999. 238. Maestre de San Juan, A.: Falta total de los nervios olfactorios con anosmia en un individuo en quien existia una atrofia congénita de los testiculos y miembro viril. Siglo Med. 131:211, 1856. 239. Manser, E., Leung, T., Salihuddin, H., et al.: A brain serine/threonine protein kinase activated by Cdc42 and Rac1. Nature 367:40, 1994. 240. Marston, D. J., Dickinson, S., and Nobes, C. D.: Racdependent trans-endocytosis of ephrinBs regulates Ephephrin contact repulsion. Nat. Cell Biol. 5:879, 2003. 241. Mathern, G. W., Babb, T. L., Leite, J. P., et al.: The pathogenic and progressive features of chronic human hippocampal epilepsy. Epilepsy Res. 26:151, 1996. 242. McWhirter, J. R., and Wang, J. Y.: An actin-binding function contributes to transformation by the Bcr-Abl oncoprotein of Philadelphia chromosome-positive human leukemias. EMBO J. 12:1533, 1993. 243. Meberg, P. J., and Bamburg, J. R.: Increase in neurite outgrowth mediated by overexpression of actin depolymerizing factor. J. Neurosci. 20:2459, 2000. 244. Merz, D. C., and Culotti, J. G.: Genetic analysis of growth cone migrations in Caenorhabditis elegans. J. Neurobiol. 44:281, 2000. 245. Miao, H., Burnett, E., Kinch, M., et al.: Activation of EphA2 kinase suppresses integrin function and causes focal-adhesion-kinase dephosphorylation. Nat. Cell Biol. 2:62, 2000. 246. Miao, H., Wei, B. R., Peehl, D. M., et al.: Activation of EphA receptor tyrosine kinase inhibits the Ras/MAPK pathway. Nat. Cell Biol. 3:527, 2001. 247. Miki, H., Suetsugu, S., and Takenawa, T.: WAVE, a novel WASP-family protein involved in actin reorganization induced by Rac. EMBO J. 17:6932, 1998.

Guidance of Axons to Targets in Development and in Disease 248. Miki, H., and Takenawa, T.: WAVE2 serves a functional partner of IRSp53 by regulating its interaction with Rac. Biochem. Biophys. Res. Commun. 293:93, 2002. 249. Miki, H., Yamaguchi, H., Suetsugu, S., and Takenawa, T.: IRSp53 is an essential intermediate between Rac and WAVE in the regulation of membrane ruffling. Nature 408:732, 2000. 250. Mikule, K., Gatlin, J. C., de la Houssaye, B. A., and Pfenninger, K. H.: Growth cone collapse induced by semaphorin 3A requires 12/15-lipoxygenase. J. Neurosci. 22:4932, 2002. 251. Millar, J. K., Wilson-Annan, J. C., Anderson, S., et al.: Disruption of two novel genes by a translocation co-segregating with schizophrenia. Hum. Mol. Genet. 9:1415, 2000. 252. Minamide, L. S., Striegl, A. M., Boyle, J. A., et al.: Neurodegenerative stimuli induce persistent ADF/cofilinactin rods that disrupt distal neurite function. Nat. Cell Biol. 2:628, 2000. 253. Ming, G., Song, H., Berninger, B., et al.: Phospholipase C-gamma and phosphoinositide 3-kinase mediate cytoplasmic signaling in nerve growth cone guidance. Neuron 23:139, 1999. 254. Ming, G. L., Song, H. J., Berninger, B., et al.: cAMPdependent growth cone guidance by netrin-1. Neuron 19:1225, 1997. 255. Mitsui, N., Inatome, R., Takahashi, S., et al.: Involvement of Fes/Fps tyrosine kinase in semaphorin3A signaling. EMBO J. 21:3274, 2002. 256. Miyoshi, K., Honda, A., Baba, K., et al.: Disrupted-InSchizophrenia 1, a candidate gene for schizophrenia, participates in neurite outgrowth. Mol. Psychiatry 8:685, 2003. 257. Murakami, Y., Suto, F., Shimizu, M., et al.: Differential expression of plexin-A subfamily members in the mouse nervous system. Dev. Dyn. 220:246, 2001. 258. Myat, A., Henry, P., McCabe, V., et al.: Drosophila Nedd4, a ubiquitin ligase, is recruited by Commissureless to control cell surface levels of the roundabout receptor. Neuron 35:447, 2002. 259. Nadler, J. V.: The recurrent mossy fiber pathway of the epileptic brain. Neurochem. Res. 28:1649, 2003. 260. Naffah-Mazzacoratti, M. G., Arganaraz, G. A., Porcionatto, M. A., et al.: Selective alterations of glycosaminoglycans synthesis and proteoglycan expression in rat cortex and hippocampus in pilocarpine-induced epilepsy. Brain Res. Bull. 50:229, 1999. 261. Nayeem, N., Silletti, S., Yang, X., et al.: A potential role for the plasmin(ogen) system in the posttranslational cleavage of the neural cell adhesion molecule L1. J. Cell Sci. 112 (Pt 24):4739, 1999. 262. Newsome, T. P., Schmidt, S., Dietzl, G., et al.: Trio combines with dock to regulate Pak activity during photoreceptor axon pathfinding in Drosophila. Cell 101:283, 2000. 263. Nguyen Ba-Charvet, K. T., Brose, K., Ma, L., et al.: Diversity and specificity of actions of Slit2 proteolytic fragments in axon guidance. J. Neurosci. 21:4281, 2001. 264. Nguyen-Ba-Charvet, K. T., Brose, K., Marillat, V., et al.: Sensory axon response to substrate-bound Slit2 is modulated by laminin and cyclic GMP. Mol. Cell. Neurosci. 17:1048, 2001.

477

265. Nguyen-Ba-Charvet, K. T., Plump, A. S., Tessier-Lavigne, M., and Chedotal, A.: Slit1 and Slit2 proteins control the development of the lateral olfactory tract. J. Neurosci. 22:5473, 2002. 266. Niethammer, M., Smith, D. S., Ayala, R., et al.: NUDEL is a novel Cdk5 substrate that associates with LIS1 and cytoplasmic dynein. Neuron 28:697, 2000. 267. Nikolic, M., Chou, M. M., Lu, W., et al.: The p35/Cdk5 kinase is a neuron-specific Rac effector that inhibits Pak1 activity. Nature 395:194, 1998. 268. Nishida, E., Maekawa, S., and Sakai, H.: Cofilin, a protein in porcine brain that binds to actin filaments and inhibits their interactions with myosin and tropomyosin. Biochemistry 23:5307, 1984. 269. Nishiki, T., Narumiya, S., Morii, N., et al.: ADP-ribosylation of the rho/rac proteins induces growth inhibition, neurite outgrowth and acetylcholine esterase in cultured PC-12 cells. Biochem. Biophys. Res. Commun. 167:265, 1990. 270. Nishiyama, M., Hoshino, A., Tsai, L., et al.: Cyclic AMP/GMP-dependent modulation of Ca2⫹ channels sets the polarity of nerve growth-cone turning. Nature 424:990, 2003. 271. Niwa, R., Nagata-Ohashi, K., Takeichi, M., et al.: Control of actin reorganization by Slingshot, a family of phosphatases that dephosphorylate ADF/cofilin. Cell 108:233, 2002. 272. Nobes, C. D., and Hall, A.: Rho, rac, and cdc42 GTPases regulate the assembly of multimolecular focal complexes associated with actin stress fibers, lamellipodia, and filopodia. Cell 81:53, 1995. 273. Ohashi, K., Nagata, K., Maekawa, M., et al.: Rho-associated kinase ROCK activates LIM-kinase 1 by phosphorylation at threonine 508 within the activation loop. J. Biol. Chem. 275:3577, 2000. 274. Oinuma, I., Katoh, H., Harada, A., and Negishi, M.: Direct interaction of Rnd1 with Plexin-B1 regulates PDZRhoGEF-mediated Rho activation by Plexin-B1 and induces cell contraction in COS-7 cells. J. Biol. Chem. 278:25671, 2003. 275. Okazaki, M. M., Evenson, D. A., and Nadler, J. V.: Hippocampal mossy fiber sprouting and synapse formation after status epilepticus in rats: visualization after retrograde transport of biocytin. J. Comp. Neurol. 352:515, 1995. 276. Olofsson, B.: Rho guanine dissociation inhibitors: pivotal molecules in cellular signalling. Cell. Signal. 11:545, 1999. 277. O’Neill, G. M., Fashena, S. J., and Golemis, E. A.: Integrin signalling: a new Cas(t) of characters enters the stage. Trends Cell Biol. 10:111, 2000. 278. Ozeki, Y., Tomoda, T., Kleiderlein, J., et al.: Disrupted-inSchizophrenia-1 (DISC-1): mutant truncation prevents binding to NudE-like (NUDEL) and inhibits neurite outgrowth. Proc. Natl. Acad. Sci. U. S. A. 100:289, 2003. 279. Palmer, A., Zimmer, M., Erdmann, K. S., et al.: EphrinB phosphorylation and reverse signaling: regulation by Src kinases and PTP-BL phosphatase. Mol. Cell 9:725, 2002. 280. Parsons, J. T.: Focal adhesion kinase: the first ten years. J. Cell Sci. 116:1409, 2003. 281. Pasterkamp, R. J., and Kolodkin, A. L.: Semaphorin junction: making tracks toward neural connectivity. Curr. Opin. Neurobiol. 13:79, 2003.

478

Neurobiology of the Peripheral Nervous System

282. Pasterkamp, R. J., Peschon, J. J., Spriggs, M. K., and Kolodkin, A. L.: Semaphorin 7A promotes axon outgrowth through integrins and MAPKs. Nature 424:398, 2003. 283. Peck, J., Douglas, G. T., Wu, C. H., and Burbelo, P. D.: Human RhoGAP domain-containing proteins: structure, function and evolutionary relationships. FEBS Lett. 528:27, 2002. 284. Pellegrini, L.: Role of heparan sulfate in fibroblast growth factor signalling: a structural view. Curr. Opin. Struct. Biol. 11:629, 2001. 285. Perosa, S. R., Porcionatto, M. A., Cukiert, A., et al.: Glycosaminoglycan levels and proteoglycan expression are altered in the hippocampus of patients with mesial temporal lobe epilepsy. Brain Res. Bull. 58:509, 2002. 286. Perrot, V., Vazquez-Prado, J., and Gutkind, J. S.: Plexin B regulates Rho through the guanine nucleotide exchange factors leukemia-associated Rho GEF (LARG) and PDZRhoGEF. J. Biol. Chem. 277:43115, 2002. 287. Plump, A. S., Erskine, L., Sabatier, C., et al.: Slit1 and Slit2 cooperate to prevent premature midline crossing of retinal axons in the mouse visual system. Neuron 33:219, 2002. 288. Polleux, F., Morrow, T., and Ghosh, A.: Semaphorin 3A is a chemoattractant for cortical apical dendrites. Nature 404:567, 2000. 289. Poltorak, M., Khoja, I., Hemperly, J. J., et al.: Disturbances in cell recognition molecules (N-CAM and L1 antigen) in the CSF of patients with schizophrenia. Exp. Neurol. 131:266, 1995. 290. Poltorak, M., Wright, R., Hemperly, J. J., et al.: Monozygotic twins discordant for schizophrenia are discordant for N-CAM and L1 in CSF. Brain Res. 751:152, 1997. 291. Raftopoulou, M., and Hall, A.: Cell migration: Rho GTPases lead the way. Dev. Biol. 265:23, 2004. 292. Rajagopalan, S., Vivancos, V., Nicolas, E., and Dickson, B. J.: Selecting a longitudinal pathway: Robo receptors specify the lateral position of axons in the Drosophila CNS. Cell 103:1033, 2000. 293. Ralevic, V., and Burnstock, G.: Receptors for purines and pyrimidines. Pharmacol. Rev. 50:413, 1998. 294. Ramon y Cajal, S.: À quelle époque apparaissent les expansions des cellules nerveuses de la moëlle épinière du poulet? Anat. Anzeiger 21–22:609, 1890. 295. Raper, J. A.: Semaphorins and their receptors in vertebrates and invertebrates. Curr. Opin. Neurobiol. 10:88, 2000. 296. Reiner, O., Carrozzo, R., Shen, Y., et al.: Isolation of a Miller-Dieker lissencephaly gene containing G protein beta-subunit-like repeats. Nature 364:717, 1993. 297. Rhee, J., Mahfooz, N. S., Arregui, C., et al.: Activation of the repulsive receptor Roundabout inhibits N-cadherin-mediated cell adhesion. Nat. Cell Biol. 4:798, 2002. 298. Ribon, V., Herrera, R., Kay, B. K., and Saltiel, A. R.: A role for CAP, a novel, multifunctional Src homology 3 domaincontaining protein, in formation of actin stress fibers and focal adhesions. J. Biol. Chem. 273:4073, 1998. 299. Ridley, A. J., and Hall, A.: The small GTP-binding protein rho regulates the assembly of focal adhesions and actin stress fibers in response to growth factors. Cell 70:389, 1992. 300. Ridley, A. J., Paterson, H. F., Johnston, C. L., et al.: The small GTP-binding protein rac regulates growth factor-induced membrane ruffling. Cell 70:401, 1992.

301. Riviello, J. J.: Classification of seizures and epilepsy. Curr. Neurol. Neurosci. Rep. 3:325, 2003. 302. Rohatgi, R., Ma, L., Miki, H., et al.: The interaction between N-WASP and the Arp2/3 complex links Cdc42dependent signals to actin assembly. Cell 97:221, 1999. 303. Rohm, B., Ottemeyer, A., Lohrum, M., and Puschel, A. W.: Plexin/neuropilin complexes mediate repulsion by the axonal guidance signal semaphorin 3A. Mech. Dev. 93:95, 2000. 304. Rohm, B., Rahim, B., Kleiber, B., et al.: The semaphorin 3A receptor may directly regulate the activity of small GTPases. FEBS Lett. 486:68, 2000. 305. Ronca, F., Andersen, J. S., Paech, V., and Margolis, R. U.: Characterization of Slit protein interactions with glypican-1. J. Biol. Chem. 276:29141, 2001. 306. Ross, M. E., and Walsh, C. A.: Human brain malformations and their lessons for neuronal migration. Annu. Rev. Neurosci. 24:1041, 2001. 307. Rothberg, J. M., and Artavanis-Tsakonas, S.: Modularity of the slit protein: characterization of a conserved carboxyterminal sequence in secreted proteins and a motif implicated in extracellular protein interactions. J. Mol. Biol. 227:367, 1992. 308. Rothberg, J. M., Hartley, D. A., Walther, Z., and ArtavanisTsakonas, S.: Slit: an EGF-homologous locus of D. melanogaster involved in the development of the embryonic central nervous system. Cell 55:1047, 1988. 309. Rothberg, J. M., Jacobs, J. R., Goodman, C. S., and Artavanis-Tsakonas, S.: Slit: an extracellular protein necessary for development of midline glia and commissural axon pathways contains both EGF and LRR domains. Genes Dev. 4:2169, 1990. 310. Rougon, G., and Hobert, O.: New insights into the diversity and function of neuronal immunoglobulin superfamily molecules. Annu. Rev. Neurosci. 26:207, 2003. 311. Rudolph, M. G., Bayer, P., Abo, A., et al.: The Cdc42/Rac interactive binding region motif of the Wiskott Aldrich syndrome protein (WASP) is necessary but not sufficient for tight binding to Cdc42 and structure formation. J. Biol. Chem. 273:18067, 1998. 312. Rugarli, E. I., Di Schiavi, E., Hilliard, M. A., et al.: The Kallmann syndrome gene homolog in C. elegans is involved in epidermal morphogenesis and neurite branching. Development 129:1283, 2002. 313. Safer, D., Elzinga, M., and Nachmias, V. T.: Thymosin beta 4 and Fx, an actin-sequestering peptide, are indistinguishable. J. Biol. Chem. 266:4029, 1991. 314. Saffell, J. L., Williams, E. J., Mason, I. J., et al.: Expression of a dominant negative FGF receptor inhibits axonal growth and FGF receptor phosphorylation stimulated by CAMs. Neuron 18:231, 1997. 315. Sagot, I., Klee, S. K., and Pellman, D.: Yeast formins regulate cell polarity by controlling the assembly of actin cables. Nat. Cell Biol. 4:42, 2002. 316. Sagot, I., Rodal, A. A., Moseley, J., et al.: An actin nucleation mechanism mediated by Bni1 and profilin. Nat. Cell Biol. 4:626, 2002. 317. Sahai, E., Olson, M. F., and Marshall, C. J.: Cross-talk between Ras and Rho signalling pathways in transformation favours proliferation and increased motility. EMBO J. 20:755, 2001.

Guidance of Axons to Targets in Development and in Disease 318. Salazar, M. A., Kwiatkowski, A. V., Pellegrini, L., et al.: Tuba, a novel protein containing bin/amphiphysin/Rvs and Dbl homology domains, links dynamin to regulation of the actin cytoskeleton. J. Biol. Chem. 278:49031, 2003. 319. Sander, E. E., ten Klooster, J. P., van Delft, S., et al.: Rac downregulates Rho activity: reciprocal balance between both GTPases determines cellular morphology and migratory behavior. J. Cell Biol. 147:1009, 1999. 320. Sanders, L. C., Matsumura, F., Bokoch, G. M., and de Lanerolle, P.: Inhibition of myosin light chain kinase by p21-activated kinase. Science 283:2083, 1999. 321. Sarmiere, P. D., and Bamburg, J. R.: Regulation of the neuronal actin cytoskeleton by ADF/cofilin. J. Neurobiol. 58:103, 2004. 322. Sasaki, S., Shionoya, A., Ishida, M., et al.: A LIS1/NUDEL/ cytoplasmic dynein heavy chain complex in the developing and adult nervous system. Neuron 28:681, 2000. 323. Sasaki, Y., Cheng, C., Uchida, Y., et al.: Fyn and Cdk5 mediate semaphorin-3A signaling, which is involved in regulation of dendrite orientation in cerebral cortex. Neuron 35:907, 2002. 324. Sattler, M., Salgia, R., Shrikhande, G., et al.: Differential signaling after beta1 integrin ligation is mediated through binding of CRKL to p120(CBL) and p110(HEF1). J. Biol. Chem. 272:14320, 1997. 325. Sayas, C. L., Avila, J., and Wandosell, F.: Glycogen synthase kinase-3 is activated in neuronal cells by Galpha12 and Galpha13 by Rho-independent and Rho-dependent mechanisms. J. Neurosci. 22:6863, 2002. 326. Schaefer, A. W., Kabir, N., and Forscher, P.: Filopodia and actin arcs guide the assembly and transport of two populations of microtubules with unique dynamic parameters in neuronal growth cones. J. Cell Biol. 158:139, 2002. 327. Schimmelpfeng, K., Gogel, S., and Klambt, C.: The function of leak and kuzbanian during growth cone and cell migration. Mech. Dev. 106:25, 2001. 328. Schmidt, A., and Hall, A.: Guanine nucleotide exchange factors for Rho GTPases: turning on the switch. Genes Dev. 16:1587, 2002. 329. Schoenwaelder, S. M., Petch, L. A., Williamson, D., et al.: The protein tyrosine phosphatase Shp-2 regulates RhoA activity. Curr. Biol. 10:1523, 2000. 330. Schwanzel-Fukuda, M., Bick, D., and Pfaff, D. W.: Luteinizing hormone-releasing hormone (LHRH)-expressing cells do not migrate normally in an inherited hypogonadal (Kallmann) syndrome. Brain Res. Mol. Brain Res. 6:311, 1989. 331. Schwanzel-Fukuda, M., and Pfaff, D. W.: Origin of luteinizing hormone-releasing hormone neurons. Nature 338:161, 1989. 332. Seeger, M., Tear, G., Ferres-Marco, D., and Goodman, C. S.: Mutations affecting growth cone guidance in Drosophila: genes necessary for guidance toward or away from the midline. Neuron 10:409, 1993. 333. Semaphorin Nomenclature Committee: Unified nomenclature for the semaphorins/collapsins. Semaphorin Nomenclature Committee. Cell 97:551, 1999. 334. Serafini, T., Colamarino, S. A., Leonardo, E. D., et al.: Netrin-1 is required for commissural axon guidance in the developing vertebrate nervous system. Cell 87:1001, 1996.

479

335. Serafini, T., Kennedy, T. E., Galko, M. J., et al.: The netrins define a family of axon outgrowth-promoting proteins homologous to C. elegans UNC-6. Cell 78:409, 1994. 336. Shamah, S. M., Lin, M. Z., Goldberg, J. L., et al.: EphA receptors regulate growth cone dynamics through the novel guanine nucleotide exchange factor ephexin. Cell 105:233, 2001. 337. Shan, W., Yoshida, M., Wu, X. R., et al.: Neural (N-) cadherin, a synaptic adhesion molecule, is induced in hippocampal mossy fiber axonal sprouts by seizure. J. Neurosci. Res. 69:292, 2002. 338. Shi, Y., Alin, K., and Goff, S. P.: Abl-interactor-1, a novel SH3 protein binding to the carboxy-terminal portion of the Abl protein, suppresses v-abl transforming activity. Genes Dev. 9:2583, 1995. 339. Shirasaki, R., Katsumata, R., and Murakami, F.: Change in chemoattractant responsiveness of developing axons at an intermediate target. Science 279:105, 1998. 340. Shneker, B. F., and Fountain, N. B.: Epilepsy. Dis. Mon. 49:426, 2003. 341. Simpson, J. H., Kidd, T., Bland, K. S., and Goodman, C. S.: Short-range and long-range guidance by slit and its Robo receptors: Robo and Robo2 play distinct roles in midline guidance. Neuron 28:753, 2000. 342. Skaper, S. D., Moore, S. E., and Walsh, F. S.: Cell signalling cascades regulating neuronal growth-promoting and inhibitory cues. Prog. Neurobiol. 65:593, 2001. 343. Sloviter, R. S.: The functional organization of the hippocampal dentate gyrus and its relevance to the pathogenesis of temporal lobe epilepsy. Ann. Neurol. 35:640, 1994. 344. Snow, P. M., Bieber, A. J., and Goodman, C. S.: Fasciclin III: a novel homophilic adhesion molecule in Drosophila. Cell 59:313, 1989. 345. Snyder, S. E., Li, J., Schauwecker, P. E., et al.: Comparison of RPTP zeta/beta, phosphacan, and trkB mRNA expression in the developing and adult rat nervous system and induction of RPTP zeta/beta and phosphacan mRNA following brain injury. Brain Res. Mol. Brain Res. 40:79, 1996. 346. Sonenberg, N., and Gingras, A. C.: The mRNA 5⬘ capbinding protein eIF4E and control of cell growth. Curr. Opin. Cell Biol. 10:268, 1998. 347. Song, H., Ming, G., He, Z., et al.: Conversion of neuronal growth cone responses from repulsion to attraction by cyclic nucleotides. Science 281:1515, 1998. 348. Song, H. J., Ming, G. L., and Poo, M. M.: cAMP-induced switching in turning direction of nerve growth cones. Nature 388:275, 1997. 349. Soussi-Yanicostas, N., de Castro, F., Julliard, A. K., et al.: Anosmin-1, defective in the X-linked form of Kallmann syndrome, promotes axonal branch formation from olfactory bulb output neurons. Cell 109:217, 2002. 350. Soussi-Yanicostas, N., Hardelin, J. P., Arroyo-Jimenez, M. M., et al.: Initial characterization of anosmin-1, a putative extracellular matrix protein synthesized by definite neuronal cell populations in the central nervous system. J. Cell Sci. 109(Pt. 7):1749, 1996. 351. Speidel, C. C.: Studies of living nerves II. Activities of ameboid growth cones, sheath cells, and myelin segments, as revealed by prolonged observation of individual nerve fibers in frog tadpoles. Am. J. Anat. 52:1, 1933.

480

Neurobiology of the Peripheral Nervous System

352. Sperry, R. W.: Chemoaffinity in the orderly growth of nerve fiber patterns and connections. Proc. Natl. Acad. Sci. U. S. A. 50:703, 1963. 353. Steigemann, P., Molitor, A., Fellert, S., et al.: Heparan sulfate proteoglycan syndecan promotes axonal and myotube guidance by slit/robo signaling. Curr. Biol. 14:225, 2004. 354. Stein, E., and Tessier-Lavigne, M.: Hierarchical organization of guidance receptors: silencing of netrin attraction by slit through a Robo/DCC receptor complex. Science 291:1928, 2001. 355. Stein, E., Zou, Y., Poo, M., and Tessier-Lavigne, M.: Binding of DCC by netrin-1 to mediate axon guidance independent of adenosine A2B receptor activation. Science 291:1976, 2001. 356. Stoeckli, E. T., Kuhn, T. B., Duc, C. O., et al.: The axonally secreted protein axonin-1 is a potent substratum for neurite growth. J. Cell Biol. 112:449, 1991. 357. Suetsugu, S., Miki, H., and Takenawa, T.: The essential role of profilin in the assembly of actin for microspike formation. EMBO J. 17:6516, 1998. 358. Sumi, T., Matsumoto, K., Takai, Y., and Nakamura, T.: Cofilin phosphorylation and actin cytoskeletal dynamics regulated by rho- and Cdc42-activated LIM-kinase 2. J. Cell Biol. 147:1519, 1999. 359. Sun, Q., Bahri, S., Schmid, A., et al.: Receptor tyrosine phosphatases regulate axon guidance across the midline of the Drosophila embryo. Development 127:801, 2000. 360. Sundaresan, V., Mambetisaeva, E., Andrews, W., et al.: Dynamic expression patterns of Robo (Robo1 and Robo2) in the developing murine central nervous system. J. Comp. Neurol. 468:467, 2004. 361. Suter, D. M., Errante, L. D., Belotserkovsky, V., and Forscher, P.: The Ig superfamily cell adhesion molecule, apCAM, mediates growth cone steering by substratecytoskeletal coupling. J. Cell Biol. 141:227, 1998. 362. Suzuki, T., Nakamoto, T., Ogawa, S., et al.: MICAL, a novel CasL interacting molecule, associates with vimentin. J. Biol. Chem. 277:14933, 2002. 363. Swiercz, J. M., Kuner, R., Behrens, J., and Offermanns, S.: Plexin-B1 directly interacts with PDZ-RhoGEF/LARG to regulate RhoA and growth cone morphology. Neuron 35:51, 2002. 364. Takahashi, T., Fournier, A., Nakamura, F., et al.: Plexinneuropilin-1 complexes form functional semaphorin-3A receptors. Cell 99:59, 1999. 365. Tamagnone, L., Artigiani, S., Chen, H., et al.: Plexins are a large family of receptors for transmembrane, secreted, and GPI-anchored semaphorins in vertebrates. Cell 99:71, 1999. 366. Tamagnone, L., and Comoglio, P. M.: Signalling by semaphorin receptors: cell guidance and beyond. Trends Cell Biol. 10:377, 2000. 367. Tani, K., Sato, S., Sukezane, T., et al.: Abl interactor 1 promotes tyrosine 296 phosphorylation of mammalian enabled (Mena) by c-Abl kinase. J. Biol. Chem. 278:21685, 2003. 368. Tauck, D. L., and Nadler, J. V.: Evidence of functional mossy fiber sprouting in hippocampal formation of kainic acid-treated rats. J. Neurosci. 5:1016, 1985. 369. Tear, G., Harris, R., Sutaria, S., et al.: Commissureless controls growth cone guidance across the CNS midline in Drosophila and encodes a novel membrane protein. Neuron 16:501, 1996.

370. Terman, J. R., Mao, T., Pasterkamp, R. J., et al.: MICALs, a family of conserved flavoprotein oxidoreductases, function in plexin-mediated axonal repulsion. Cell 109:887, 2002. 371. Tessier-Lavigne, M., Placzek, M., Lumsden, A. G., et al.: Chemotropic guidance of developing axons in the mammalian central nervous system. Nature 336:775, 1988. 372. Tomek, W., Melo Sterza, F. A., Kubelka, M., et al.: Regulation of translation during in vitro maturation of bovine oocytes: the role of MAP kinase, eIF4E (cap binding protein) phosphorylation, and eIF4E-BP1. Biol. Reprod. 66:1274, 2002. 373. Tong, J., Killeen, M., Steven, R., et al.: Netrin stimulates tyrosine phosphorylation of the UNC-5 family of netrin receptors and induces Shp2 binding to the RCM cytodomain. J. Biol. Chem. 276:40917, 2001. 374. Torres, R., Firestein, B. L., Dong, H., et al.: PDZ proteins bind, cluster, and synaptically colocalize with Eph receptors and their ephrin ligands. Neuron 21:1453, 1998. 375. Uchida, N., Honjo, Y., Johnson, K. R., et al.: The catenin/cadherin adhesion system is localized in synaptic junctions bordering transmitter release zones. J. Cell Biol. 135:767, 1996. 376. van de Wetering, M., de Lau, W., and Clevers, H.: WNT signaling and lymphocyte development. Cell 109(Suppl.):S13, 2002. 377. Vawter, M. P., Usen, N., Thatcher, L., et al.: Characterization of human cleaved N-CAM and association with schizophrenia. Exp Neurol 172:29, 2001. 378. Vicente, A. M., Macciardi, F., Verga, M., et al.: NCAM and schizophrenia: genetic studies. Mol Psychiatry 2:65, 1997. 379. Vielmetter, J., Kayyem, J. F., Roman, J. M., and Dreyer, W. J.: Neogenin, an avian cell surface protein expressed during terminal neuronal differentiation, is closely related to the human tumor suppressor molecule deleted in colorectal cancer. J Cell Biol 127:2009, 1994. 380. Vikis, H. G., Li, W., and Guan, K. L.: The plexin-B1/Rac interaction inhibits PAK activation and enhances Sema4D ligand binding. Genes Dev 16:836, 2002. 381. Vikis, H. G., Li, W., He, Z., and Guan, K. L.: The semaphorin receptor plexin-B1 specifically interacts with active Rac in a ligand-dependent manner. Proc. Natl. Acad. Sci. U. S. A. 97:12457, 2000. 382. Vindis, C., Cerretti, D. P., Daniel, T. O., and Huynh-Do, U.: EphB1 recruits c-Src and p52Shc to activate MAPK/ERK and promote chemotaxis. J. Cell Biol. 162:661, 2003. 383. Vojtek, A., Haarer, B., Field, J., et al.: Evidence for a functional link between profilin and CAP in the yeast S. cerevisiae. Cell 66:497, 1991. 384. Waddington, J. L., Lane, A., Scully, P. J., et al.: Neurodevelopmental and neuroprogressive processes in schizophrenia: antithetical or complementary, over a lifetime trajectory of disease? Psychiatr. Clin. North Am. 21:123, 1998. 385. Wahl, S., Barth, H., Ciossek, T., et al.: Ephrin-A5 induces collapse of growth cones by activating Rho and Rho kinase. J. Cell Biol. 149:263, 2000. 386. Wang, K. H., Brose, K., Arnott, D., et al.: Biochemical purification of a mammalian slit protein as a positive regulator of sensory axon elongation and branching. Cell 96:771, 1999. 387. Wang, Y., Thekdi, N., Smallwood, P. M., et al.: Frizzled-3 is required for the development of major fiber tracts in the rostral CNS. J. Neurosci. 22:8563, 2002.

Guidance of Axons to Targets in Development and in Disease 388. Waskiewicz, A. J., Flynn, A., Proud, C. G., and Cooper, J. A.: Mitogen-activated protein kinases activate the serine/threonine kinases Mnk1 and Mnk2. EMBO J. 16:1909, 1997. 389. Watanabe, N., Madaule, P., Reid, T., et al.: p140mDia, a mammalian homolog of Drosophila diaphanous, is a target protein for Rho small GTPase and is a ligand for profilin. EMBO J. 16:3044, 1997. 390. Waterman-Storer, C. M., and Salmon, E. D.: Actomyosinbased retrograde flow of microtubules in the lamella of migrating epithelial cells influences microtubule dynamic instability and turnover and is associated with microtubule breakage and treadmilling. J. Cell Biol. 139:417, 1997. 391. Welch, M. D., Iwamatsu, A., and Mitchison, T. J.: Actin polymerization is induced by Arp2/3 protein complex at the surface of Listeria monocytogenes. Nature 385:265, 1997. 392. Wennerberg, K., Forget, M. A., Ellerbroek, S. M., et al.: Rnd proteins function as RhoA antagonists by activating p190 RhoGAP. Curr. Biol. 13:1106, 2003. 393. Wenzel, H. J., Woolley, C. S., Robbins, C. A., and Schwartzkroin, P. A.: Kainic acid-induced mossy fiber sprouting and synapse formation in the dentate gyrus of rats. Hippocampus 10:244, 2000. 394. Whitford, K. L., Marillat, V., Stein, E., et al.: Regulation of cortical dendrite development by Slit-Robo interactions. Neuron 33:47, 2002. 395. Wilkinson, D. G.: Multiple roles of EPH receptors and ephrins in neural development. Nat. Rev. Neurosci. 2:155, 2001. 396. Williams, E. J., Furness, J., Walsh, F. S., and Doherty, P.: Activation of the FGF receptor underlies neurite outgrowth stimulated by L1, N-CAM, and N-cadherin. Neuron 13:583, 1994. 397. Wills, Z., Bateman, J., Korey, C. A., et al.: The tyrosine kinase Abl and its substrate enabled collaborate with the receptor phosphatase Dlar to control motor axon guidance. Neuron 22:301, 1999. 398. Wills, Z., Emerson, M., Rusch, J., et al.: A Drosophila homolog of cyclase-associated proteins collaborates with the Abl tyrosine kinase to control midline axon pathfinding. Neuron 36:611, 2002. 399. Winberg, M. L., Noordermeer, J. N., Tamagnone, L., et al.: Plexin A is a neuronal semaphorin receptor that controls axon guidance. Cell 95:903, 1998. 400. Winberg, M. L., Tamagnone, L., Bai, J., et al.: The transmembrane protein Off-track associates with Plexins and functions downstream of Semaphorin signaling during axon guidance. Neuron 32:53, 2001. 401. Witke, W., Podtelejnikov, A. V., Di Nardo, A., et al.: In mouse brain profilin I and profilin II associate with regulators of the endocytic pathway and actin assembly. EMBO J. 17:967, 1998. 402. Wodarz, A., and Nusse, R.: Mechanisms of Wnt signaling in development. Annu. Rev. Cell Dev. Biol. 14:59, 1998. 403. Wong, E. V., Schaefer, A. W., Landreth, G., and Lemmon, V.: Casein kinase II phosphorylates the neural cell adhesion molecule L1. J. Neurochem. 66:779, 1996. 404. Wong, E. V., Schaefer, A. W., Landreth, G., and Lemmon, V.: Involvement of p90rsk in neurite outgrowth mediated by the cell adhesion molecule L1. J. Biol. Chem. 271:18217, 1996.

481

405. Wong, K., Ren, X. R., Huang, Y. Z., et al.: Signal transduction in neuronal migration: roles of GTPase activating proteins and the small GTPase Cdc42 in the Slit-Robo pathway. Cell 107:209, 2001. 406. World Health Organization: The ICD-10 Classification of Mental and Behavioural Disorders: Clinical Descriptions and Diagnostic Guidelines. Geneva, World Health Organization, 1992. 407. Wybenga-Groot, L. E., Baskin, B., Ong, S. H., et al.: Structural basis for autoinhibition of the Ephb2 receptor tyrosine kinase by the unphosphorylated juxtamembrane region. Cell 106:745, 2001. 408. Xu, X. M., Fisher, D. A., Zhou, L., et al.: The transmembrane protein semaphorin 6A repels embryonic sympathetic axons. J. Neurosci. 20:2638, 2000. 409. Yamasaki, M., Thompson, P., and Lemmon, V.: CRASH syndrome: mutations in L1CAM correlate with severity of the disease. Neuropediatrics 28:175, 1997. 410. Yang, J., Si, T., Ling, Y., et al.: Association study of the human FZD3 locus with schizophrenia. Biol. Psychiatry 54:1298, 2003. 411. Yang, N., Higuchi, O., Ohashi, K., et al.: Cofilin phosphorylation by LIM-kinase 1 and its role in Rac-mediated actin reorganization. Nature 393:809, 1998. 412. Yang-Snyder, J., Miller, J. R., Brown, J. D., et al.: A frizzled homolog functions in a vertebrate Wnt signaling pathway. Curr. Biol. 6:1302, 1996. 413. Yu, D. H., Qu, C. K., Henegariu, O., et al.: Protein-tyrosine phosphatase Shp-2 regulates cell spreading, migration, and focal adhesion. J. Biol. Chem. 273:21125, 1998. 414. Yu, H. H., Zisch, A. H., Dodelet, V. C., and Pasquale, E. B.: Multiple signaling interactions of Abl and Arg kinases with the EphB2 receptor. Oncogene 20:3995, 2001. 415. Yu, T. W., Hao, J. C., Lim, W., et al.: Shared receptors in axon guidance: SAX-3/Robo signals via UNC-34/Enabled and a Netrin-independent UNC-40/DCC function. Nat. Neurosci. 5:1147, 2002. 416. Zanata, S. M., Hovatta, I., Rohm, B., and Puschel, A. W.: Antagonistic effects of Rnd1 and RhoD GTPases regulate receptor activity in Semaphorin 3A-induced cytoskeletal collapse. J. Neurosci. 22:471, 2002. 417. Zeng, L., Fagotto, F., Zhang, T., et al.: The mouse Fused locus encodes Axin, an inhibitor of the Wnt signaling pathway that regulates embryonic axis formation. Cell 90:181, 1997. 418. Zhang, X. F., Schaefer, A. W., Burnette, D. T., et al.: Rhodependent contractile responses in the neuronal growth cone are independent of classical peripheral retrograde actin flow. Neuron 40:931, 2003. 419. Zhou, L., White, F. A., Lentz, S. I., et al.: Cloning and expression of a novel murine semaphorin with structural similarity to insect semaphorin I. Mol. Cell. Neurosci. 9:26, 1997. 420. Zimmer, M., Palmer, A., Kohler, J., and Klein, R.: EphBephrinB bi-directional endocytosis terminates adhesion allowing contact mediated repulsion. Nat. Cell Biol. 5:869, 2003. 421. Zisch, A. H., Stallcup, W. B., Chong, L. D., et al.: Tyrosine phosphorylation of L1 family adhesion molecules: implication of the Eph kinase Cek5. J. Neurosci. Res. 47:655, 1997.

22 Aging in the Peripheral Nervous System T. COWEN, B. ULFHAKE, AND R. H. M. KING

Nature of Aging in the PNS Primary Sensory Neurons Nerve Trunks Autonomic Neurons Mechanisms of Aging in the PNS

Mechanisms Underlying Sensory Impairments during Aging Mechanisms Underlying Age-Related Changes in Nerve Trunks

A characteristic of aging in the peripheral nervous system (PNS), as in other areas of the nervous system, is the locally specific nature of the changes that occur. ‘Selective vulnerability’ in aging, as this phenomenon has been called, seems to characterize the effects of aging in all areas of the PNS, to a greater or lesser extent. Long, myelinated nerve fibers are more vulnerable to age-related change compared with shorter fibers. Large sensory neurons appear to be more vulnerable than small ones. Sympathetic neurons supplying some target tissues are more vulnerable than those supplying others. Enteric neurons are particularly vulnerable to age-related neuronal cell death in which diet plays an important role. All these groups of neurons undergo dying back of peripheral axon terminals and also probably of fine dendritic branches to some extent. Interestingly, in peripheral as in central neurons, the definitive demonstration of neuron loss during aging is relatively rare, at least in the mammalian nervous system, whereas neuronal atrophy and dying back of fine terminal axonal and dendritic fibers are much more common. In at least some of these instances, disturbance of trophic interactions between peripheral nerve fibers and their target sensory receptors or effectors is thought to contribute to age-related selective vulnerability.

Mechanisms Underlying Age-Related Impairments in Autonomic Neurons Conclusions

NATURE OF AGING IN THE PNS Primary Sensory Neurons Functional Impairments in Aging Sensory Neurons Disturbances of gait cycle and balance control as well as increased thresholds for exteroceptive, proprioceptive, and vibratory sensations are some of the most common problems seen in the elderly.21,30,51,54,68,71,84,104,127,137,144,177,183,212,222 These impairments contribute to characteristic traits of aging such as an increased frequency of falls and fractures, impediment of the activities of daily life, and loss of independence.136 Falls and their complications constitute a major cause of hospitalization in senescent individuals, and a significant correlation has been found between the occurrence of falls in the elderly and decreased proprioceptive function.30 Structural Impairments in Aging Sensory Neurons Age-related changes are more conspicuous in myelinated than in unmyelinated axons, and there appears to be a good agreement between the severity of sensory nerve lesions and functional sensory deficits. Numbers of myelinated 483

484

Neurobiology of the Peripheral Nervous System

sensory axons are either unaltered or moderately decreased in peripheral nerves and spinal roots of aged subjects.126,129,157,180,187,194,197,213 It is at present unclear whether unmyelinated sensory axons are affected to the same extent. Consistent with a biased distribution of sensory deficits, lesions in both humans (bipedal) and rodents (quadripedal) are more widespread in lumbar than in cervical nerves.17,32,182,212,213,222 Despite relatively restricted losses of preterminal sensory nerve fibers and sensory neurons, there are numerous and selective losses of sensory receptors and nerve endings in the different target tissues of aged individuals. A major specific effect of aging on tactile skin receptors has been demonstrated in both rodents and humans. For example, human finger and footpads show a pronounced decline in the number of pacinian corpuscles, Meissner corpuscles, and Merkel cell–neurite complexes in senescence.26,34,36,192,225 Consistent with these morphologic changes, Gescheider and collaborators84 found that the thresholds for vibrotactile stimuli, mediated by pacinian, Meissner, and Ruffini corpuscles and Merkel endings, increased with age, with the threshold of pacinian corpuscles the most affected. Moreover, tests of tactile responses in the finger pulp194 indicate that the distal axon, including the endings, seems to be more severely affected than the proximal axon. The receptors most vulnerable in aging are therefore primarily mechanoreceptors in skin and muscle whose axons are the large myelinated A␣ and A␤ fibers. These peripheral fibers originate in the large, pale neurons of the dorsal root ganglia (DRG)140 and terminate either in laminae III through VI or, for some muscle afferents, in laminae VII and IX of the deep dorsal horn of the spinal cord.224 Taken together, it seems as if the decrease in tactile sensitivity with advancing age may be explained by a decreased receptor density resulting from a degenerative process that probably involves mainly the peripheral receptor/nerve ending rather than more proximal regions of the parent sensory neuron. In addition, the possibility cannot be excluded that conventional morphologic techniques are less sensitive than functional tests in detecting age-related aberrations in mechanoreception. Further evidence of functional impairments of aging sensory neurons comes from studies showing that conduction velocity declines, albeit at a slower rate and more uniformly than in motor nerve fibers (see Functional Impairments in Axons below). In the median nerve, sensory axon conduction velocity declines at a rate of 2 m/s per decade, whereas in the ulnar nerve the decrease is 1m/s per decade between 20 and 55 years of age, increasing to 3 m/s per decade after 55 years of age.31 In addition to altered conduction velocity, the amplitude of nerve action potentials in sensory nerves from humans at 70 years of age has declined to half the value at 20 years. Whereas data on the effects of age in peripheral sensory nerves are abundant, much less is known about how these

processes affect the central terminations of primary sensory neurons.20 Taking advantage of tracer substances with differing affinities for different classes of primary sensory neurons, evidence was obtained for a dramatic and selective decrease in the density of myelinated, but not unmyelinated, primary afferents terminating in the dorsal horn of aged rodents (Fig. 22–1).22 The substantial reduction in mechanoreceptive input to the central nervous system (CNS) includes muscle spindle and Golgi tendon organ afferents terminating in the deep dorsal horn and the spinal motor nucleus. Furthermore, the extent of the loss of mechanosensory input to the spinal cord agreed with the degree as well as the distribution of sensorimotor impairment of the animals. Although the major causes of age-related behavioral changes remain to be determined, loss of neurons has often been considered to be a prime cause of functional deficits during aging. However, evidence from a growing number of studies shows that cell loss is far less prominent than previously thought.4,39,43,145,159,191,223 Unbiased estimates of primary sensory neurons in rodents show only a small decrease (⬍15%) of both large neurons with myelinated processes and small neurons with unmyelinated processes at advanced age. More importantly, there is no clear association between the degree of cell loss and sensory deficits among the individual animals (Fig. 22–2).21 Earlier studies of numbers of DRG neurons are contradictory, with some showing a notable loss of neurons during aging,78,161 whereas others report no change.135,169 Part of the inconsistency in these studies can probably be attributed to the fact that they were performed using indirect counting techniques based on two-dimensional probes, thereby producing biased results.38 Signaling Molecules in Aging Sensory Neurons The loss of epidermal and dermal innervation involves both sensory and autonomic components (see sections on Myelinated Axons and on Selective Vulnerability of Autonomic Neurons to Age-Related Neurodegeneration below). Aminergic, cholinergic, and peptidergic nerve fibers are decreased in number, with a decreased innervation of blood vessels, subepidermis, and sweat glands.2,45,48,59,75,155,164 Impaired innervation of sweat glands and skin blood vessels will have a negative impact on wound healing and body temperature control118 (see below). In this context it is important to consider that the absence of signaling molecules such as neuropeptides in peripheral nerves does not necessarily reflect loss of the axons. At the cellular level, primary sensory neurons show a complex pattern of changes during aging. Besides the axon aberrations and the small unselective loss of neurons (see above), there is evidence for a selective cell body atrophy (large pale cell bodies)21,171,180 and phenotypic alterations in the expression pattern of signaling molecules, receptors, and cytoskeletal proteins (reviewed by Ulfhake et al.211). Primary sensory neurons

Aging in the Peripheral Nervous System

FIGURE 22–1 Images showing cholera toxin B subunit (CTB) labeling in the lower lumbar spinal cord dorsal horns (A and B), ventral horn (C and D), and dorsal root ganglia (E and F) 3 days after injection of CTB into the tibial nerve of young adult (Ad; A, C, and E) or aged (Ag; B, D, and F) rats. Tibial nerve CTB injections in young adult rats resulted in a moderate to dense labeling of myelinated primary afferent axons terminating in the medial part of the deep dorsal horn (laminae III through VI; A) and in the ventral horn, including its motor nucleus (VII and IX; C) as well as in motoneurons (C). Aged rats disclosed a dramatic decrease in CTB labeling both in the deep dorsal horn (B) and in the ventral horn (D), whereas the labeling appeared unaffected in motoneurons (D). CTB labeling in the corresponding dorsal root ganglia was confined mainly to large neurons (E and F), with no difference observed between young adult and aged individuals. Insets in C and D show motoneurons at higher magnification. Scale bar: 150 ␮m (A–D); 75 ␮m (E and F); 50 ␮m (insets C and D). (From Bergman, E., and Ulfhake, B.: Evidence for loss of myelinated input to the spinal cord in senescent rats. Neurobiol. Aging 23:271, 2002.)

485

486

Neurobiology of the Peripheral Nervous System

20 103

Total number of DRG neurons

18 103

16 103 L4

14 103

12 103

C5

10 103

8 103

Young adult

Stage I

Stage II

Stage III

FIGURE 22–2 Graphic representation of the relation between the total number of neurons in the fifth cervical (C5; below hatched line) and fourth lumbar (L4; above hatched line) dorsal root ganglia (DRG), age (young adult: open symbols; aged: filled symbols), and stage of sensory impairment (stages I through III, in which stage I represents symptoms of a mild degree of deficits and stage III corresponds to symptoms of advanced deficits). Male and female rats are indicated with squares and circles, respectively. Note the lack of correlation between neuron loss and symptoms among the aged rats at both DRG levels studied. (From Bergman, E., and Ulfhake, B.: Loss of primary sensory neurons in the very old rat: neuron number estimates using the disector method and confocal optical sectioning. J. Comp. Neurol. 396:211, 1998.)

synthesize neuropeptides, which act as transmitters and/or modulators of sensory transmission at the central termination of these neurons.102,139,224 Consistent with the decrement of neuropeptides in peripheral sensory nerves and endings, there is a decreased expression of both calcitonin gene–related peptide and substance P in aged DRG neurons.18,75 In addition to their central effects, these neuropeptides exert effects in the periphery, where they are thought to be involved in inflammatory responses and wound healing.94,103,119,165 The lower level of neuropeptide expression in subsets of sensory neurons during aging provides a molecular and cellular substrate for the previously described diminished axon reflex in senescence.98,120

Nerve Trunks Functional Impairments in Axons As previously stated (see Functional Impairments in Aging Sensory Neurons above), the main clinical signs of functional impairment of peripheral nerves in the elderly are

loss of ankle jerks and impaired vibration sense starting in the feet, combined with pain, temperature, or light touch impairment in 25% of patients between 70 and 85 years of age.51,104 The pain and touch impairments were ascribed to peripheral nerve degeneration.51,104 Although not all the earlier studies excluded impaired glucose tolerance, several reliable studies have concluded that there is a significant decline in the threshold of vibratory sensation in healthy people after the age of 50 years.88,173,203 This correlates with evidence of loss of mechanoreceptors (see Structural Impairments in Aging Sensory Neurons above) and axonal loss affecting predominantly large myelinated fibers (including those involved in mechanoreception) in older people over the age of 60 years, together with abnormalities in the vasa nervorum.207 However, age-related functional deficits in peripheral nerves are not a consistent observation, and differentiating clinically between the effects of age on central versus peripheral neurons and on neurons versus the muscles or receptors they innervate is difficult. Thus a recent, exhaustive study of 200 people with a mean age of 80 years found only 3 people with absent ankle jerks that could not be explained by disease, trauma, or other factors.105 This compares well with a study on elderly patients with chronic inflammatory demyelinating polyneuropathy in whom only 2% of the normal controls lacked sural nerve action potentials.219 Effects on motor nerves are more marked, as previously stated, with conduction velocity declining with age by up to 30% in humans and animals.167,220,221 In sensory nerves, the decrease in conduction velocity is less than in motor nerves, although recording techniques do not allow reliable distinction between responses in large myelinated sensory and motor axons (see sections on Myelinated Axons and on Unmyelinated Axons below). In another study, sural responses were unobtainable in 10% of subjects age 80 and over, indicating the lack of consistency between different studies of similar nerve populations. Despite their shorter lifespan, experimental animals also show an increasing number of abnormalities in their peripheral nerves with advancing age.9,28,29,126,131,205 Electrophysiological studies in animals indicate a functional decline similar to that seen in humans. For example, a study in old cats showed a decrease in motor nerve conduction velocity with age and myelin abnormalities.3 Although conduction velocity in the nerve to the flexor hallucis longus muscle of aged rats was reported to be unchanged,221 it was reduced in the muscle spindles.158 These fibers are among the largest, indicating that the size of the fibers may be important in determining which nerves are vulnerable to aging, rather than whether they are motor or sensory. Structural Impairments in Axons Examination of biopsy or necropsy specimens helps to define the morphological changes that underlie functional impairments. Few studies have examined all the different

Aging in the Peripheral Nervous System

parts of the PNS, including primary sensory neurons, peripheral motor nerve fibers, and autonomic neurons. Despite the difficulties associated with obtaining wellpreserved human material for light and electron microscopy, several earlier as well as more recent studies in humans, and also in experimental animals, have consistently described age-related axonal loss in spinal roots and peripheral nerves.65,168 There are indications that the loss of axons does not affect all areas to the same extent; a preferential loss of nerve fibers in the hind limbs of old rats has been demonstrated compared with those supplying the forelimbs.97 Neurogenic atrophy in human skeletal muscle has been reported as more marked in muscles of the legs than in the arms, with intermediate findings in the trunk.208 Both observations suggest that the larger and longer axons supplying the muscles of the hind limbs are more vulnerable to aging than the smaller and shorter axons supplying the forelimbs. Age-related abnormalities of motor end plates in the middle portion of thigh muscles, consistent with axonal atrophy, appeared different from those produced by dying-back neuropathies such as are caused by acrylamide.72 Axonal loss may, of course, be due to neuronal death. However, relatively few studies prove this correlation. Axonal degeneration in the cochlear nerves was observed at 6 months in Sprague-Dawley rats, while axonal atrophy was not seen before 35 months.101 These changes are probably related to the loss of spiral ganglion cells.116 Changes in the plantar nerves of Wistar rats were also first evident at 6 months, and by 24 months 55% of teased nerve fibers were abnormal.197 Both axonal degeneration and regeneration and segmental demyelination occurred, but the decrease in total fiber counts was minimal and the fate of the associated neurons was not examined. In the same animals, changes in the tibial nerve were not found until after 18 months of age and by 24 months only involved 30% of fibers.197 Teased fiber studies preferentially examine the larger fibers, so these figures cannot be extrapolated to the whole fiber population. No significant axonal loss has been found in the inferior alveolar nerve of the rat (C. S. Johansson, personal communication, 2002). Possible explanations for the variation in changes found between different sites could be the effects of local trauma on the foot nerves as suggested by work on guinea pigs.74 The high proportion (72%) of degenerating fibers in the distal tail nerves of 2-year-old rats may also be partly the result of damage produced by handling the animals by the tail123,125 (Fig. 22–3). Alternatively, damage to foot and tail nerves may result from their greater axonal length, as suggested above. This would not, however, explain the extensive degeneration and fiber loss found in the cochlear nerve,101 where distal axonal atrophy may be involved. Besides the apparent loss of axons, major stigmata of aging in peripheral axons are axon (neuroaxonal) dystrophy, axon atrophy, and disturbed myelination. Axon dystrophy

487

and atrophy result in a loss of peripheral target connections both in somatosensory (see Primary Sensory Neurons above) and autonomic pathways193 (i.e., a loss of innervation). Disturbed myelination will affect not only axon conduction (impulse propagation as well as propagation velocity) but also other aspects of axon biology, including the structural integrity of the axon. Axonal dystrophy is a process initially confined to the terminal part of the axon, which again involves certain axons but not others. For example, the extended projections of myelinated primary sensory neurons terminating in the dorsal column of the spinal cord are among the most severely affected.73,111 Axon atrophy, that is, a reduction in diameter of axons resulting in abnormally thick myelin, has been reported in both roots and peripheral nerves,115,126,184 but unequivocal evidence of dying-back atrophy is still lacking. Within aging axons, light microscopic and ultrastructural studies on 24- to 30-month-old Sprague-Dawley rats show abnormalities such as axonal glycogenosomes, degenerate mitochondria, dense membranous bodies, polyglucosan or Lafora bodies, and decorated particles (Fig. 22–4). These abnormalities occur with increasing frequency distally in the tibial and plantar nerves. These abnormalities can, however, also be found close to the nerve cell body in the DRG (Fig. 22–5). Glycogen granules are first seen during aging in axonal mitochondria. These develop into glycogenosomes and are then transformed into polyglucosan bodies (Fig. 22–6).174,214 Active axonal degeneration, indicated by bands of Büngner, and demyelination are rarely prominent and are only seen with electron microscopy. Schwann cell infoldings into the axon are sometimes found, as are Hirano bodies (eosinophilic fusiform bodies with a paracrystalline fine structure100) in the adaxonal Schwann cell cytoplasm.123 It has been suggested that trauma, particularly chronic pressure, may be a factor in the changes found in the plantar nerves, especially because these changes were not seen in branches of the tibial nerve to the gastrocnemius muscle.92 However, the plantar nerves are also among the longest nerves, and the occurrence of loss, dystrophy, and atrophy of other extended axons in nerves with more protected paths, such as visceral autonomic nerves (see Autonomic Neurons below), the cochlear nerve (see below), and the distal tail nerves,123,125 indicates that axon length is more strongly associated with selective vulnerability in aging than trauma. Myelinated Axons In addition to axonal degeneration, peripheral nerves and spinal nerve roots of aged animals show evidence of abnormal myelin, including demyelination and the formation of large intramyelinic swellings.17,32,41,85,126,156,180,197,205,213 Myelin abnormalities occur relatively early in life, appearing soon after 6 months in rats,186 suggesting that nerves have barely finished growing before they start to exhibit signs of regression.

488

Neurobiology of the Peripheral Nervous System

A

B

Myelinated fiber profiles become more irregular with age, with extensive infoldings and outpouchings of the myelin sheath. Such appearances are particularly prominent in the spinal roots and may be explained by axonal shrinkage.126,130 In support of the idea that large myelinated fibers are more vulnerable to the effects of aging compared with small unmyelinated nerves, morphological and teased fiber studies

FIGURE 22–3 A, Tail nerve from an old rat. There is considerable loss of myelinated fibers, many of which have been replaced by clumps of Schwann cell processes (black arrow) distinguishable by their lack of axons from normal Remak fibers containing unmyelinated axons (yellow arrow) (shown in more detail in inset). Most of the myelinated fibers have inappropriately thin myelin for axon diameter. There is no evidence of regeneration or demyelination. Resin section stained with thionin and acridine orange. Bar: 20 ␮m. B, Tail nerve from a 6-month-old normal rat. The myelinated fibers are closely packed with very little collagen between them. Resin section stained with thionin and acridine orange. Bar: 20 ␮m.

showed a reduction in numbers of larger myelinated fibers and an increased variability in internodal length. These changes only appeared after 60 years of age in humans.138 In contrast, several studies found little change in the population of small myelinated fibers.185,199,207 The increased variability in internodal length may result from demyelination and remyelination, or from axonal degeneration and regeneration.

Aging in the Peripheral Nervous System

489

A FIGURE 22–4 A, Plantar nerve from a 24-month-old male Sprague-Dawley rat showing a reduction in myelinated fiber density. There are round inclusions in the axons of many of the fibers. The yellow inclusions are polyglucosan bodies (red arrows); the blue-gray ones are glycogenosomes (black arrows). Some myelin sheaths are inappropriately thin, suggesting remyelination (green arrow). There are occasional macrophages (asterisk). Even the apparently normal myelin sheaths have a more irregular profile with numerous infoldings and outpouchings. Resin section stained with thionin and acridine orange. Bar: 20 ␮m. B, Plantar nerve from a normal 6-month-old rat. The fibers are closely packed. Some have a crenated appearance because they have been sectioned through the paranode (arrow). Resin section stained with thionin and acridine orange. Bar: 20 ␮m.

B

When teased fibers are examined, regenerated fibers show many consecutive, similarly myelinated internodes of the same short length, whereas remyelinated fibers have differing internodal lengths and myelin thickness. Both regeneration and remyelination result in an increased variability in the thickness of the myelin sheath relative to the diameter of its

axon.109 The most dramatic change seen in nerves from older animals is the formation of the large intramyelinic edematous spaces, commonly referred to as myelin bubbles or balloons, as previously mentioned. These appear from 6 months onward and are largest in the spinal roots (Fig. 22–7). Myelin changes are not necessarily associated with axon atrophy,

490

Neurobiology of the Peripheral Nervous System

FIGURE 22–5 DRG from a 26-month-old SpragueDawley rat. Several of the axons contain inclusions. Yellowish ones are polyglucosan bodies (red arrow) and the bluish ones are glycogenosomes (black arrows). The very thinly myelinated fibers are probably remyelinated (green arrows). Resin section stained with thionin and acridine orange. Bar: 20 ␮m.

being found more frequently in the spinal roots, whereas axonal abnormalities are more prominent in the peripheral nerves. The often dense and shrunken appearance of axons running through the myelin bubbles could be due to compression by the edematous fluid that fills them. Segmental demyelination has been reported from a relatively early age in the hind limb nerves of experimental animals. Abnormalities of both axon and myelin are far fewer and appear later in life in nerves such as the phrenic that are relatively protected from trauma and have less extended axons. Focal intramyelinic edema131,156 is more prominent in the spinal roots and phrenic nerve than in the sciatic nerve and its branches but, again, only in older animals. Myelin bubbles are most conspicuous in lumbar spinal roots from approximately 2 years of age in rats.17 This leads to demyelination (Fig. 22–8), and repeated episodes may produce onion bulb formation.64 The axons are frequently unaffected apart from reduction in diameter.131 Extensive axonal degeneration seems to be more common in Charles River rats compared with the SpragueDawley and Wistar strains.41,85 Rarely, severe cystic changes may be found in spinal roots, with axonal degeneration and sprouting, deposition of cholesterol crystals, and macrophage infiltration.32,123 These may result from unrecognized pathologies of the vertebrae and intervertebral discs.32 The cause of focal intramyelinic edema (Fig. 22–8) is still unclear. It was first reported in rats proximal to neuromas, suggesting a relationship with axonal atrophy,202 although axonal atrophy more commonly produces small, dense axons with an abnormally thick myelin sheath. The animals in question were 18 months old. The myelin abnormalities may therefore be due to age or to an age-related alteration in the reaction to injury. Two independent processes may affect myelination: one primary, involving alterations in Schwann cells, and the other secondary to axonal shrinkage. It is generally considered that the primary process is more important. Axonal atrophy

FIGURE 22–6 High-power view of a longitudinally sectioned myelinated nerve fiber in the median plantar nerve of a 24-month-old rat. There are several glycogenosomes in the axon (arrows). The myelin sheath appears normal (asterisk). Ultrathin section contrasted with uranyl acetate and lead citrate. Bar: 1 ␮m.

Aging in the Peripheral Nervous System

491

A

FIGURE 22–7 A, Ventral root from a 24-month-old rat containing many myelinated fibers affected by intramyelinic edema (asterisks). There is little axonal abnormality. Resin section stained with thionin and acridine orange. Bar: 50 ␮m. B, Ventral root from a 4-month-old rat showing closely packed normal myelinated fibers. Resin section stained with thionin and acridine orange. Bar: 50 ␮m.

B

leads to paranodal demyelination, seen in teased fiber preparations as intercalated short, thinly myelinated segments alternating in a regular pattern with segments of normal thickness and nearly normal length. Recovery may occur following this form of demyelination in animals,69 but the situation in human myelinated nerves where this has been studied is less clear.67 Changes occur with age in the proportions of the various myelin proteins in the sciatic nerve and spinal roots of old rats. The relative amounts of P0 and P1 decrease similarly at both sites.210 The relationship of these changes to the intramyelinic edema is uncertain because myelin balloons (or bubbles; see above) are much more prominent in the spinal roots than in the sciatic nerve and thus do not match the pattern of alterations in myelin proteins.

Unmyelinated Axons Relatively few studies have attempted to quantify the effects of age on the unmyelinated fiber population in peripheral nerves. The most obvious changes are in their associated Schwann cells. A unit of Schwann cell processes and associated unmyelinated axons can be conveniently called a Remak fiber after the original description.181 A reduction in the fiber population is seen as the occurrence of Schwann cell processes unassociated with axons. The density of these correlates with age in normal subjects.113 The morphological picture resulting from unmyelinated axon loss is rather different from that associated with myelinated fibers. When unmyelinated axons degenerate, the surrounding Schwann cell processes collapse inward, forming flattened sheets.204 These stacks of flattened

492

Neurobiology of the Peripheral Nervous System

ial,109,113,168 suggesting that the difference was not the result of poor preservation or other artifact.113 Regarding fiber size, one study found that size distribution of unmyelinated axon diameters was unimodal in younger subjects but developed a lower second peak representing smaller fibers from the fifth decade,109 whereas another found a defined age-related bimodal distribution of fiber size. The increase in proportion of very small axons may result from degeneration of the larger axons and their subsequent replacement by regenerative sprouts of smaller diameter. Contamination can also occur, with unmyelinated axon sprouts arising after myelinated fiber degeneration,168 but these will probably not be as small as those resulting from regeneration of unmyelinated fibers per se.

FIGURE 22–8 Electron micrograph of a dorsal root from a 24-month-old rat demonstrating demyelinated fibers (black arrow), thinly remyelinated ones (white arrows), and intramyelinic edema (asterisk). Section contrasted with lead citrate and uranyl actetate. Bar: 5 ␮m.

processes do not resemble bands of Büngner left by myelinated fiber degeneration, which consist of rounded cell processes often within the original Schwann cell basal lamina. Collections of Schwann cell sheets were reported in studies of changes of the aging human sural nerve occurring after the fourth decade, where they were interpreted to indicate loss of axons.168 A study of younger subjects between 17 and 47 years of age reported no difference in incidence of Schwann cell sheets,14 probably because they were too young, since Schwann cell sheets were numerous in older subjects aged 83 and 88 years.196 In addition, Schwann cell processes encircling bundles of collagen fibrils, apparently in place of axons, may be found, although the association with axon loss is not proven (see Chapter 3). Morphometric studies of age changes in unmyelinated axon size and diameter give varying results between animals and humans. In the pelvic nerve of rats aged 30 to 37 months, the number of small fibers (⬍0.7 ␮m) was selectively decreased while larger ones were unaffected. The distributions, however, remained unimodal.162 In the mouse tibial nerve, the smallest unmyelinated fibers also declined by 50% at 27 months of age, with the distribution again remaining unimodal.35 These results contrast with a lack of any effect of age on density of unmyelinated axons in the human sural nerve in biopsy and postmortem mater-

Vascular Changes Age-related changes are observed in several features of the blood vessels supplying nerve trunks. For example, endoneurial blood vessels in human nerves often show a thickening of their basal lamina ensheathment during aging,109 which resembles that seen at earlier ages in patients with diabetic neuropathy,24 and may result from the accumulation of less degradable glycated basal laminal proteins.124 Although not confirmed by animal studies, this may contribute to age-related episodes of local ischemia sufficient to damage the associated nerve trunk. The formation and accumulation of advanced glycation end products (AGEs) during aging may be particularly important because they reduce nitric oxide–dependent vasodilatation as well as altering the physical properties of vascular connective tissue with significant pathological effects. In support of this possibility, administration of AGE-modified albumin to normal rats and rabbits produced an increase in vascular permeability and a decrease in vasodilator responses.217 Other changes involving peripheral blood vessels include a decreased axon reflex associated with impairments in inflammatory responses and wound healing such as are involved in aging.118,146,172 In addition, in 24-month-old Sprague-Dawley rats there is altered sensitivity of perineurial blood vessels to the vasoconstrictors endothelin-1, noradrenaline, and angiotensin II, possibly as a result of alterations in the function or expression of their receptors.121,122 The same authors showed reduced vasoconstriction in aged rats in response to vasopressin.121,122

Autonomic Neurons Selective Vulnerability of Autonomic Neurons to Age-Related Neurodegeneration Of the different branches of the autonomic nervous system, the sympathetic and enteric systems have been most widely studied in relation to aging. The majority of these studies are in rodents. Like other peripheral neurons, sympathetic neurons exhibit the phenomenon of selective vulnerability.42

493

Aging in the Peripheral Nervous System

FIGURE 22–9 Selective age-related axon loss in sympathetic neurons. Immunofluorescence staining for the pan-neuronal marker PGP 9.5 shows a substantial and obvious loss of axons in cerebral blood vessels taken from 24-month-old (B) compared with 2-month-old (A) Sprague-Dawley rats.

A

Neighboring groups of nerve cells within the same ganglion undergo markedly different trajectories of change during their life history. An example is the superior cervical ganglion from which neurons project to target tissues in the head and neck. Superior cervical ganglion neurons projecting to cerebral blood vessels,206 the pineal gland,132 sweat glands of the footpad, skin, and dermal blood vessels2 all exhibit age-related axonal atrophy (Fig. 22–9). In one of these instances,132 the axon loss has been characterized as resulting from loss of whole axon branches with the associated varicosities—the sites of neurotransmitter release. In contrast, neurons projecting to the iris,83 submandibular gland, and other tissues appear to maintain their axonal structure and even continue to grow in old age. Although studied in less detail, sympathetic neurons of the coeliac ganglion also exhibit selective vulnerability during aging.10,45 Axons of neurons projecting to the gut wall10 and renal blood vessels45 exhibit loss of branches and/or reduced neurotransmitter levels during aging, whereas those projecting to the mesenteric vasculature do not. The majority of peripheral autonomic and sensory neurons exhibit continued growth and retraction on a day-to-day basis during development and in early adult life, as was elegantly demonstrated by Purves and co-workers.96,175 This dynamic state extends into adulthood (see Signaling Molecules in Aging Sensory Neurons above). Vulnerability to age-related neuronal atrophy may therefore be the result of a negative balance between ongoing nerve growth and retraction. In the main, studies of this kind are difficult to reproduce in human tissues. Therefore, the question of the adequacy of rodent models for understanding human aging processes as they affect the nervous system remains unresolved. However, in one of the few cases in which direct comparisons

B have been made, losses of axons projecting to human cerebral blood vessels have been observed25 that partly resemble those observed in the aging rat.206 Losses of peripheral fibers were accentuated in tissues from patients with Alzheimer’s disease compared with those from healthy aged individuals.25 In some cases, axon loss has been shown to be matched by age-related loss of dendritic branches of the same neurons. Thus superior cervical ganglion neurons projecting to cerebral blood vessels exhibit loss of dendrites while those projecting to the iris do not (Fig. 22–10).8 Furthermore, the preganglionic neurons of peripheral autonomic neurons may be similarly affected. Thus both pre- and postganglionic sympathetic neurons associated with spinal cord projections to the pelvic ganglia in rats exhibit age-related atrophy of dendrites and loss of synaptic contacts and inputs.58,190 These declines are associated with selective loss of sympathetic, but not parasympathetic, postganglionic neurons in the pelvic ganglia. This is one of the only sites where sympathetic neuron loss has been clearly demonstrated. Studies of the rat vagus nerve indicate that the numbers of parasympathetic and sensory fibers are unaffected by aging.188 There have been few systematic investigations of the effects of age on autonomic neuroeffector junctions. However, electron microscopic studies of aging cerebral blood vessels have demonstrated a proportional loss of nerve axons with age,106 which implies a reduction of functional nerve supply to the vascular smooth muscle. Apart from the major pelvic ganglia (see above), no widespread loss of sympathetic neurons has been demonstrated in other sympathetic ganglia such as the superior cervical and coeliac ganglia of the aging rat. Despite the absence of neuron loss, there is evidence of decreased sympathetic

494

Neurobiology of the Peripheral Nervous System

Young: Iris. Aged: Iris.

*

* *

Young: Middle cerebral artery

Aged: Middle cerebral artery.

*

*

*

FIGURE 22–10 Selective age-related loss of dendrites in sympathetic neurons. Intracellular injection and immunolabeling demonstrate loss of length and complexity in dendritic arborizations of sympathetic neurons projecting to cerebral blood vessels, but not in neurons projecting to the iris, showing that sympathetic neurons vulnerable to age-related degeneration exhibit loss of both axons (see Fig. 22–9) and dendrites. (From Andrews, T. J., Thrasivoulou, C., Nesbit, W., and Cowen, T.: Target-specific differences in the dendritic morphology and neuropeptide content of neurons in the rat SCG during development and aging. J. Comp. Neurol. 368:33, 1996.)

innervation and functional control of the intestines10,11 suggestive of axon atrophy. However, because of the technical difficulties in identifying neuron loss in small groups of neurons, the possibility cannot be discounted that loss of small subpopulations of sympathetic neurons may occur in these and other peripheral ganglia during aging. The only major branch of the autonomic nervous system where large-scale neuron loss has been observed is the enteric nervous system. Losses of up to 50% of neurons have been observed in the myenteric plexus of rats (Fig. 22–11)47,189 and guinea pigs.76 These losses appear to be selective, affecting principally the cholinergic population,47 although this group is not homogeneous and comprises a large proportion of the total. The timing of age-related neuronal impairment remains poorly understood, largely because many studies

rely on comparisons between ‘young’ and ‘old’ rather than examining samples over a wider age range. However, the timing of age-related neuronal loss in the rat enteric nervous system has been examined and found to begin unexpectedly early in the lifespan of the Sprague-Dawley rat, at around 13 months, and to be complete by approximately 16 months.47 Thus neuron loss occurs long before the more obvious indicators of senescence, such as increased incidence of cancer, cataract, and changes of hair color, which are seen between 18 and 21 months in SpragueDawley rats and rather later in Wistar and other strains. Neurodegeneration and Loss of Function As we have stated, it is often difficult to establish a clear link between age-related neurodegeneration and loss of function either in the innervated end organ or in the

Aging in the Peripheral Nervous System

495

FIGURE 22–11 Age-related loss of myenteric neurons. Immunofluorescence staining with the panneuronal marker PGP 9.5 demonstrates substantial loss of neurons in 24-month-old (B) compared with 3-month-old (A) SpragueDawley rats. Note large holes (arrows, B) in aged myenteric ganglion indicating cell loss. (From Cowen, T., and Thrasivoulou, C.: Cerebrovascular nerves in old rats show reduced accumulation of 5-hydroxytryptamine and loss of nerve fibers. Brain Res. 513:237, 1990.)

neurons themselves. In a few situations, such a correlation has been attempted or has been made possible by parallel functional and morphological studies in autonomic neurons. For example, studies of human skin have shown loss of sweating responses to cholinergic agonists during aging, which correlate with observed patterns of neurodegeneration.1 Similar losses of nerves have been demonstrated in aging rat skin and sweat glands.2 In humans, loss of motility in the large intestine—something that commonly affects the elderly95—correlates with observations of agerelated neuron loss in the myenteric plexus of large90 and small55 intestines similar to those observed in the gastrointestinal tract of aging rodents.47 These examples suggest that neurodegeneration is an underlying cause of age-related impairments in neural function in animals and humans.

MECHANISMS OF AGING IN THE PNS Mechanisms Underlying Sensory Impairments during Aging Age-related loss and degeneration of cutaneous receptors and sensory nerve terminals is extensive enough to explain most of the sensory deficits observed in elderly individuals. Thus sensory impairments seem to reflect a process starting in the distal axon domain, possibly instigated by changes in nerve-target interactions22,23,31,75,82,109,206 (see Mechanisms Underlying Age-Related Impairments in Autonomic Neurons below). Incapacitation of the intrinsic machinery that maintains neuron integrity may also contribute to age-related neuronal dysfunction. Given that neuronal longevity in the main matches that of the host organism, postmitotic cells of the nervous system may be especially susceptible to the cumulative damaging effects of aging. Mechanical ‘wear and tear’ probably plays a role23,66; however, it is clearly not the only factor.75,205 Several lines of evidence suggest that aging has a more deleterious effect on myelinated primary afferents than their unmyelinated

counterparts. Neuronal atrophy, axonal lesions, and loss of peripheral nerve endings and receptor organs as well as of centrally projecting nerve terminals preferentially affect large myelinated primary afferent neurons. Although the mere difference in neuronal volume between myelinated and unmyelinated primary afferents may explain the selective vulnerability of the former, perhaps because of the greater metabolic load, it is highly likely that other factors contribute. One possible explanation for the difference in vulnerability may be sought in age-related changes in neurotrophin signaling. The development of the nervous system is characterized by the establishment of neuronal numbers, neuronal connectivity, and a differentiated phenotype through close interactions with target cells. In the adult animal, extension of developmental processes enables the nervous system to adapt to new or altered demands or experiences through plasticity, which includes differential growth of neurons and their axonal and dendritic arborizations.176 In this context, old age may be considered as the final stage of development, wherein neurons, with greater or lesser success, continue to respond to altered functional demands. An inability to maintain appropriate neuronal function in senescence may result from a disturbance in the trophic signaling between neurons and their target cell.43,211,212 This possibility is supported by evidence for a decreased capacity of peripheral target tissues such as skeletal muscle and skin23,154 to synthesize neurotrophic factors of the nerve growth factor (NGF) family of neurotrophins. In parallel, there is a lowered expression of the cognate (i.e., trk) receptors in primary sensory neurons.19,154 Neurotrophins have been shown to regulate the expression of neurofilaments,89 thereby affecting neuronal plasticity. Expression of neurofilaments, the major determinants of axon caliber, is depressed in primary sensory neurons of aged rats,133,171 which may contribute to the impaired capacity of, in particular, large primary sensory neurons to maintain their axonal arborizations in senescence. Neurotrophins also influence the capacity of neurons to withstand the

496

Neurobiology of the Peripheral Nervous System

damaging effects of free radicals166 (see Role of Changes in Autonomic Neurons below). Moreover, by virtue of its capacity to regulate expression of neuropeptides and other neurotransmitter substances, attenuated neurotrophin signaling may influence important functional aspects of primary sensory neurons.18,20,216 Changes in neurotrophin signaling may therefore influence several aspects of the aging process as it affects sensory neurons. In contrast to the decrease in neurotrophin signaling, glialderived neurotrophic factor (GDNF) signaling increases with advancing age.19,153 GDNF has been shown to protect the phenotype of nociceptive primary sensory neurons, as well as preserving the conduction velocity of unmyelinated, but not myelinated, axons.16,160 Thus increased GDNF signaling during senescence, in parallel with decreased neurotrophin signaling, may explain the preserved phenotype and the lack of cell body atrophy and loss of terminals among small unmyelinated primary afferents.22,75,212 An important unresolved issue concerns whether alterations in target tissues, nerve-target interactions, or neuronal/axonal aberrations represent the primary event underlying the changes observed in sensory neurons with advancing age. Recent studies on two sensory pathways, both relying on peripheral receptor cells, may shed some light on this issue. The primary event during aging of the Merkel cell–neurite complex seems to be failure in the maintenance of the Merkel cells and their production of neurotrophin 3 (NT3). In contrast, the Merkel cell axons, although reduced in number and showing decreased expression of trkC (the NT3 receptor), show signs of regenerative growth toward alternative targets that exhibit preserved expression of NT3.23 Thus the loss of Merkel cell–neurite complexes during aging appears to be a peripheral process, starting with the Merkel cell itself, possibly caused by mechanical wear and tear.23 Another interesting model is the hair cells of the cochlea and the sensory neurons of the spiral ganglion. Hearing impairment is common in senescence, and a reduction in both hair cells and in their innervation from the spiral ganglion has been noted. The spiral ganglion cells rely on target-derived BDNF (type II cells) and NT3 (type I cells) for their development. It was recently shown that the process causing loss or damage of hair cells might differ from the mechanism inducing damage or loss of the innervating neurons in the spiral ganglion. Thus the initial process involves damage to the hair cells and their contact with auditory neurons resulting from glutamate–N-methyl- D -aspartate (NMDA) receptor mechanisms that can be blocked by MK801. As a result of this loss of contact, spiral ganglion neurons are deprived of NT3 produced by the target hair cells and are consequently lost, although they can be rescued by treatment with NT3.12,61,70 In both of these examples, the initial process appears to take place in the target Merkel or hair cells, which affects their contacts with the

innervating neurons. As a result of the breakdown of target-neuron contact, the target becomes unable to sustain the sensory innervation, which consequently exhibits secondary degeneration.

Mechanisms Underlying Age-Related Changes in Nerve Trunks The causes of the observed morphological and electrophysiological changes reported in peripheral nerves from aging humans are still unclear. There are several possible mechanisms that could have a deleterious effect on nerve axons or their Schwann cells. These include changes in uptake of metabolites, reduction in protein synthesis by the cell body, disturbances of fast and slow axonal transport, and defects in ‘turnaround’ at the axonal end organ (where molecular cargoes are delivered by anterograde axonal transport and collected by retrograde axonal transport). Defects in turnaround create an imbalance between anterograde and retrograde flow, resulting in dying-back neuropathies such as that induced by p-bromophenylacetylurea.110 In addition, it has been suggested that age changes are the result of repeated minor episodes of trauma or ischemia.109 Ischemia-induced reductions in metabolism or in axonal transport may also contribute to dying-back neuropathy. The electrophysiological studies already discussed provide partial support for this view.31 Alternatively, axonal loss may result from neuronal death, as suggested by reports of loss of anterior horn cells in older subjects,143 although these observations have not been confirmed using unbiased counting methods. Studies of age-related changes in axonal transport have produced contradictory results. For example, axoplasmic transport of cholinesterase was reduced in sciatic nerves of old rats,148 whereas a study in aging mice failed to find a reduction in axonal transport rates despite demonstrating a reduction in Na⫹K⫹-ATPase in sciatic nerve and dorsal roots.184 Because ATPases are transported by fast anterograde transport from the neuron, the reduction found by these authors could indicate a decrease in protein synthesis in the neuron, which might also contribute to the reported slowing of nerve conduction velocity.184 Confusingly, other studies on axons closer to the cell bodies in the DRG suggested that a reduced rate of slow axonal transport may result in an increase in axon diameter and, therefore, an increase in conduction velocity.128,149 Failure of the turnaround system could contribute to atrophy of axon terminals or, alternatively, to an abnormal accumulation of material. Neurons with extended axon arborizations are likely to be most sensitive to age-related impairments of axonal transport and/or of the turnaround system, perhaps explaining the greater vulnerability of neurons of this character. A reduction in medium-molecularweight neurofilaments (NF-M) compared with the light

Aging in the Peripheral Nervous System

form (NF-L) has been demonstrated.209 This may reduce the neurofilament packing density, and thus the diameter of the axon. This is supported by morphometric studies showing that age-related reductions in slow axoplasmic transport are associated with reduced spacing rather than numbers of cytoskeletal elements.33 Dystrophy and Atrophy of Axons, Demyelination, and Selective Vulnerability As discussed above (see Structural Impairments in Axons), large myelinated primary afferents are selectively vulnerable to axon dystrophy. Large myelinated primary afferents use glutamate as the fast neurotransmitter. Studies in the CNS have shown a high incidence of axon dystrophy in glutamatergic neurons, and that the induction of axon dystrophy may relate to oxidative stress in excitatory (glutamatergic) pathways (see Ramirez-Leon et al.179 and references therein). Glutamate signaling can initiate oxidative challenge through stimulation of NMDA receptors, which are also expressed presynaptically141,150; thus glutamate may be detrimental to the glutamatergic terminals themselves.147 Further evidence for this autolytic hypothesis comes from observations that the total content of glutathione, a major scavenger of reactive oxygen species (ROS) expressed by neurons, is increased in dystrophic axons.179 Moreover, vitamin E deficiency has been shown to accelerate the emergence and extent of dystrophic axon lesions in sensory but not autonomic pathways195,201 (see, however, Johnson et al.112). The major contribution to ROS in neurons derives from cell respiration. Thus the metabolic demand on neurons with large and/or extensive axon arbors, such as myelinated sensory fibers to the distal hind limb, may increase the risk for oxidative damage compared with smaller neurons with a less extensive axon arbor. It cannot be excluded that glutamate and ROS act in concert to make some neurons vulnerable to axon dystrophy. Another, not mutually exclusive, mechanism (discussed in the sections on Mechanisms Underlying Sensory Impairments during Aging above and on Mechanisms Underlying Age-Related Impairments in Autonomic Neurons below) is that peripherally projecting neurons may face reduced access to target-derived neurotrophic factors in senescence (for reviews, see Ulfhake et al.211,212 and references below). This is highly relevant in the context of oxidative stress because the available evidence indicates that neurotrophic factors increase neuronal resistance to ROS challenge (see Role of Changes in Autonomic Neurons below). Moreover, deprivation of target-derived neurotrophic factors may decrease the capacity of dependent neurons to respond to ongoing adaptive demands, which include maintaining contact with the target. It is conceivable that axon dystrophy as well as axon atrophy may be induced by this mechanism. This notion is supported by observations that chronic axotomy,

497

a situation in which the axon faces permanent disruption of contact with its target, induces axon atrophy.69 Axon dystrophy and atrophy appear to be primarily cellautonomously regulated, as suggested by the evidence that early signs of axon dystrophy seen in the dorsal column system seem to be confined to the terminals.73 However, myelinated axons are also dependent on Schwann cells for their integrity. Thus axon dystrophy, as well as atrophy, is associated with demyelination.73,111 Studies on multiple sclerosis and experimentally induced encephalomyelitis have shown that axon dystrophy appears in the wake of demyelination, suggesting that demyelination can induce axon dystrophy.86,178,226 A possible explanation for this is that compact myelin or one of its components is necessary for neurofilament organization in the axonal cytoskeleton.27,56,227 Subsequently, demyelination may interfere with axonal transport. Despite controversy regarding the changes that take place in the different components of axonal transport with age, there is evidence suggesting that the rate of slow anterograde transport, which is responsible for the movement of cytoskeleton and enzymes of intermediary metabolism, may be decreased.108,128,149 This has implications for axon plasticity and may reduce activity at the axon terminal, resulting in neurofilament accumulation and subsequent distortion of the axon architecture. Moreover, interactions between compact myelin and the axon are thought to influence neurofilament organization through activation of kinases and phosphatases.27 Hence, it is possible that demyelination itself, by disrupting the compact myelin, may result in abnormally phosphorylated epitopes that are more resistant to ubiquitin-proteasomal degradation and that subsequently can accumulate in the distal axon. However, the evidence regarding the extent and specificity of phosphorylation in the different groups of neurofilaments in aging axons is at present contradictory. Indeed, there is evidence for increased phosphorylation of neurofilaments in senescence.91,209 Interestingly, overload or disturbance of the ubiquitin-proteasomal pathway, which is responsible for the cellular handling of damaged or toxic proteins, has been implicated in the progress of neurodegenerative as well as demyelinating diseases.37,86 Even if circumstantial evidence implies that axon changes are the primary event during aging, altered Schwann cell function, including the production and properties of myelin, cannot be excluded as a potential influence. ROS challenge may cause oxidative damage in the PNS in a manner similar to that affecting the white matter of the CNS during aging,198 in which oxidative damage induces lipid peroxidation and altered membrane properties13 as well as generation of cytotoxic compounds such as 4-hydroxynonenal.77 Another possible cause of myelin sheath disruption during aging is suggested by evidence for an age-related increase in the length of very-long-chain fatty acids (VLCFA) as well as an increased proportion of unsaturated

498

Neurobiology of the Peripheral Nervous System

compared with saturated VLCFA.87 This pattern resembles the situation in adrenoleukodystrophy, an X-linked inherited demyelinating disease in which VLCFA accumulate as a result of disturbed transport into the peroxisomes and impaired ␤-oxidation of VLCFAs. The accumulation of VLCFA is thought to destabilize the myelin and cause an activation of macrophages.62 Recovery after Injury Failure of neuronal protein synthesis may underlie agerelated alterations in motor nerve regeneration. In young animals, the latent period before reinnervation of the nerve stump after transection is only approximately 1 day. However, in aged rats, this period may extend to 8 days. Once this phase is completed, there is little difference between young and old animals in rate of growth of the regenerating axons. Autoradiographic labeling showed that 3 days posttransection, proteosynthesis in the motor neurons increased considerably in young rats, whereas the rate did not increase in 28-month-old rats, perhaps explaining the extended latent period in older animals.93 Regeneration in rat sciatic nerves sectioned at 15 months of age also differed from that in nerves sectioned at 3 months, and the retrograde changes were more marked. Changes in growth factor signaling in older animals may underlie the observed age differences in axonal atrophy and neuronal vacuolation.117 Although investigated less extensively than myelinated axons, regeneration of unmyelinated axons in older animals seems to be slower and less complete than in younger animals.163,215

Mechanisms Underlying Age-Related Impairments in Autonomic Neurons The underlying causes of nonpathological changes in the aging autonomic nervous system remain obscure. Unanswered questions include problems similar to those addressed above in relation to sensory neurons; for example, are neuronal changes secondary to changes in the innervated target tissue? Also, a related question, are neuronal changes secondary adaptations to altered functional demand in the aging organism, or are neuronal changes partly the result of intrinsic changes in the capacity of neurons to cope with the stresses of aging? Neuron-Target Interactions in the Aging Autonomic Nervous System As already stated, neurotrophic theory53,176 provides a model with powerful credentials to help in understanding neuronal aging. Neurotrophic theory states that neurons depend for their survival, and perhaps also for their growth and regulation of other aspects of phenotype, on neurotrophic factors derived from their target tissues. In the autonomic nervous system, ‘target tissues’ may include

peripheral effectors such as muscle and glands, or other neurons. The archetypal neurotrophic factor is NGF.142 However, in recent years a family of relatives of NGF, the neurotrophins, have been discovered with effects in many different regions of the CNS and PNS.15 In addition, a large number of cytokines and other growth factors have been identified with neurotrophic activities on central as well as peripheral neurons (see, e.g., Airaksinen and Saarma5). Based on the view that neurons require trophic interactions with their target tissues, the hypothesis has been advanced that availability of target-derived neurotrophic factors and/or expression of the appropriate receptors may continue to determine neuronal growth, survival, and phenotype during adult life and, perhaps, on into old age.82 In order to test this hypothesis, transplants of target tissues exhibiting age-related changes in the density of innervating sympathetic nerve fibers were taken from 6- or 24-month-old donor rats and placed onto the iris in the anterior eye chamber of young host animals. In this situation, the sympathetic and other nerves innervating the host iris extend over the implanted target tissue, allowing the interaction between host nerves and implanted target tissues of different ages to be studied. Donor blood vessels,206 sweat glands,49 and other tissues became organotypically reinnervated by irideal nerves from the host, testifying to the capacity of the target tissues to organize ingrowing nerves into an appropriate pattern. Cerebral blood vessels transplanted in this way from aged donors exhibited a 50% lower density of reinnervation by young host nerves compared with young transplanted vessels. The pattern as well as the density of reinnervation closely resembled that seen in vivo. It was concluded that target tissues influence both pattern and density of innervation during adult life, including the reductions of nerve density seen in old age. Examination of the responses of host and transplanted neurons revealed that age changes in neuronal plasticity also contribute to the outcome of trophic interactions between nerve and target tissue. Old and young sympathetic ganglia transplanted in oculo in young hosts extended new innervation over the denervated host iris with equal success, indicating that aged sympathetic neurons were unimpaired in their capacity to respond to the axotomy caused by transplanting the ganglia.81 However, examination of the responses of the iris-projecting sympathetic neurons of old hosts to implanted target tissues of different ages revealed that aged host nerves were impaired in their capacity to reinnervate all target tissues, irrespective of age,79 suggesting that intrinsic changes in neurons, as well as alterations in targets, contribute to age-related changes in neuronal plasticity. Treatment of the transplants with NGF failed to counteract the age-related loss of plasticity of aging sympathetic neurons in the heterochronic transplantation model. These apparently contradictory results are likely to reflect the differential effects of age on

499

Aging in the Peripheral Nervous System

NGF mRNA (RT-PCR)

fg/pg GAP DH

30

8–12wk 24m

20 10 0 CV

B

Iris

NGF protein (ELISA) 6

fpg/mg wet wt.

Role of Changes in Target Tissues Reduced expression of neurotrophic factors has been shown in some regions of the brain228 and spinal cord.154 In addition, there is evidence from sensory systems of peripheral changes in neurotrophic factors (see Mechanisms Underlying Sensory Impairments during Aging above). Other studies in the CNS have shown no changes,52 leading to the view that there is no clear association between altered expression of neurotrophic factors and vulnerability to age-related neurodegeneration. Models using sympathetic neurons and their targets, which often consist of regionally distinct tissues, therefore provide an advantage in analyzing and distinguishing the contributions of neurons versus targets to neuronal aging processes. Levels of expression of neurotrophic factors have been examined in target tissues that are associated with sympathetic neurons that are vulnerable to, or protected from, age-related atrophy or degeneration.44,132 These studies failed to show any age-related reduction in neurotrophic factors in those tissues, such as the rat pineal and cerebral blood vessels, that are innervated by neurons exhibiting atrophy during aging. There is also no change in expression of neurotrophic factors in tissues such as the rat iris, which is innervated by neurons that do not exhibit neurodegeneration. However, it has recently been shown that target tissues innervated by neurons vulnerable to age-related atrophy exhibit much lower levels of expression of neurotrophic factors throughout life compared with target tissues supplied by neurons that are not so vulnerable (Fig. 22–12).50 Thus the lifelong level of availability of target-derived neurotrophic factors to the innervating neurons may enhance the capacity of neurons to cope with age-related stressors. This model suggests that capacity to cope with the stresses of aging results partly from the developmental program which leads to functional differentiation of neuronal phenotype. For a fuller discussion of this model, the reader is referred to Cowen et al.50 Further evidence that neurotrophic factors remain important in the aging nervous system comes from experiments involving treatment of different groups of neurons with neurotrophic factors. Sympathetic nerves supplying cerebral blood vessels have been shown to sprout new fibers in response to treatment with NGF.7,106 This growth response is retained in old age6,106 when high doses of NGF are used. However, growth responses to NGF have been shown to be dose dependent,107 and one study has demonstrated that sympathetic nerves in aging rats exhibit

A

8–12wk 24m

4 2 0 CV

C

Iris

NT3mRNA (RT-PCR) 50

tg/pg GAP DH

regeneration of sympathetic neurons, which is unaffected by age, versus the capacity for collateral sprouting, which is impaired with age.43,60 In light of this evidence that changes in targets as well as in neurons contribute to age-related neurodegeneration, it is appropriate to enquire as to the nature of these changes and whether they occur independently of each other or as part of an ‘aging program’ affecting neurons.

8–12wk 24m

40 30 20 10 0 CV

Iris

FIGURE 22–12 Age changes in expression of neurotrophic factors. Reverse transcriptase–polymerase chain reaction for mRNA (A and C) and enzyme-linked immunosorbent assay of protein (B) for NGF (A and B) and for NT3 (C) demonstrate lack of age-related changes of neurotrophin expression in cerebral blood vessels (CV) or iris from 6-month-old and 24-month-old Sprague-Dawley rats. However, neurotrophin expression (mRNA and protein) is markedly higher in iris, where innervating neurons are unaffected by age, compared with CV, where neurons exhibit vulnerability to age-related atrophy of axons (see Fig. 22–9) and dendrites (see Fig. 22–10). This could indicate a protective effect of lifelong exposure to high levels of neurotrophin expression. (From Cowen, T., Woodhoo, A., Sullivan, C. D., et al.: Reduced age-related plasticity of neurotrophin receptor expression in sympathetic neurones of the rat. Aging Cell 2:59, 2003.)

decreased growth responsiveness to NGF compared with neurons from young animals.83 The extracellular matrix (ECM) may make a significant contribution to the progress of neurodegeneration during aging. Laminin is a key component of the ECM with wellknown nerve growth and survival-promoting capabilities. In vitro studies of sympathetic neurons show that laminin and NGF act synergistically to promote neurite outgrowth in mature as well as in aged neurons.46 Examination of the

500

Neurobiology of the Peripheral Nervous System

levels and distribution of laminin in cerebral blood vessels from aged rats shows that high levels of laminin are found in the basal lamina interposed between nerve and the target muscle cell at all ages.80 Laminin levels are reduced in the aged blood vessel wall, making the target tissue less attractive for the maintenance of nerve fibers in that area. Assuming that autonomic nerves undergo continual remodeling throughout life (see Selective Vulnerability of Autonomic Neurons to Age-Related Neurodegeneration above), laminin is likely to be a key player in determining neuronal vulnerability to aging, perhaps through its interactions with neurotrophin signaling. Role of Changes in Autonomic Neurons There is strong evidence that neuronal responsiveness to neurotrophic factors alters during aging and that this is likely to contribute in some way to age-related neurodegeneration. Evidence includes reduced uptake or retrograde axonal transport of NGF in aged rat sympathetic nerves,44 including those projecting to cerebral blood vessels.134 Similar observations have been made in NGFresponsive neurons of the aged rat basal forebrain.40 A decrease in neurofilament light (NF-L) gene expression in the superior cervical ganglion of aging rats could be related to altered axonal transport, as well as to axonal hypotrophy or atrophy.133 Furthermore, as stated above, there is evidence of reduced growth responsiveness to NGF in aging rat sympathetic neurons. Responsiveness to the neurotrophin family of neurotrophic factors is mediated by the trk family of receptor tyrosine kinases, each with specific binding for particular neurotrophins, and by the pan-specific p75 receptor (a member of the tumor necrosis factor receptor family), which binds all neurotrophins.114 The role of p75 includes a role in apoptotic neuronal cell death when expressed alone.151 However, when it is coexpressed with trkA, as in the majority of sympathetic neurons, p75 is generally considered to enhance responsiveness to the neurotrophins.99 Reduced levels of expression of the p75 NGF receptor and the inability to upregulate receptor expression in response to NGF characterize those sympathetic neurons vulnerable to age-related neurodegeneration.50 During development and early postnatal life, neurotrophic factors regulate survival as well as growth of sympathetic, enteric, and other neuron types. These responses are reinforced by an upregulation of expression of the appropriate receptors, notably p75 in the case of sympathetic neurons,152 and this is considered to provide a mechanism by which growing target tissues can influence their innervation. During maturation, however, changes occur in neuronal responsiveness to neurotrophic factors. Sympathetic and NGF-responsive sensory neurons become relatively independent of NGF for their survival57,170,218 while retaining their responsiveness for growth, a change of obvious significance for the functional

preservation of mature neurons. Intrinsic downregulation of neurotrophin receptors and impaired regulation of receptor expresssion50 are therefore the likely mechanisms underlying reduced growth responsiveness of aging autonomic neurons (see above). Neurotrophin signaling may impact on neuronal aging processes in at least one further way. ROS play an important part in aging processes in general, having well-established damaging effects on the structure and function of many cells and organ systems, as previously discussed (see Mechanisms Underlying Sensory Impairments during Aging above).13 Neurons are particularly vulnerable to ROS-induced damage. A novel role for neurotrophic factors in antioxidant defense in neurons, including in sympathetic and enteric neurons of the autonomic nervous system, has recently been discovered.63 In enteric neurons, ROS levels increase with age. Dietary restriction, which increases longevity in rodents and many other species,200 also protects neurons against the age-related increase in ROS levels, apparently by enhancing the antioxidant role of neurotrophic factors. It therefore seems likely that intrinsic changes in neuronal gene expression, including the capacity to express neurotrophin receptors, are an essential part of an antioxidant defense mechanism, which may be adversely influenced by ongoing, age-related increases in intracellular ROS.

CONCLUSIONS Old age exerts highly selective effects on peripheral as on central areas of the nervous system. In the periphery, large myelinated sensory neurons, their axons, and their central connections appear to be particularly vulnerable during aging. The effects of age on the structure of these particular neurons correlates with some of the principal signs of aging in the elderly, which include a range of functional deficits affecting, notably, mechanoreceptive systems. We present evidence that the sensory receptors that these neurons innervate also degenerate with age and that the distal axons appear to be affected before the more proximal parts of the neuron. Based on this evidence, the view is advanced that interactions between sensory neurons and their receptor targets are crucial initiators of an agerelated nervous deterioration, which extends centripetally from the periphery to affect whole neuronal circuits, including central connections in the spinal cord. Defects in neurotrophin signaling between the neuron and its target, associated with impaired axonal transport and disturbed myelination, may provide the mechanism underlying functional and structural impairments of peripheral neurons. Other groups of sensory and motor neurons are affected by age, but in a less obvious way than in the large myelinated axons of sensory neurons. In autonomic neurons, defects in neurotrophin signaling also contribute significantly to the selective impairments that

Aging in the Peripheral Nervous System

affect these neurons in old age. Lifelong exposure to high or low levels of neurotrophic factors respectively provides protection or renders neurons vulnerable. As in sensory neurons, failure of axon transport probably contributes to vulnerability in aging. However, in other respects, autonomic neurons are different from sensory and somatic motor neurons. For example, larger autonomic neurons seem to cope with the stresses of aging better than their smaller relatives, unlike sensory neurons, in which the reverse is the case. Important clues are surfacing that neurotrophic factors, in addition to their well-established role in relation to neuronal growth and survival, help to defend autonomic neurons against free radical damage. ROS-induced damage is also implicated in age-related sensory deficits, as well as in damage to neurons of the CNS. In both areas, glutamatergic neurotransmission appears to confer particular vulnerability to ROS. The strengthening connection between aging and ongoing free radical damage therefore links agerelated neurodegeneration in central, peripheral, and autonomic nervous systems.

REFERENCES 1. Abdel-Rahman, T. A., Collins, K. J., Cowen, T., and Rustin, M.: Immunohistochemical, morphological and functional changes in the peripheral sudomotor neuro-effector system in elderly people. J. Auton. Nerv. Syst. 37:187, 1992. 2. Abdel-Rahman, T. A., and Cowen, T.: Neurodegeneration in sweat glands and skin of aged rats. J. Auton. Nerv. Syst. 46:55, 1993. 3. Adinolfi, A., Yamuy, J., Morales, F. R., and Chase, M. H.: Segmental demyelination in peripheral nerves of old cats. Neurobiol. Aging 12:173, 1991. 4. Ahmad, A., and Spear, P. D.: Effects of aging on the size, density, and number of rhesus monkey lateral geniculate neurons. J. Comp. Neurol. 334:631, 1993. 5. Airaksinen, M. S., and Saarma, M.: The GDNF family: signalling, biological functions and therapeutic value. Nat. Rev. Neurosci. 3:383, 2002. 6. Andrews, T. J., and Cowen, T.: In vivo infusion of NGF induces the organotypic regrowth of perivascular nerves following their atrophy in aged rats. J. Neurosci. 14:3048, 1994. 7. Andrews, T. J., and Cowen, T.: Nerve growth factor enhances the dendritic arborization of sympathetic ganglion cells undergoing atrophy in aged rats. J. Neurocytol. 23:234, 1994. 8. Andrews, T. J., Thrasivoulou, C., Nesbit, W., and Cowen, T.: Target-specific differences in the dendritic morphology and neuropeptide content of neurons in the rat SCG during development and aging. J. Comp. Neurol. 368:33, 1996. 9. Ansved, T., and Larsson, L.: Quantitative and qualitative morphological properties of the soleus motor nerve and the L5 ventral root in young and old rats. J. Neurol. Sci. 96:269, 1990. 10. Baker, D. M., and Santer, R. M.: A quantitative study of the effects of age on the noradrenergic innervation of Auerbach’s plexus in the rat. Mech. Ageing Dev. 42:147, 1988.

501

11. Baker, D. M., Watson, S. P., and Santer, R. M.: Evidence for a decrease in sympathetic control of intestinal function in the aged rat. Neurobiol. Aging 12:363, 1991. 12. Basile, A. S., Huang, J. M., Xie, C., et al.: N-methyl-Daspartate antagonists limit aminoglycoside antibioticinduced hearing loss. Nat. Med. 2:1338, 1996. 13. Beckman, K. B., and Ames, B. N.: The free radical theory of aging matures. Physiol. Rev. 78:547, 1998. 14. Behse, F., Buchthal, F., Carlsen, F., and Knappeis, G. G.: Unmyelinated fibers and Schwann cells of sural nerve in neuropathy. Brain 98:493, 1975. 15. Bennet, M. R., Gibson, W. G., and Lemon, G.: Neuronal cell death, nerve growth factor and neurotrophic models: 50 years on. Auton. Neurosci. 95:1, 2002. 16. Bennett, D. L., Michael, G. J., Ramachandran, N., et al.: A distinct subgroup of small DRG cells express GDNF receptor components and GDNF is protective for these neurons after nerve injury. J. Neurosci. 18:3059, 1998. 17. Berg, B. N., Wolf, A., and Simms, H. S.: Degenerative lesions of spinal roots and peripheral nerves in aging rats. Gerontologia 6:72, 1962. 18. Bergman, E., Carlsson, K., Liljeborg, A., et al.: Neuropeptides, nitric oxide synthase and GAP-43 in B4binding and RT97 immunoreactive primary sensory neurons: normal distribution pattern and changes after peripheral nerve transection and aging. Brain Res. 832:63, 1999. 19. Bergman, E., Fundin, B. T., and Ulfhake, B.: Effects of aging and axotomy on the expression of neurotrophin receptors in primary sensory neurons. J. Comp. Neurol. 410:368, 1999. 20. Bergman, E., Johnson, H., Zhang, X., et al.: Neuropeptides and neurotrophin receptor mRNAs in primary sensory neurons of aged rats. J. Comp. Neurol. 375:303, 1996. 21. Bergman, E., and Ulfhake, B.: Loss of primary sensory neurons in the very old rat: neuron number estimates using the disector method and confocal optical sectioning. J. Comp. Neurol. 396:211, 1998. 22. Bergman, E., and Ulfhake, B.: Evidence for loss of myelinated input to the spinal cord in senescent rats. Neurobiol. Aging 23:271, 2002. 23. Bergman, E., Ulfhake, B., and Fundin, B. T.: Regulation of NGF-family ligands and receptors in adulthood and senescence: correlation to degenerative and regenerative changes in cutaneous innervation. Eur. J. Neurosci. 12:2694, 2000. 24. Bischoff, A.: Die Ultrastruktur peripherer Nerven bei der diabetischen Neuropathie. Dtsch. Ges. Inn. Med. 18:1138, 1967. 25. Bleys, R. L., Cowen, T., Groen, G. J., and Hillen, B.: Perivascular nerves of the human basal cerebral arteries: II. Changes in aging and Alzheimer’s disease. J. Cereb. Blood Flow Metab. 16:1048, 1996. 26. Bolton, C. F., Winkelmann, R. K., and Dyck, P. J.: A quantitative study of Meissner’s corpuscles in man. Neurology 16:1, 1966. 27. Brady, S. T., Witt, A. S., Kirkpatrick, L. L., et al.: Formation of compact myelin is required for maturation of the axonal cytoskeleton. J. Neurosci. 19:7278, 1999. 28. Braund, K. G., McGuire, K. A., and Lincoln, C. E.: Agerelated changes in the peripheral nerves of the dog. I. A morphologic and morphometric study of single-teased fibers. Vet. Pathol. 19:365, 1982.

502

Neurobiology of the Peripheral Nervous System

29. Braund, K. G., McGuire, K. A., and Lincoln, C. E.: Agerelated changes in the peripheral nerves of the dog. II. A morphologic and morphometric study of cross-sectioned nerve. Vet. Pathol. 19:398, 1982. 30. Brocklehurst, J. C., Robertson, D., and James-Groom, P.: Clinical correlates of sway in old age-sensory modalities. Age Aging 11:1, 1982. 31. Buchthal, F., Rosenfalck, A., and Behse, F.: Sensory potentials of normal and diseased human nerves. In Dyck, P. J., Thomas, P. K., and Lambert, E. H. (eds.): Peripheral Neuropathy. Philadelphia, W. B. Saunders, p. 442, 1975. 32. Burek, J. D., Van der Kogel, A. J., and Hollander, C. F.: Degenerative myelopathy in three strains of aging rats. Vet. Pathol. 13:321, 1976. 33. Caselli, U., Bertoni-Freddari, C., Paoloni, R., et al.: Morphometry of axon cytoskeleton at internodal regions of rat sciatic nerve during aging. Gerontologist 45:307, 1999. 34. Cauna, N.: The effects of aging on the receptor organs of the human dermis. In Montagna, W. (ed.): Advances in the Biology of the Skin. New York, Pergamon Press, p. 63, 1965. 35. Ceballos, D., Cuadras, J., Verdu, E., and Navarro, X.: Morphometric and ultrastructural changes with ageing in mouse peripheral nerve. J. Anat. 195:563, 1999. 36. Cerimele, D., Celleno, L., and Serri, F.: Physiological changes in ageing skin. Br. J. Dermatol. 122:13, 1990. 37. Chung, K. K., Dawson, V. L., and Dawson, T. M.: The role of the ubiquitin-proteasomal pathway in Parkinson’s disease and other neurodegenerative disorders. Trends Neurosci. 24:S7, 2001. 38. Coggeshall, R. E.: A consideration of neural counting methods. Trends Neurosci. 15:9, 1992. 39. Coleman, P. D., and Flood, D. G.: Neuron numbers and dendritic extent in normal aging and Alzheimer’s disease. Neurobiol. Aging 8:521, 1987. 40. Cooper, J. D., Lindholm, D., and Sofroniew, M. V.: Reduced transport of [125I]nerve growth factor by cholinergic neurons and down-regulated TrkA expression in the medial septum of aged rats. Neuroscience 62:625, 1994. 41. Cotard-Bartley, M. P., Secchi, J., Glomot, R., and Cavanagh, J. B.: Spontaneous degenerative lesions of peripheral nerves in aging rats. Vet. Pathol. 18:110, 1981. 42. Cowen, T.: Selective vulnerability in adult and aging mammalian neurons. Auton. Neurosci. 96:20, 2002. 43. Cowen, T., and Gavazzi, I.: Plasticity in adult and aging sympathetic neurons. Prog. Neurobiol. 54:249, 1998. 44. Cowen, T., Gavazzi, I., Weingartner, J., and Crutcher, K. A.: Levels of NGF protein do not correlate with changes in innervation of the rat iris in old age. Neuroreport 7:2216, 1996. 45. Cowen, T., Haven, A. J., Wen Qin, C., et al.: Development and ageing of perivascular adrenergic nerves in the rabbit: a quantitative fluorescence histochemical study using image analysis. J. Auton. Nerv. Syst. 5:317, 1982. 46. Cowen, T., Jenner, C., Song, G. X., et al.: Responses of mature and aged sympathetic neurons to laminin and NGF: an in vitro study. Neurochem. Res. 22:1003, 1997. 47. Cowen, T., Johnson, R. J. R., Soubeyre, V., and Santer, R. M.: Restricted diet rescues rat enteric motor neurones from age related cell death. Gut 47:653, 2000.

48. Cowen, T., and Thrasivoulou, C.: Cerebrovascular nerves in old rats show reduced accumulation of 5-hydroxytryptamine and loss of nerve fibers. Brain Res. 513:237, 1990. 49. Cowen, T., Thrasivoulou, C., Shaw, S. A., and AbdelRahman, T. A.: Transplanted sweat glands from mature and aged donors determine cholinergic phenotype and altered density of host sympathetic nerves. J. Auton. Nerv. Syst. 60:215, 1996. 50. Cowen, T., Woodhoo, A., Sullivan, C. D., et al.: Reduced age-related plasticity of neurotrophin receptor expression in sympathetic neurones of the rat. Aging Cell 2:59, 2003. 51. Critchley, M.: The neurology of old age. Lancet 1:1221, 1931. 52. Crutcher, K. A., and Weingartner, J.: Hippocampal NGF levels are not reduced in the aged Fischer-344 rat. Neurobiol. Aging 12:449, 1991. 53. Davies, A. M.: The neurotrophic hypothesis: where does it stand? Philos. Trans. R. Soc. Lond. B 351:389, 1996. 54. de Neeling, J. N., Beks, P. J., Bertelsmann, F. W., et al.: Sensory thresholds in older adults: reproducibility and reference values. Muscle Nerve 17:454, 1994. 55. de Souza, R. R., Moratelli, H. B., Borges, N., and Liberti, E. A.: Age-induced nerve cell loss in the myenteric plexus of the small intestine in man. Gerontology 39:183, 1993. 56. De Waegh, S. M., Lee, V. M., and Brady, S. T.: Local modulation of neurofilament phosphorylation, axonal caliber, and slow axonal transport by myelinating Schwann cells. Cell 68:451, 1992. 57. Deckwerth, T. L., and Johnson, E. M.: Temporal analysis of events associated with programmed cell death (apoptosis) of sympathetic neurons deprived of nerve growth factor. J. Cell Biol. 123:1207, 1993. 58. Dering, M. A., Santer, R. M., and Watson, A. H.: Agerelated changes in the morphology of preganglionic neurons projecting to the rat hypogastric ganglion. J. Neurocytol. 25:555, 1996. 59. Dhall, U., Cowen, T., Haven, A. J., and Burnstock, G.: Perivascular noradrenergic and peptide-containing nerves show different patterns of change during development and ageing in the guinea-pig. J. Auton. Nerv. Syst. 16:109, 1986. 60. Diamond, J., Foerster, A., Holmes, M., and Coughlin, M.: Sensory nerves in adult rats regenerate and restore sensory function to the skin independently of endogenous NGF. J. Neurosci. 12:1467, 1992. 61. Duan, M., Agerman, K., Ernfors, P., and Canlon, B.: Complementary roles of neurotrophin 3 and a N-methylD-aspartate antagonist in the protection of noise and aminoglycoside-induced ototoxicity. Proc. Natl. Acad. Sci. U. S. A. 97:7597, 2000. 62. Dubois-Dalcq, M., Feigenbaum, V., and Aubourg, P.: The neurobiology of X-linked adrenoleukodystrophy, a demyelinating peroxisomal disorder. Trends Neurosci. 22:4, 1999. 63. Dugan, L. L., Creedon, D. J., Johnson, E. M., and Holtzman, D. M.: Rapid suppression of free radical formation by nerve growth factor involves the mitogen-activated protein kinase pathway. Proc. Natl. Acad. Sci. U. S. A. 94:4086, 1997. 64. Duncan, I. D.: Age-related neuropathy in animals. In Thomas, P. K. (ed.): Peripheral Nerve Changes in the Elderly. Chichester, UK, John Wiley, p. 213, 1988. 65. Dunn, E. H.: The influence of age, sex, weight and relationship upon the number of medullated nerve fibers and

Aging in the Peripheral Nervous System

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76. 77.

78. 79.

80.

81.

82.

83.

on the size of the largest fibers in the ventral root of the second cervical nerve of the albino rat. J. Comp. Neurol. 22:131, 1912. Dyck, P. J., Classen, S. M., Stevens, J. C., and O’Brien, P. C.: Assessment of nerve damage in the feet of long-distance runners. Mayo Clin. Proc. 62:568, 1987. Dyck, P. J., Johnson, W. J., Lambert, E. H., and O’Brien, P. C.: Segmental demyelination secondary to axonal degeneration in uremic neuropathy. Mayo Clin. Proc. 46:400, 1971. Dyck, P. J., Karnes, J., O’Brien, P. C., and Zimmerman, I.: Detection thresholds of cutaneous sensations in humans. In Dyck, P. J., Thomas, P. K., Lambert, E. H., and Bunge, R. (eds.): Peripheral Neuropathy, 2nd ed. Philadelphia, W. B. Saunders, p. 1103, 1984. Dyck, P. J., Lais, A. C., Karnes, J. L., et al.: Permanent axotomy, a model of axonal atrophy and secondary segmental demyelination and remyelination. Ann. Neurol. 9:575, 1981. Ernfors, P., Duan, M. L., ElShamy, W. M., and Canlon, B.: Protection of auditory neurons from aminoglycoside toxicity by neurotrophin-3. Nat. Med. 2:463, 1996. Ferrell, W. R., Crighton, A., and Sturrock, R. D.: Agedependent changes in position sense in human proximal interphalangeal joints. Neuroreport 3:259, 1992. Fujisawa, K.: Some observations on the skeletal musculature of aged rats. III. Abnormalities of terminal axons found in motor end-plates. Exp. Gerontol. 11:43, 1976. Fujisawa, K., and Shiraki, H.: Study of axonal dystrophy. II. Dystrophy and atrophy of the presynaptic boutons: a dual pathology. Neuropathol. Appl. Neurobiol. 6:387, 1980. Fullerton, P. M., and Gilliatt, R. W.: Pressure neuropathy in the hind foot of the guinea pig. J. Neurol. Neurosurg. Psychiatry 30:18, 1967. Fundin, B. T., Bergman, E., and Ulfhake, B.: Alterations in mystacial pad innervation in the aged rat. Exp. Brain Res. 117:324, 1997. Gabella, G.: Fall in the number of myenteric neurons in aging guinea pigs. Gastroenterology 96:1487, 1989. Gard, A. L., Solodushko, V. G., Waeg, G., and Majic, T.: 4-Hydroxynonenal, a lipid peroxidation byproduct of spinal cord injury, is cytotoxic for oligodendrocyte progenitors and inhibits their responsiveness to PDGF. Microsc. Res. Tech. 52:709, 2001. Gardener, E.: Decrease in human neurones with age. Anat. Rec. 77:529, 1940. Gavazzi, I.: Collateral sprouting and responsiveness to nerve growth factor of ageing neurons. Neurosci. Lett. 189:47, 1995. Gavazzi, I., Boyle, K. S., Edgar, D., and Cowen, T.: Reduced laminin immunoreactivity in the blood vessel wall of aging rats correlates with reduced innervation in vivo and following transplantation. Cell Tissue Res. 281:23, 1995. Gavazzi, I., and Cowen, T.: Axonal regeneration from transplanted sympathetic ganglia is not impaired by age. Exp. Neurol. 122:57, 1993. Gavazzi, I., and Cowen, T.: Can the neurotrophic hypothesis explain degeneration and loss of plasticity in mature and ageing autonomic nerves? J Auton. Nerv. Syst. 58:1, 1996. Gavazzi, I., Railton, K. L., Ong, E., and Cowen, T.: Responsiveness of sympathetic and sensory irideal nerves

84.

85. 86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

99.

503

to NGF treatment in young and aged rats. Neurobiol. Aging 22:287, 2001. Gescheider, G. A., Beiles, E. J., Checkosky, C. M., et al.: The effects of aging on information-processing channels in the sense of touch: II. Temporal summation in the P channel. Somatosens. Mot. Res. 11:359, 1994. Gilmore, S. A.: Spinal nerve root degeneration in aging laboratory rats: a light microscopic study. Anat. Rec. 174:251, 1972. Giordana, M. T., Richiardi, P., Trevisan, E., et al.: Abnormal ubiquitination of axons in normally myelinated white matter in multiple sclerosis brain. Neuropathol. Appl. Neurobiol. 28:35, 2002. Giusto, N. M., Roque, M. E., and Ilincheta de Boschero, M. G.: Effects of aging on the content, composition and synthesis of sphingomyelin in the central nervous system. Lipids 27:835, 1992. Goff, G. D., Rosner, B. S., Detre, T., and Kennard, D.: Vibration perception in normal man and medical patients. J. Neurol. Neurosurg. Psychiatry 28:503, 1965. Gold, B. G., Mobley, W. C., and Matheson, S. F.: Regulation of axonal caliber, neurofilament content, and nuclear localization in mature sensory neurons by nerve growth factor. J. Neurosci. 11:943, 1991. Gomes, O. A., de Souza, R. R., and Liberti, E. A.: A preliminary investigation of the effects of aging on the nerve cell number in the myenteric ganglia of the human colon. Gerontology 43:210, 1997. Gou, J. P., Eyer, J., and Leterrier, J. F.: Progressive hyperphosphorylation of neurofilament heavy subunits with aging: possible involvement in the mechanism of neurofilament accumulation. Biochem. Biophys. Res. Commun. 215:368, 1995. Grover-Johnson, N., and Spencer, P. S.: Peripheral nerve abnormalities in aging rats. J. Neuropathol. Exp. Neurol. 40:155, 1981. Gutmann, E., Jakoubek, B., Hájek, I., et al.: Effect of age on proteosynthesis in spinal motoneurons following nerve interruption as shown by autoradiography of S35 labelled methionine. Physiol. Bohemoslov. 11:437, 1962. Haegerstrand, A., Dalsgaard, C. J., Jonson, B., et al.: Calcitonin gene-related peptide stimulates proliferation of human endothelial cells. Proc. Natl. Acad. Sci. U. S. A. 87:3299, 1990. Hall, K. E.: Aging and neural control of the GI tract. II. Neural control of the aging gut: can an old dog learn new tricks? Am. J. Physiol. Gastrointest. Liver Physiol. 283:G827, 2002. Harris, L. W., and Purves, D.: Rapid remodelling of sensory endings in the corneas of living mice. J. Neurosci. 9:2210, 1989. Hashizume, K., and Kanda, K.: Differential effects of aging on motoneurons and peripheral nerves innervating the hindlimb and forelimb muscles of rats. Neurosci. Res. 22:189, 1995. Helme, R. D., and McKernan, S.: Effects of age on the axon reflex response to noxious chemical stimulation. Clin. Exp. Neurol. 22:57, 1985. Hempstead, B., Martin-Zanca, D., Kaplan, D. R., et al.: High-affinity NGF binding requires coexpression of the trk proto-oncogene and the low affinity NGf receptor. Nature 350:678, 1991.

504

Neurobiology of the Peripheral Nervous System

100. Hirano, A.: Hirano bodies and related neuronal inclusions. Neuropathol. Appl. Neurobiol. 20:3, 1994. 101. Hoeffding, V., and Feldman, M. L.: Changes with age in the morphology of the cochlear nerve in rats: light microscopy. J. Comp. Neurol. 276:537, 1988. 102. Hökfelt, T.: Neuropeptides in perspective: the last ten years. Neuron 7:867, 1991. 103. Holzer, P.: Local effector functions of capsaicin-sensitive sensory nerve endings: involvement of tachykinins, calcitonin gene-related peptide and other neuropeptides. Neuroscience 24:739, 1988. 104. Howell, T. H.: Senile deterioration of the central nervous system: a clinical study. Br. Med. J. 1:56, 1949. 105. Impallomeni, M., Kenny, R. A., Flynn, M. D., et al.: The elderly and their ankle jerks. Lancet 1:670, 1984. 106. Isaacson, L. G., and Crutcher, K. A.: Uninjured aged sympathetic neurons sprout in response to exogenous NGF in vivo. Neurobiol. Aging 19:333, 1998. 107. Isaacson, L. G., Mareska, M., Nixdorf, W., and Oris, J. T.: Dose-dependent response of mature cerebrovascular axons in vivo following intracranial infusion of nerve growth factor. Neurosci. Lett. 222:21, 1997. 108. Jacob, J. M.: Fast axonal transport rates are unchanged in 6- and 24-month F344 rats. Brain Res. 699:154, 1995. 109. Jacobs, J. M., and Love, S.: Qualitative and quantitative morphology of human sural nerve at different ages. Brain 108:897, 1985. 110. Jakobsen, J., and Brimijoin, S.: Axonal transport of enzymes and labeled proteins in experimental axonopathy induced by p-bromophenylacetylurea. Brain Res. 229:103, 1981. 111. Jellinger, K., and Jirasek, A.: Neuroaxonal dystrophy in man: character and natural history. Acta Neuropathol. (Berl.) Suppl. 5:16, 1971. 112. Johnson, J. E., Mehler, W. R., and Miquel, J.: A fine structural study of degenerative changes in the dorsal column nuclei of aging mice: lack of protection by vitamin E. J. Gerontol. 30:395, 1975. 113. Kanda, T., Tsukagoshi, H., Oda, M., et al.: Morphological changes in unmyelinated nerve fibres in the sural nerve with age. Brain 114:585, 1991. 114. Kaplan, D. R., and Miller, F. D.: Neurotrophin signal transduction in the nervous system. Curr. Opin. Neurobiol. 10:381, 2000. 115. Kazui, H., and Fujisawa, K.: Radiculoneuropathy of aging rats: a quantitative study. Neuropathol. Appl. Neurobiol. 14:137, 1988. 116. Keithley, E. M., Ryan, A. F., and Feldman, M. L.: Cochlear degeneration in aged rats of four strains. Hear. Res. 59:171, 1992. 117. Kerezoudi, E., King, R. H. M., Muddle, J. R., et al.: Influence of age on the late retrograde effects of sciatic nerve section in the rat. J. Anat. 187:27, 1995. 118. Khalil, Z., and Helme, R.: Sensory peptides as neuromodulators of wound healing in aged rats. J. Gerontol. A Biol. Sci. Med. Sci. 51:354, 1996. 119. Khalil, Z., and Helme, R. D.: Sequence of events in substance P plasma extravasation in rat skin. Brain Res. 500:256, 1989. 120. Khalil, Z., Ralevic, V., Bassirat, M., et al.: Effects of ageing on sensory nerve function in rat skin. Brain Res. 641:265, 1994.

121. Kihara, M., Nakasaka, Y., Mitsui, Y., et al.: Aging differentially modifies sensitivity of nerve blood flow to vasocontractile agents (endothelin-1, noradrenaline and angiotensin II) in sciatic nerve. Mech. Ageing Dev. 114:5, 2000. 122. Kihara, M., Shioyama, M., Okuda, K., and Takahashi, M.: The impact of aging on vasa nervorum, nerve blood flow and vasopressin responsiveness. Can. J. Neurol. Sci. 29:164, 2002. 123. King, R. H. M.: Age changes in the peripheral nervous system. In Mohr, U., Dungworth, D. L., and Capen, C. C. (eds.): Pathobiology of the Aging Rat. Washington, DC, ILSI Press, p. 35, 1994. 124. King, R. H. M., Llewelyn, J. G., Thomas, P. K., et al.: Diabetic neuropathy: abnormalities of Schwann cell and perineurial basal laminae. Implications for diabetic vasculopathy. Neuropathol. Appl. Neurobiol. 15:339, 1989. 125. King, R. H. M., and Thomas, P. K.: Distal axonal degeneration in aging rats. Neuropathol. Appl. Neurobiol. 9:73, 1983. 126. Knox, C. A., Kokmen, E., and Dyck, P. J.: Morphometric alteration of rat myelinated fibers with aging. J. Neuropathol. Exp. Neurol. 48:119, 1989. 127. Kokmen, E., Bossemeyer, R. W. J., Barney, J., and Williams, W. J.: Neurological manifestations of aging. J. Gerontol. 32:411, 1977. 128. Komiya, Y.: Slowing with age of the rate of slow axon flow in bifurcating axons of rat dorsal root ganglion cells. Brain Res. 183:477, 1980. 129. Krinke, G.: Spinal radiculoneuropathy in aging rats: demyelination secondary to neuronal dwindling? Acta Neuropathol. (Berl.) 59:63, 1983. 130. Krinke, G., Froehlich, E., Herrmann, M., et al.: Adjustment of the myelin sheath to axonal atrophy in the rat spinal root by the formation of infolded myelin loops. Acta Anat. 131:182, 1988. 131. Krinke, G., Suter, J., and Hess, R.: Radicular myelinopathy in aging rats. Vet. Pathol. 8:335, 1981. 132. Kuchel, G. A., Crutcher, K. A., Naheed, U., et al.: NGF expression in the aged rat pineal gland does not correlate with loss of sympathetic axonal branches and varicosities. Neurobiol. Aging 20:685, 1999. 133. Kuchel, G. A., Poon, T., Irshad, K., et al.: Decreased neurofilament gene expression is an index of selective axonal hypotrophy in aging. Neuroreport 8:799, 1996. 134. Kudwa, A., Shoemaker, S., Crutcher, K., and Isaacson, L.: Evidence for reduced accumulation of exogenous neurotrophin by aged sympathetic neurons. Brain Res. 948:24, 2002. 135. La Forte, R. A., Melville, S., Chung, K., and Coggeshall, R. E.: Absence of neurogenesis of adult rat dorsal root ganglion cells. Somatosens. Mot. Res. 8:3, 1991. 136. Lamberts, S. W. J., van den Beld, A. W., and van der Lely, A. J.: The endocrinology of aging. Science 278:419, 1997. 137. Larish, D. D., Martin, P. E., and Mungiole, M.: Characteristic patterns of gait in the healthy old. Ann. N. Y. Acad. Sci. 515:18, 1988. 138. Lascelles, R. G., and Thomas, P. K.: Changes due to age in internodal length in the sural nerve in man. J. Neurol. Neurosurg. Psychiatry 29:40, 1966. 139. Lawson, S. L.: Morphological and biochemical cell types of sensory neurons. In Scott, S. E. (ed.): Sensory Neurons:

Aging in the Peripheral Nervous System

140.

141. 142. 143.

144. 145.

146. 147. 148.

149.

150.

151.

152.

153.

154.

155.

156.

157.

158.

Diversity, Development and Plasticity. New York, Oxford University Press, p. 27, 1992. Lawson, S. N., and Waddell, P. J.: Soma neurofilament immunoreactivity is related to cell size and fiber conduction velocity in rat primary sensory neurons. J. Physiol. (Lond.) 435:41, 1991. Leist, M., and Nicotera, P.: Apoptosis, excitotoxicity, and neuropathology. Exp. Cell Res. 239:183, 1998. Levi-Montalcini, R.: The nerve growth factor: 35 years later. Science 237:1154, 1987. Low, P. A., Okazaki, H., and Dyck, P. J.: Splanchnic preganglionic neurons in man. I. Morphometry of preganglionic cytons. Acta Neuropathol. (Berl.) 40:55, 1977. Macintosh, S. R., and Sinclair, D. C.: Age-related changes in the innervation of the rat snout. J. Anat. 125:149, 1978. Madeira, M. D., Sousa, N., Santer, R. M., et al.: Age and sex do not affect the volume, cell numbers, or cell size of the suprachiasmatic nucleus of the rat: an unbiased stereological study. J. Comp. Neurol. 361:585, 1995. Makinodan, T., and Kay, M. M. B.: Age influence on the immune system. Adv. Immunol. 29:287, 1989. Mattson, M. P., Keller, J. N., and Begley, J. G.: Evidence for synaptic apoptosis. Exp. Neurol. 153:35, 1998. McMartin, D. N., and O’Connor, J. A. J.: Effect of age on axoplasmic transport of cholinesterase in rat sciatic nerves. Mech. Ageing Dev. 10:241, 1979. McQuarrie, I. G., Brady, S. T., and Lasek, R. J.: Retardation of the rate of slow axonal transport of cytoskeletal elements during maturation and aging. Neurobiol. Aging 10:359, 1989. Michaelis, E. K.: Molecular biology of glutamate receptors in the central nervous system and their role in excitotoxicity, oxidative stress and aging. Prog. Neurobiol. 54:369, 1998. Miller, F. D., and Kaplan, D. R.: Neurotrophin signaling pathways regulating neuronal apoptosis. Cell. Mol. Life Sci. 58:1045, 2001. Miller, F. D., Speelman, A., Mathew, T. C., et al.: Nerve growth factor derived from terminals selectively increases the ratio of p75 to trkA NGF receptors on mature sympathetic neurons. Dev. Biol. 161:206, 1994. Ming, Y., Bergman, E., Edström, E., and Ulfhake, B.: Evidence for increased GDNF signaling in aged sensory and motor neurons. Neuroreport 10:1529, 1999. Ming, Y., Bergman, E., Edström, E., and Ulfhake, B.: Reciprocal changes in the expression of neurotrophin mRNAs in target tissues and peripheral nerves of aged rats. Neurosci. Lett. 273:187, 1999. Mione, M. C., Dhital, K. K., Amenta, F., and Burnstock, G.: An increase in the expression of neuropeptidergic vasodilator, but not vasoconstrictor, cerebrovascular nerves in aging rats. Brain Res. 460:103, 1988. Mitsumori, K., Maita, K., and Shirasu, Y.: An ultrastructural study of spinal nerve roots and dorsal root ganglia in aging rats with spontaneous radiculoneuropathy. Vet. Pathol. 18:714, 1981. Mittal, K. R., and Logmani, F. H.: Age-related reduction in 8th cervical ventral nerve root myelinated fibre diameters and numbers in man. J. Gerontol. 42:8, 1987. Miwa, T., Miwa, Y., and Kanda, K.: Dynamic and static sensitivities of muscle spindle primary endings in aged rats to ramp stretch. Neurosci. Lett. 201:179, 1995.

505

159. Monji, A., Morimoto, N., Okuyama, I., et al.: The number of noradrenergic and adrenergic neurons in the brain stem does not change with age in male Sprague-Dawley rats. Brain Res. 641:171, 1994. 160. Munson, J. B., and McMahon, S. B.: Effects of GDNF on axotomized sensory and motor neurons in adult rats. Eur. J. Neurosci. 9:1126, 1997. 161. Nagashima, K., and Oota, K.: A histopathological study of the human spinal ganglia. 1. Normal variations in aging. Acta Pathol. Jpn. 24:333, 1974. 162. Nakayama, H., Noda, K., Hotta, H., et al.: Effects of aging on numbers, sizes and conduction velocities of myelinated and unmyelinated fibers of the pelvic nerve in rats. J. Auton. Nerv. Syst. 69:148, 1998. 163. Navarro, X., Kamei, H., and Kennedy, W. R.: Effect of age and maturation on sudomotor nerve regeneration in mice. Brain Res. 447:133, 1988. 164. Navarro, X., and Kennedy, W. R.: Changes in sudomotor nerve territories with aging in the mouse. J. Auton. Nerv. Syst. 31:101, 1990. 165. Nilsson, J., von Euler, A. M., and Dalsgaard, C. J.: Stimulation of connective tissue cell growth by substance P and substance K. Nature 315:61, 1985. 166. Nistico, G., Ciriolo, M. R., Fiskin, K., et al.: NGF restores decreases in catalase activity and increases superoxide dismutase and glutathione peroxidase activity in the brain of aged rats. Free Radical Biol. Med. 12:177, 1992. 167. Norris, A. H., Shock, N. W., and Wagman, I. H.: Age changes in the maximum conduction velocity of motor fibres of human ulnar nerves. J. Appl. Physiol. 5:589, 1953. 168. Ochoa, J., and Mair, W. G.: The normal sural nerve in man. II. Changes in the axons and Schwann cells due to ageing. Acta Neuropathol. 13:217, 1969. 169. Ohta, M., Offord, K., and Dyck, P. J.: Morphometric evaluation of first sacral ganglia of man. J. Neurol. Sci. 22:73, 1974. 170. Orike, N., Thrasivoulou, C., Wrigley, A., and Cowen, T.: Differential regulation of survival and growth in adult sympathetic neurons: an in vitro study of neurotrophin responsiveness. J. Neurobiol. 47:295, 2001. 171. Parhad, I. M., Scott, J. N., Cellars, L. A., et al.: Axonal atrophy in aging is associated with a decline in neurofilament gene expression. J. Neurosci. Res. 41:355, 1995. 172. Parkhouse, N., and LeQuesne, P. M.: Impaired neurogenic vascular response in patients with diabetes and neuropathic foot lesions. N Engl. J. Med. 318:1306, 1988. 173. Perret, E., and Regli, F.: Age and the perceptual threshold for vibratory stimuli. Eur. Neurol. 4:65, 1970. 174. Powell, H. C., Ward, H. W., Garrett, R. S., et al.: Glycogen accumulation in the nerves and kidney of chronically diabetic rats. J. Neuropathol. Exp. Neurol. 38:114, 1979. 175. Purves, D., Hadley, R. D., and Voyvodic, J. T.: Dynamic changes in the dendritic geometry of individual neurons visualized over periods of up to three months in the superior cervical ganglion of living mice. J. Neurosci. 6:1051, 1986. 176. Purves, D., Snider, W. D., and Voyvodic, J. T.: Trophic regulation of nerve cell morphology and innervation in the autonomic nervous system. Nature 336:123, 1988. 177. Quoniam, C., Hay, L., Roll, J. P., and Harlay, F.: Age effects on reflex and postural responses to propriomuscular inputs

506

178.

179.

180.

181.

182.

183.

184.

185.

186.

187. 188.

189.

190.

191.

192.

193.

194.

195.

Neurobiology of the Peripheral Nervous System generated by tendon vibration. J. Gerontol. A Biol. Sci. Med. Sci. 50:155, 1995. Raine, C. S., and Cross, A. H.: Axonal dystrophy as a consequence of long-term demyelination. Lab. Invest. 60:714, 1989. Ramirez-Leon, V., Kullberg, S., Hjelle, O. P., et al.: Increased glutathione levels in neurochemically identified fibre systems in the aged rat lumbar motor nuclei. Eur. J. Neurosci. 11:2935, 1999. Rao, R. S., and Krinke, G.: Changes with age in the number and size of myelinated axons in the rat L4 dorsal spinal root. Acta Anat. 117:187, 1983. Remak, R.: Observationes Anatomicae et Microscopicae de Systematis Nervosi Structura. Berlin, Sumptibus et Formis Reimerianis, 1838. Rexed, B.: Contributions to the knowledge of the postnatal development of the peripheral nervous system in man. Acta Psychiat. Neurol. Suppl. 33:1, 1944. Robbins, S., Waked, E., and McClaran, J.: Proprioception and stability: foot position awareness as a function of age and footwear. Age Ageing 24:67, 1995. Robertson, A., Day, B., Pollock, M., and Collier, P.: The neuropathy of elderly mice. Acta Neuropathol. (Berl.) 86:163, 1993. Rosenberg, S. I., Malmgren, L. T., and Woo, P.: Agerelated changes in the internal branch of the rat superior laryngeal nerve. Arch. Otolaryngol. Head Neck Surg. 115:78, 1989. Saitua, F., and Alvarez, J.: Do axons grow during adulthood? A study of the caliber and microtubules of sural nerve axons in young, mature, and aging rats. J. Comp. Neurol. 269:203, 1988. Samorajski, T.: Age differences in the morphology of posterior tibial nerves of mice. J. Comp. Neurol. 157:439, 1974. Santer, R. M.: Morphological evidence for the maintenance of the cervical sympathetic system in aged rats. Neurosci. Lett. 130:248, 1991. Santer, R. M., and Baker, D. M.: Enteric neuron numbers and sizes in Auerbach’s plexus in the small and large intestine of young adult and aged rats. J. Auton. Nerv. Syst. 25:59, 1988. Santer, R. M., Dering, M. A., Ranson, R. N., et al.: Differential susceptibility to ageing of rat preganglionic neurones projecting to the major pelvic ganglion and of their afferent inputs. Auton. Neurosci. 96:73, 2002. Satorre, J., Cano, J., and Reinoso-Suarez, F.: Stability of the neuronal population of the dorsal lateral geniculate nucleus (LGNd) of aged rats. Brain Res. 339:375, 1985. Schimrigk, K., and Ruttinger, H.: The touch corpuscles of the plantar surface of the big toe: histological and histometrical investigations with respect to age. Eur. J. Neurosci. 19:49, 1980. Schmidt, R. E., Beaudet, L., Plurad, S. B., et al.: Pathologic alterations in pre- and postsynaptic elements in aged mouse sympathetic ganglia. J. Neurocytol. 24:189, 1995. Schmidt, R. E., Chae, H. Y., Parvin, C. A., and Roth, K. A.: Neuroaxonal dystrophy in aging human sympathetic ganglia. Am. J. Pathol. 136:1327, 1990. Schmidt, R. E., Coleman, B. D., and Nelson, J. S.: Differential effect of chronic vitamin E deficiency on the development of neuroaxonal dystrophy in rat gracile/cuneate

196.

197. 198.

199.

200. 201.

202.

203. 204.

205.

206.

207.

208.

209.

210.

211.

212.

213. 214.

nuclei and prevertebral sympathetic ganglia. Neurosci. Lett. 123:102, 1991. Schroder, J. M., and Gibbels, E.: Marklöse Nervernfasern im Senium und Spätstadium der Thalidomid-polyneuropathie: quantitativ-electronen-mikroscopische Untersuchungen. Acta Neuropathol. 39:271, 1977. Sharma, A. K., Bajada, S., and Thomas, P. K.: Age changes in the tibial and plantar nerves of the rat. J. Anat. 130:417, 1980. Sloane, J. A., Hollander, W., Moss, M. B., et al.: Increased microglial activation and protein nitration in white matter of the aging monkey. Neurobiol. Aging 20:395, 1999. Smith, D. O., and Rosenheimer, J. L.: Factors governing speed of action potential conduction and neuromuscular transmission in aged rats. Exp. Neurol. 83:358, 1984. Sohal, R. S., and Weindruch, R.: Oxidative stress, caloric restriction, and aging. Science 273:59, 1996. Southam, E., Thomas, P. K., King, R. H. M., et al.: Experimental vitamin E deficiency in rats: morphological and functional evidence of abnormal axonal transport secondary to free radical damage. Brain 114:915, 1991. Spencer, P. S., and Thomas, P. K.: The examination of isolated nerve fibres by light and electron microscopy, with observations on demyelination proximal to neuromas. Acta Neuropathol. (Berl.) 16:177, 1970. Steiness, I. B.: Vibratory perception in normal subjects: a biothesiometric study. Acta Med. Scand. 198:315, 1957. Thomas, P. K., and King, R. H. M.: The degeneration of unmyelinated axons following nerve section: an ultrastructural study. J. Neurocytol. 3:497, 1974. Thomas, P. K., King, R. H. M., and Sharma, A. K.: Changes with age in the peripheral nerves of the rat: an ultrastructural study. Acta Neuropathol. (Berl.) 52:1, 1980. Thrasivoulou, C., and Cowen, T.: Regulation of rat sympathetic nerve density by target tissues and NGF in maturity and old age. Eur. J. Neurosci. 7:381, 1995. Toghi, H., Tsukagoshi, H., and Toyokura, Y.: Quantitative changes with age in normal sural nerves. Acta Neuropathol. (Berl.) 38:213, 1977. Tomonaga, M.: Histochemical and ultrastructural changes in senile human skeletal muscle. J. Am. Geriatr. Soc. 25:125, 1977. Uchida, A., Yorifuji, H., Lee, V. M., et al.: Neurofilaments of aged rats: the strengthened interneurofilament interaction and the reduced amount of NF-M. J. Neurosci. Res. 58:337, 1999. Uchida, Y., Tomonaga, M., and Nomura, K.: Age-related changes of myelin proteins in the rat peripheral nervous system. J. Neurochem. 46:1376, 1986. Ulfhake, B., Bergman, E., Edstrom, E., et al.: Regulation of neurotrophin signaling in aging sensory and motoneurons: dissipation of target support? Mol. Neurobiol. 21:109, 2000. Ulfhake, B., Bergman, E., and Fundin, B. T.: Impairment of peripheral sensory innervation in senescence. Auton. Neurosci. 96:43, 2002. van Steenis, G., and Kroes, R.: Changes in the nervous system and musculature of old rats. Vet. Pathol. 8:320, 1971. Vanneste, J., and van den Bosch de Aguilar, P.: Mitochondrial alterations in the spinal ganglion neurons in ageing rats. Acta Neuropathol. 54:83, 1981.

Aging in the Peripheral Nervous System 215. Verdu, E., Buti, M., and Navarro, X.: Functional changes of the peripheral nervous system with aging in the mouse. Neurobiol. Aging 17:73, 1996. 216. Verge, V. M., Gratto, K. A., Karchewski, L. A., and Richardson, P. M.: Neurotrophins and nerve injury in the adult. Philos. Trans. R. Soc. B 351:423, 1996. 217. Vlassara, H., Fuh, H., Makita, Z., et al.: Exogenous advanced glycosylation end products induce complex vascular dysfunction in normal animals: a model for diabetic and aging complications. Proc. Natl. Acad. Sci. U. S. A. 89:12043, 1992. 218. Vogelbaum, M. A., Tong, J. X., and Rich, K. M.: Developmental regulation of apoptosis in dorsal root ganglion neurons. J. Neurosci. 18:8928, 1998. 219. Vrancken, A. F. J. E., Franssen, H., Wokke, J. H. J., et al.: Chronic idiopathic axonal polyneuropathy and successful aging of the peripheral nervous system in elderly people. Arch. Neurol. 59:533, 2002. 220. Wagman, I. H., and Lesse, H.: Maximum conduction studies of motor fibers of ulnar nerve in human subjects of various ages and sizes. J. Neurophysiol. 15:235, 1952. 221. Wayner, M. J. J., and Emmers, R.: Spinal synaptic delay in young and aged rats. Am. J. Physiol. 19:403, 1958.

507

222. Whanger, A. D., and Wang, H. S.: Clinical correlates of the vibratory sense in elderly psychiatric patients. J. Gerontol. 29:39, 1974. 223. Wickelgren, I.: Is hippocampal cell death a myth? Science 271:1229, 1996. 224. Willis, W. D., and Coggeshall, R. E.: Sensory Mechanisms of the Spinal Cord. New York, Plenum Press, 1991. 225. Winkelmann, R. K.: Nerve changes in the aging skin. In Montagna, W. (ed.): Advances in Biology of Skin, Vol. 6: Aging: Symposium on the Biology of Skin. Oxford, Pergamon Press, p. 51, 1965. 226. Woodruff, R. H., and Franklin, R. J.: Demyelination and remyelination of the caudal cerebellar peduncle of adult rats following stereotaxic injections of lysolecithin, ethidium bromide, and complement/anti-galactocerebroside: a comparative study. Glia 25:216, 1999. 227. Yin, X. H., Crawford, T. O., Griffin, J. W., et al.: Myelinassociated glycoprotein is a myelin signal that modulates the caliber of myelinated axons. J. Neurosci. 18:1953, 1998. 228. Yurek, D. M., and Fletcher-Turner, A.: Differential expression of GDNF, BDNF, and NT-3 in the aging nigrostriatal system following a neurotoxic lesion. Brain Res. 891:228, 2001.

23 Oxidative Stress and Excitatory Neurotoxins in Neuropathy PHILLIP A. LOW

Overview Free Radical Biology Biologic Generators Protective Mechanisms Oxidative Stress Oxidative Stress and Peripheral Nerve Disease Diabetic Neuropathy Evidence of Oxidative Stress in Non-neural Tissue Integrative View of Oxidative Mechanisms of Diabetic Neuropathy

Evidence of Oxidative Stress in Diabetic Peripheral Nerve Antioxidants Molecular Pathogenesis of Endothelial Dysfunction Molecular Pathogenesis of Sensory Neuropathy: Pivotal Role of the Mitochondrion Molecular Pathogenesis of Schwann Cell Injury Oxidative Stress Related to Nerve Ischemia

OVERVIEW The pivotal role of oxidative stress and injury to peripheral nerve has become increasingly appreciated in the past decade. A number of interacting mechanisms cause oxidative injury. Especially important are the roles of ischemia, ischemia-reperfusion, the inflammatory response, and hyperglycemia. All these mechanisms cause oxidative stress. Oxidative DNA damage in turn can cause cellular dysfunction and apoptosis. Among the disorders, the role of oxidative stress has been best studied in diabetic neuropathy and, to a lesser degree, in ischemic neuropathy. It is likely also important in the toxic and inflammatory neuropathies, but little information is available at this time. This chapter focuses on oxidative stress as applied to diabetic neuropathy and ischemic injury. To a lesser extent we explore the role of excitotoxic injury, especially as it applies to painful neuropathies.

FREE RADICAL BIOLOGY Reactive oxygen species (ROS), or free radicals, comprise unpaired electrons that are highly reactive. They

Nerve Ischemia Reperfusion Injury Neuropathology Inflammatory Response Role of Cytokines Genetic Susceptibility Molecular Pathogenesis Neuroprotection of Ischemia-Reperfusion Injury N-Methyl-D-Aspartate and Glutamate Receptors

are generated by a number of chemical reactions, and the flow of electrons is shown in Equations 1 through 536: enz-H2 ⫹ 2O2 : enz ⫹ 2O2⫺ ⫹ 2H⫹

(1)

enz-H2 ⫹ O2 : enz ⫹ H2O2

(2)

O2⫺ ⫹ O2⫺ ⫹ 2H⫹ : H2O2 ⫹ O2

(3)

Me(n⫹)chelate ⫹ O2 : Me(n⫺1)chelate ⫹ O2

(4)

Me(n⫺1)chelate ⫹ H2O2 :Me(n⫹)chelate ⫹ OH ⫹ OH•

(5)

The enzyme (enz) can reduce O2 univalently to yield the superoxide anion (O2 : O2⫺) in Equation 1 (often called the univalent pathway) or divalently to yield hydrogen peroxide (O2 : H2O2) in Equation 2 (divalent pathway). The superoxide anion can undergo spontaneous dismutation (Equation 3), yielding hydrogen peroxide.113 Another fate of the superoxide anion is the reduction of metal (Me) chelates (Equation 4). The H2O2 may react with reduced 509

510

Neurobiology of the Peripheral Nervous System

metal chelates by the Haber-Weiss reaction (Equation 5) to generate hydroxyl radical (OH•). OH• reacts with several compounds, including lipids.131 Carbon-centered lipid radical (L•) is formed from polyunsaturated fatty acid by hydrogen abstraction (Fig. 23–1). Molecular rearrangement may occur to yield conjugated diene (a relatively stable footprint of lipid peroxidation). This compound reacts with oxygen rapidly to give lipid peroxy radical (LO2•). This radical attacks another lipid molecule and abstracts a hydrogen atom to give lipid hydroperoxide. At the same time it attaches another lipid radical and starts the process all over again (propagation). Molecular rearrangement yields lipid endoperoxide and fragmentation yields malondialdehyde, a frequently measured but relatively nonspecific footprint of lipid peroxidation. These small organic radicals are reactive and therefore short lived.151 The hydroxyl radical is so reactive that it reacts within 1 to 5 molecular diameters. The half-life of the hydroxyl radical has been estimated to be about 10⫺9 seconds.

Biologic Generators There are a number of biologic generators of free radicals.57 One source is sympathetic adrenergic postganglionic terminals. For instance, in response to ischemia, catecholamines are released from these nerve terminals and undergo oxidation, releasing free radicals.12 However, this contribution is quantitatively probably small.57 A more important source is the mitochondrion (mitochondrial leak mechanism). In normal mitochondria 1% of electron flow results in O2,181 thought to occur at the NADH

dehydrogenase step and near the ubiquinone component.11 This mechanism is enhanced in the diabetic state and in aging, in which free radical leakage is increased. A third source of ROS is the leukocyte. Leukocytes can produce large amounts of O2. Indeed, as much as 70% of O2 consumed by the activity of leukocytes may be converted to O2.186 This mechanism is an important source of ROS in reperfusion injury following ischemia to nerve. It is also an important source in the inflammatory neuropathies.

Protective Mechanisms A number of mechanisms reduce the toxicity of ROS.143 The toxicity of mitochondrial superoxide leak is greatly attenuated by the composition of the electron transport chain. Electrons and oxidant species so formed are tightly bound so that they do not participate in redox reactions. The electronic structure of free oxygen is also protective. Oxygen is a bi-radical that contains two separate orbitals each housing an electron with identical spin. Most molecules have two electrons with opposite spin. This difference in rotation creates a spin restriction that limits the ability of O2 to accept electrons directly. For oxygen to accept both electrons, an energy-requiring spin inversion must take place first. Toxic transitional metals are rendered nontoxic by certain proteins, yielding nontoxic proteins such as ferritin, transferrin, and ceruloplasmin. A series of antioxidants (Table 23–1) reduce chain initiation and suppress free radical generation, hence these are sometimes referred to as preventive antioxidants.131 Examples (with the relevant radical in parentheses)

Fatty acid with 3 double bonds Hydrogen abstraction

–H

Lipid radical Molecular rearrangement

Conjugated diene O2 uptake

Peroxy radical RH

OOH +R

Lipid hydroperoxide

Lipid endoperoxide Fragmentation

Malondialdehyde

FIGURE 23–1 Molecular steps of lipid peroxidation from polyunsaturated free fatty acids.

Oxidative Stress and Excitatory Neurotoxins in Neuropathy

Table 23–1. Some Free Radical Defenses of Peripheral Nerve Cytosolic

Membrane

Ascorbate, urate, cysteine, glutathione, transferrin, albumin; ␤-carotene, ceruloplasmin, reduced glutathione, glutathione peroxidase, catalase, superoxide dismutase, glutathione reductase ␣-Tocopherol

are glutathione peroxidase (hydroperoxides), catalase (H2O2), transferrin (Fe2⫹), albumin (Cu2⫹), ␤-carotene (singlet oxygen), and ceruloplasmin (Fe2⫹). Chain-breaking antioxidants suppress free radical chain oxidation by molecular oxygen. These are often typically divided into water-soluble and lipid-soluble antioxidants. Examples of water-soluble antioxidants are ascorbate, urate, cysteine, and glutathione. The only significant lipid-soluble chain-breaking antioxidant is ␣-tocopherol.

Oxidative Stress There are a number of factors that increase free radical activity in tissue. The first is the increased generation of ROS by mechanisms such as ischemia, hyperglycemia, and inflammation. The second mechanism of increased free radical activity is reduced scavenging capacity. Free radical scavengers include superoxide dismutase (SOD), catalase, reduced glutathione (GSH), glutathione S-transferase, glutathione peroxidase, glutathione reductase, and ␣-tocopherol. Another mechanism leading to oxidative stress is an alteration of pro-oxidant status of tissue, increasing susceptibility of the cell to ROS. These include an accumulation of free iron,39 reduction in pH,4 and increased lipolysis of phospholipids, resulting in increased release of polyunsaturated fatty acids.48,152,196

OXIDATIVE STRESS AND PERIPHERAL NERVE DISEASE Oxidative stress is likely to play major pathogenetic roles in several types of neuropathy. These include neuropathies associated with ischemia (e.g., angiopathy, diabetes, peripheral vascular disease), inflammation (e.g., infections, inflammatory-demyelinating, sarcoid, paraneoplastic disease), neurotoxicity (e.g., heavy metals, cisplatin), and hyperglycemia. In this chapter we focus on diabetic and ischemic neuropathy. Earlier studies have focused on measurements of footprints of oxidative stress, typically in plasma. Subsequently measurements in nerve have been done. More recently, emphasis has shifted to measuring

511

indices that indicate tissue and especially DNA damage. Immunohistochemistry of 8-hydroxydeoxyguanosine and hydroxynonenal (HNE) permit simultaneous cellular localization and indexing of DNA damage.

DIABETIC NEUROPATHY With the combination of macrovascular and microvascular disease and chronic hyperglycemia, it is not surprising that evidence of oxidative stress has accumulated for multiple tissues. Much of the early data derived from non-neural tissue. We review some of this information here, followed by data on neural tissue. A number of major pathogenetic mechanisms have been postulated. These mechanisms converge in a final common pathway of oxidative injury. Finally we provide molecular schemes for the pathogenesis of diabetic sensory neuropathy, diabetic endothelial dysfunction, and Schwann cell apoptosis.

Evidence of Oxidative Stress in Non-neural Tissue Human Diabetes Plasma levels of lipid peroxide are increased in human diabetes.1,72,111,159 The highest levels were found in patients with microvascular angiopathy, manifested as retinopathy or microalbuminuria, and the lowest in those diabetic patients without angiopathy27,159; levels were normal in patients with well-controlled diabetes.159 The relationship may relate to the observation that low-density lipoproteins of diabetic patients are significantly more oxidizable than those of controls, an abnormality that is correctable by 6 weeks’ treatment with the antioxidant probucol.3 Presumably diabetics with angiopathy, who have higher levels of low-density lipoproteins, will have the greatest lipid peroxidation. GSH is reduced in erythrocytes from patients with type 2 diabetes mellitus, and there is a corresponding increase in oxidized glutathione (GSSG).122 In subsequent studies, these workers additionally demonstrated reductions in the glutathione-synthesizing enzyme ␥-glutamylcysteine synthetase, and in transport of the thiol [S-(s,4-dintrophenyl)glutathione] in erythrocytes of these patients. These abnormalities were reversible with improved glycemic control. They also demonstrated that a high-glucose medium augments the toxicity of xenobiotics on K562 cells, associated with reduction in both the enzyme and its messenger RNA (mRNA).197 Erythrocyte cuprozinc SOD is reduced in patients with type 2 diabetes.111 This reduction is suggested to be mediated by the accumulation of intracellular H2O2.99 ␣-Tocopherol levels are reported to be reduced in the platelets78,99 and erythrocytes but not plasma of insulindependent diabetes mellitus patients.

512

Neurobiology of the Peripheral Nervous System

Wolff 190 has emphasized the role of decompartmentalized transitional metals in diabetic patients in addition to the effects of hyperglycemia in producing auto-oxidative lipid peroxidation. Particular emphasis has been placed on copper and iron. Copper levels have been reported to be higher in diabetics than in normal subjects and are highest in those with angiopathy.110,134 Experimental Diabetes Oxidative stress occurs in experimental diabetes induced by streptozotocin and alloxan, and in the diabetic BB Wistar rat. Lipid peroxidation in chemical diabetes appears to be due to hyperglycemia and not the agent, because the pattern of changes is not agent specific. Streptozotocin and alloxan cause similar changes, and additionally the spontaneously diabetic BB Wistar rat shows virtually identical patterns of tissue antioxidant enzyme changes.50 Furthermore, these alterations are preventable by insulin treatment.10,50,188,189 The common mechanism of increased oxidative stress is, therefore, diabetes (hyperglycemia) and not its mode of induction. Plasma and liver lipid peroxides are increased in streptozotocin- or alloxan-induced diabetes111,150 and improved by ␣-tocopherol supplementation.150 Tissue levels of lipid peroxide, estimated by the thiobarbituric acid method, were increased in kidney and retina and accompanied by a reduction in fat-soluble antioxidants as determined by the ferric chloride–bipyridyl reaction. These changes were eliminated by insulin treatment.132 Complex patterns of changes in antioxidant enzymes have been described in different tissues in streptozocin diabetes.188,189 Liver and kidney have reduced catalase and SOD. Glutathione peroxidase and GSH are reduced in liver; glutathione peroxidase is increased in kidney. Catalase and glutathione reductase are increased in heart and pancreas and SOD is additionally increased in pancreas. One of the most common alterations is a reduction in cuprozinc SOD. This reduction has been reported in numerous tissues, including erythrocyte32,43,100,111,201; liver8,111; retina and kidney43,100; and spleen, heart, testis, pancreas, and skeletal muscle in rats with streptozotocin and/or alloxan diabetes.8,43,100,111 The loss of SOD appears to be a function of duration and severity of diabetes.85,99 ␣-Tocopherol is reported to be reduced in streptozotocin diabetes.85,99 Our observations are that plasma ␣-tocopherol is very variable and is greatly dependent on dietary intake; it can be increased in experimental diabetes because of polyphagia.130

Integrative View of Oxidative Mechanisms of Diabetic Neuropathy A major area of research in diabetic neuropathy is on oxidative stress. Earlier studies have focused on the demonstration of the presence of footprints of oxidative stress in serum

and in nerve tissue. Subsequent studies have focused on cellular localization of oxidative DNA injury and on the pathogenetic role of oxidative stress leading to peripheral neuronal injury and apoptosis. A number of potential pathogenetic mechanisms of diabetic neuropathy have been identified. These include ischemia/hypoxia, auto-oxidative lipid peroxidation, advanced glycation end products (AGEs), polyol pathway overactivity, protein kinase C (PKC) overactivity, excessive lipolysis, growth factor deficiency, and the inflammatory response. Of interest is that each of these can cause oxidative stress. Ischemia/Hypoxia in Experimental Diabetic Neuropathy A nerve blood flow deficit of 50% in experimental diabetic neuropathy (EDN)20,126,180 results in the generation of hypoxanthine from ATP, NADPH (the cofactor), and conversion of the inactive enzyme to xanthine oxidase.104,113 Figure 23–2 shows a reduction in nerve blood flow by 50% and a dosedependent prevention by treatment with ␣-lipoic acid. There is a concomitant reduction in GSH, the best index of current oxidative stress. Concomitantly, GSSG increases. Treatment with the antioxidant ␣-lipoic acid results in a dose-dependent normalization of GSH (Fig. 23–3). The diabetic state and endoneurial ischemia increase lipolysis, resulting in an increase in ␻-6 fatty acids such as linoleic acid and arachidonic acid,196 whose peroxidation results in 4-HNE, and causes prominent cytotoxic effects in cultured endothelial cells manifested by morphologic changes, diminished cellular viability, and impaired endothelial barrier function.60 HNE has relatively long half-lives within cells (minutes to hours), allowing for multiple interactions with cellular components.82 It impairs glutamate transport and mitochondrial function in neurons148 and mediates oxidative stress–induced neuronal apoptosis.95 Immunohistochemical labeling of HNE is a reliable indicator of oxidative injury to cells, including dorsal root ganglion (DRG) neurons and Schwann cells179 (see Figure 23–10 later). Auto-oxidative Lipid Peroxidation Hyperglycemia, by a process of auto-oxidation in the presence of decompartmentalized redox-active trace transitional metals, can generate highly reactive oxidants and result in lipid peroxidation.191 We have demonstrated, using an in vitro lipid peroxidation model (ascorbateiron–ethylenediaminetetraacetic acid [EDTA] preparation), that a high-glucose medium will result in lipid peroxidation, in vitro, of brain and sciatic nerve. The addition of 20-mM glucose to the incubation medium increased lipid peroxidation fourfold, confirming rapid and marked glucose-mediated auto-oxidative lipid peroxidation.129 These studies confirm the observation of auto-oxidative glycation/oxidation in plasma.7,190 The relationship between oxidative glycation and free radical production has been explored.66 Glucose auto-oxidation

Oxidative Stress and Excitatory Neurotoxins in Neuropathy

513

FIGURE 23–2 Nerve blood flow (NBF) and nerve vascular resistance of controls (Con), animals with streptozotocin diabetic neuropathy (STZ), and animals supplemented with lipoic acid at doses of 20 mg/kg (STZ20), 50 mg/kg (Con50; STZ50), and 100 mg/kg (STZ100). Lipoic acid supplementation results in normal NBF and nerve vascular resistance. Significance of difference, a, P ⬍ .05. (From Nagamatsu, M., Nickander, K. K., Schmelzer, J. D., et al.: Lipoic acid improves nerve blood flow, reduces oxidative stress, and improves distal nerve conduction in experimental diabetic neuropathy. Diabetes Care 18:1160, 1995, with permission.)

results in the production of protein-reactive ketoaldehydes, hydrogen peroxide, and other highly reactive oxidants and the fragmentation of proteins (indicative of free radical mechanisms). Glycation and oxidation are simultaneous and were considered to be inextricably linked.66 There are numerous indices of oxidative stress. These include malondialdehyde, lipid hydroperoxide, and conju-

gated dienes. Malondialehyde is rather nonspecific and lipid hydroperoxides are somewhat unstable.101 The formation of the double bond in conjugated dienes (see Fig. 23–1) confers stability, and its measurement has been a preferred index of oxidative stress. Sciatic nerve conjugated dienes were significantly increased in diabetic rats that had been diabetic for 1, 4, or 12 months (Fig. 23–4).

FIGURE 23–3 Sciatic nerve reduced glutathione (GSH) concentrations in controls (Con) and in animals with restricted caloric intake (Con[R]), streptozotocin diabetes (STZ), ␣-tocopherol depletion (⫺), and supplementation with lipoic acid at 20, 50, and 100 mg/kg. Lipoic acid supplementation resulted in a dose-dependent prevention of GSH. Significance of difference versus control, *P ⬍ .05; ***P ⬍ .001; a, P ⬍ .05; t, P ⬍ .001. (From Nagamatsu, M., Nickander, K. K., Schmelzer, J. D., et al.: Lipoic acid improves nerve blood flow, reduces oxidative stress, and improves distal nerve conduction in experimental diabetic neuropathy. Diabetes Care 18:1160, 1995, with permission.)

514

Neurobiology of the Peripheral Nervous System

Protein Kinase C There is some uncertainty as to whether activity of PKC is increased in nerve and as to which subtype is altered. Our own studies found an increase but did not study specific isoforms.90 It is known that inhibition of PKC-␤ will reduce oxidative stress109 and will normalize the deficits in blood flow and nerve conduction.19 In endothelial cells, high glucose causes NF-␬B activation. Co-incubation with a selective PKC inhibitor, calphostin C, produced a concentration-dependent inhibition of glucose-induced NF-␬B activation, suggesting that PKC is important at the endothelial cell level in the activation of adhesion molecules and generation of ROS. FIGURE 23–4 Sciatic nerve conjugated dienes at 1, 4, and 12 months of diabetes. (From Low, P. A., and Nickander, K. K.: Oxygen free radical effects in sciatic nerve in experimental diabetes. Diabetes 40:873, 1991, with permission.)

AGE, Its Receptor, and Nuclear Factor-␬B Glycation and the formation of AGE is followed by binding with its receptor (RAGE), the activation of nuclear factor-␬B (NF-␬B),164 and the generation of ROS and an inflammatory response.83 There is induction of specific DNA binding activity for NF-␬B in the vascular cell adhesion molecule-1 promoter region. The necessity for RAGE and the role of ROS was demonstrated by a block of this induction by antiRAGE immunoglobulin G or N-acetylcysteine (GSH donor). The application of this finding to humans is supported by the finding that peripheral blood mononuclear cells isolated from patients with diabetic nephropathy show increased activation of NF-␬B.64 There is a vicious cycle involving AGE-producing superoxide, superoxide-accelerating AGE generation, and AGE-quenching nitric oxide (NO).14 Generation of ROS by Polyol Pathway Overactivity Polyol pathway overactivity generates ROS in a number of ways. Depletion in NADPH results in NO deficiency and an increase in leukocyte superoxide anion generation. Because NADPH is also required for the regeneration of GSH, its depletion results in a reduction of GSH.96 Sorbitol dehydrogenase, the second enzyme in the polyol pathway that converts sorbitol to fructose, also contributes to oxidative stress most likely because depletion of its cofactor NAD⫹ leads to more glucose being channeled through the polyol pathway.96 Oxidative stress results in poly(ADP-ribose) polymerase (PARP) activation. In conditions with reduced NAD⫹, PARP activation leads to caspase-3 activation and apoptosis.172 Polyol pathway overactivity results in reductions in myoinositol and taurine. The latter is a potent antioxidant, and its depletion174 further contributes to oxidative stress.

Excessive Lipolysis The diabetic state and endoneurial ischemia increase lipolysis, resulting in an increase in ␻-6 fatty acids such as linoleic acid or arachidonic acid,196 whose peroxidation results in 4-HNE60 (discussed earlier). Malondialdehyde is also generated in the cyclooxygenase pathway. Growth Factor Deficit Lipid peroxidation is aggravated by a reduction in nerve growth factor.58 Nerve growth factor reduction will reduce, and its administration will restore, glutathione peroxidase and catalase.133,157 Inflammatory Response The main sources of ROS in mammals are the leukocyte and the mitochondrion. The quantitative role of the leukocyte, cytokines, and catecholamine oxidation in diabetic nerve is uncertain. Leukocytic infiltration is not a feature in most cases of EDN. In human diabetic neuropathy, there is some evidence of an immune-mediated process as suggested by the presence of iritis and inflammatory infiltrates in sympathetic ganglia.41,103 Some forms of diabetic neuropathy, such as acute autonomic neuropathy and the subacute radiculoplexus neuropathies, might be associated with prominent round cell infiltration.44,147 A role of associated ROS and oxidative stress should be considered.

Evidence of Oxidative Stress in Diabetic Peripheral Nerve Oxidative stress in diabetic nerves can develop by (1) an impairment of free radical defenses, (2) an increase in pro-oxidant status, and (3) increased free radical generation. Evidence for changes in all three areas are present in diabetic nerves and combine to generate oxidative stress and injury to cause diabetic peripheral neuropathy. Reduction of Free Radical Defenses in Diabetic Peripheral Nerve Peripheral nerve has a number of cytosolic and lipophilic antioxidant defenses, as described earlier (see Table 23–1).

515

Oxidative Stress and Excitatory Neurotoxins in Neuropathy

Table 23–2. Enzymatic Free Radical Scavenger Activity in Nerve Relative to Brain and Liver Activity† Antioxidant* GSH-GSSG GSH-Px (: H2O2) GSH-Px (: t-BOOH) GSSG Reductase GST (: CDNB) GST (: 4-HNE) DT-diaphorase SO

Nerve

% of Brain

Brain

261.0 ⫾ 24.0 6.4 ⫾ 2.1 4.6 ⫾ 1.4 2.7 ⫾ 0.3 9.4 ⫾ 2.4 5.4 ⫾ 1.2 9.9 ⫾ 2.2 93.8 ⫾ 12.4

10% 13% 9% 13% 4% 5%

2620.0 ⫾ 124.0 48.5 ⫾ 5.5 50.3 ⫾ 1.4 20.5 ⫾ 0.6 232.5 ⫾ 14.0 116.5 ⫾ 10.4

Liver 74.9 ⫾ 7.1 144.0 ⫾ 16.9 56.3 ⫾ 6.1 171.3 ⫾ 55.4 345.2 ⫾ 92.0 208.0 ⫾ 49.9 171.3 ⫾ 55.4

GSH-GSSG ⫽ total glutathione; GSH-Px ⫽ glutathione peroxidase; t-BOOH ⫽ tert-butyl-hydroperoxide; GSSG Reductase ⫽ gluathione disulfide reductase; GST ⫽ glutathione transferase; CDNB ⫽ 1-chloro-2,4-dinitrobenzene; 4-HNE ⫽ 4-hydroxy-2,3-trans-nonenol; SO ⫽ superoxide dismutase. † Results are expressed as nanomoles per minute per milligram of protein, ⫾SEM. *

The cytosolic and membrane antioxidants work in concert in a well-organized interacting chain. Activity of all components is needed to maintain these antioxidants in their reduced state. There is an enormous literature on free radical biology, but information on peripheral nerve, in particular diabetic peripheral nerve, is quite limited. Free radical defenses in nerve are selectively reduced. Compared with that of brain, SOD activity is as high,153,154 but GSH and glutathione-containing enzymatic scavengers (glutathione peroxidase and reductase) are only about 10% that of brain61 (Table 23–2). Under conditions of increased oxidative stress, diabetic nerves undergo a reduction in GSH and neuropathy. The diabetic state results in additional alterations in these defenses (Table 23–3). Cuprozinc SOD is reduced in sciatic nerve in EDN, and this reduction is improved by insulin treatment.102 Glutathione peroxidase is further reduced in EDN.61,101 These workers were able to regress glutathione peroxidase activity against blood glucose concentration (y ⫽ 69.3 ⫺ 0.9x, where y is enzyme activity measured as nanomoles per milligram of protein per minute and x is blood glucose in millimoles per liter; r ⫽ .9). Plasma and leukocyte ascorbic acid are reduced and oxidation of this

antioxidant is increased.68,171 GSH is reduced in diabetic nerves.126,130 SOD,102 GSH,130 glutathione peroxidase,61 and catalase and total quinone reductase53,141 are reduced in diabetic rodent nerves. Increased Free Radical Generation in EDN Potential sources of free radical generation in diabetes include hyperglycemia, ischemia, increased mitochondrial leak, catecholamine oxidation, and leukocytes. A nerve blood flow deficit of 50% in EDN20,162,165,180 results in the generation of hypoxanthine from ATP and NADPH (the cofactor), and conversion of the inactive enzyme to xanthine oxidase.101,113 Increased Diabetic Pro-oxidant Status Diabetic pro-oxidant status is increased because diabetic sciatic nerve has increased polyunsaturated fatty acids with excessive lipolysis.196 It has also been suggested that metals, especially copper, might be decompartmentalized or increased.190 Although there is good evidence that oxidative stress occurs in the diabetic state, the demonstration of lipid peroxidation alone is insufficient evidence for a free

Table 23–3. Changes in Antioxidant Defenses in EDN Antioxidant

Change

Reference(s)

Superoxide dismutase (cuprozinc) Glutathione peroxidase Ascorbic acid content

Reduced Reduced Reduced

Ascorbic oxidation

Increased

Reduced glutathione

Reduced

Low and Nickander (1991)102 Hermenegildo et al. (1993)61 Sodhi et al. (1989),170 Greene et al. (1989)55 Sodhi et al. (1989),170 Greene et al. (1989)55 Nickander et al. (1994)130 Nagamatsu et al. (1995)126

516

Neurobiology of the Peripheral Nervous System

radical role in the etiopathogenesis of neuropathy. We propose that the following four criteria should be satisfied before accepting that neuropathy is due to excessive lipid peroxidation: 1. An increase in lipid peroxidation in previously normal nerves results in neuropathy. 2. An increase in lipid peroxidation in diabetic neuropathic nerves further worsens function. 3. Antioxidant improves or prevents neuropathy. 4. This improvement or prevention is associated with an improvement in the indices of lipid peroxidation of peripheral nerve. When weanling rats were fed an ␣-tocopherol–deficient diet, plasma ␣-tocopherol became unmeasurable.130 Endoneurial oxidative stress developed, as indicated by a reduction in GSH and increased lipid peroxidation (conjugated dienes, lipid hydroperoxides). These findings were associated with the development of a sensory neuropathy in previously normal nerves.130 A second line of evidence derives from the worsening of neuropathy in EDN130 associated with an increase in lipid peroxidation. The pro-oxidant primaquine caused conduction deficit in the hind limb nerves of previously normal rats, a reduction in sciatic nerve blood flow, and endoneurial hypoxia; these deficits were prevented by treatment with the antioxidant probucol.18 Evidence of increased nerve lipid peroxidation is now available. Diabetic peripheral nerve has increased conjugated dienes,102,130 reduced GSH,130 and reduced glutathione peroxidase.61,101 Among the indices of increased oxidative stress in a chronic in vivo situation, malondialdehyde, conjugated dienes, and lipid hydroperoxides are increased in peripheral nerve in EDN,33,130,162 and the increase is more consistent in lumbar dorsal root and superior cervical ganglia.130 The most reliable index of increased oxidative stress is reduction in GSH.130

Antioxidants Several antioxidants have shown promise in the treatment of EDN (Table 23–4). Probucol, a powerful free radical scavenger,73 normalizes both nerve blood flow and electrophysiology.18,77 These workers also reported an improvement in nerve perfusion with a 1% vitamin E diet.77 ␣-Lipoic acid is a powerful lipophilic antioxidant.16,145 It has a number of additional properties relevant to neuropathy. It stimulates nerve growth factor123 and promotes fiber regeneration in a culture system.40 ␣-Lipoic acid will prevent the deficit in nerve blood flow (see Fig. 23–2) and the distal electrophysiologic changes in EDN.126 These benefits are associated with improved indices of lipid peroxidation, the most sensitive of which is a prevention of the reduction in GSH (see Fig. 23–3). Mild improvement in EDN has been reported following treatment with GSH.13 One percent butylated hydroxytoluene for 2 months completely prevented the conduction deficit in EDN.21 Carvedilol, an antioxidant and vasodilator, is reported to be efficacious in preventing the deficits in both blood flow and conduction.29 Cameron and Cotter17 have reported improvement in both nerve blood flow and nerve conduction in EDN following treatment with the iron chelator deferoxamine. We have also found that modified SOD will prevent the deficit in nerve blood flow. However, a significant deficiency in most antioxidant studies in EDN is the lack of measurements of oxidative stress, so that it is not known if antioxidants, all of which have multiple mechanisms of action, improve neuropathy by their antioxidant or other properties. We evaluated the gene expression of glutathione peroxidase, catalase, and SOD (cuprozinc and manganese separately) in L4/5 DRG and superior cervical ganglion, as well as enzyme activity of glutathione peroxidase in DRG and sciatic nerve, in EDN of 3-month and 12-month durations. Gene expression of glutathione peroxidase, catalase, cuprozinc SOD, and manganese SOD was not reduced in EDN at either 3 or 12 months. Catalase mRNA was

Table 23–4. Antioxidants in the Treatment of Experimental Diabetic Neuropathy Antioxidant

Lipid Peroxidation

NBF Deficit

NCS Deficit

Reference(s)

Probucol ␣-Tocopherol

Not studied Not studied

Prevented Prevented

Prevented Prevented

␣-Lipoic acid GSH Butylated hydroxytoluene Carvedilol Deferoxamine SOD ␤-Carotene Ascorbic acid

Improved Not studied Not studied Not studied Not studied Improved Not studied Not studied

Prevented Not studied Not studied Prevented Prevented Prevented Prevented Partial prevention

Prevented Partial prevention Prevented Prevented Prevented Partial prevention Prevented Partial prevention

Cameron et al. (1994)18 Cameron et al. (1994),18 Cotter et al. (1995),30 Karasu et al. (1995)77 Nagamatsu et al. (1995)126 Bravenboer et al. (1992)13 Cameron et al. (1993)21 Cotter and Cameron (1995)29 Cotter and Cameron (1995)28 J. F. Poduslo, unpublished Cotter et al. (1995)30 Cotter et al. (1995)30

GSH ⫽ reduced glutathione; NBF ⫽ nerve blood flow; NCS ⫽ nerve conduction study; SOD ⫽ superoxide dismutase.

Oxidative Stress and Excitatory Neurotoxins in Neuropathy

significantly increased in EDN at 12 months. Glutathione peroxidase enzyme activity was normal in sciatic nerve. We concluded that gene expression is not reduced in peripheral nerve tissues in very chronic EDN. Changes in enzyme activity may be related to duration of diabetes or due to posttranslational modifications.89 Experimental depletion of tissue ␣-tocopherol results in depletion in GSH, increased lipid peroxidation, and the development of a sensory neuropathy in previously normal nerves and worsening of neuropathy in diabetic nerves.130 As a result, diabetic peripheral nerve has increased malondialdehyde, conjugated dienes (see Fig. 23–4), and lipid hydroperoxides33,130,162; the increase is more consistent in lumbar dorsal root and superior cervical ganglia.130,162 The most reliable index of increased oxidative stress is reduction in GSH.130,162 Finally, oxidative stress can cause neuropathy, because experimental depletion of tissue ␣-tocopherol results in a reduction in GSH, increased lipid peroxidation,

517

and the development de novo of a distal sensory neuropathy, and worsens EDN.130 Recent emphasis is on immunocytochemical labeling of products of oxidative stress that are known to be neurotoxic.81 Ischemia activates phospholipase A2 and the arachidonic acid cascade, resulting in superoxide anion and 4-HNE, known to cause neuronal apoptosis and mitochondrial damage.81,148 Additional support for a role of oxidative stress comes from improvement in neuropathy following treatment with probucol,18,73 vitamin E,77 and especially ␣-lipoic acid. With increasing age, there are myelin changes in ventral root.92 More recently, advanced changes in both dorsal and ventral roots have been described in EDN with a duration beyond 6 months.178 We demonstrated pathologic alteration in spinal roots and DRGs, showing demyelination, vacuolar degeneration, and pigmentary changes158 (Fig. 23–5). These observations have been furthered by recent pathophysiologic

FIGURE 23–5 Electron micrographs of representative ventral root fibers showing a progression of myelinopathy in experimental diabetes. A, Myelin decompaction. B, A rim of intact myelin surrounds degenerating myelin, with early myelin balls and a denuded atrophic axon. C, An atrophic axon is surrounded by myelin showing residual rims separated by myelin degeneration assuming a prominent honeycombed appearance. D, A completely demyelinated axon. (From Sasaki, H., Schmelzer, J. D., Zollman, P. J., and Low, P. A.: Neuropathology and blood flow of nerve, spinal roots and dorsal root ganglia in longstanding diabetic rats. Acta Neuropathol. 93:118, 1997, with permission.)

518

Neurobiology of the Peripheral Nervous System

studies supporting the sensory neuron as a target, in particular with the mitochondrion as the most relevant organelle. In a series of studies, the Michigan group has confirmed ballooning of mitochondria and disruption of the internal cristae as a result of the neurotoxic effects of glucose. They demonstrated apoptosis with sequential steps, including a reduction of mitochondrial membrane potential, leakage of cytochrome c, and caspase-3 activation.45,54,146 The molecular pathogenesis of diabetic neuropathy has focused on oxidative DNA damage to three targets. Most of the interest has focused on the endothelial cell and DRG, and to a lesser degree the Schwann cell.

Molecular Pathogenesis of Endothelial Dysfunction A simplified model of the role of oxidative stress in the pathogenesis of endothelial dysfunction and reduced nerve blood flow is shown in Figure 23–6. Nerve blood flow is impaired as a result of reduced action of vasodilators (including nitric oxide and prostacyclin) and enhanced activity of vasoconstrictors (including endothelin). The roles of endothelin, vasopressin, and prostaglandins are all important but are not covered here. Hyperglycemia in vivo and in vitro generates ROS and lipid peroxidation129 and single-strand DNA breaks. This in turn activates PARP, a profuse nuclear enzyme that has been implicated in response to DNA injury.172 PARP initiates an energy-consuming cycle by transferring ADP-ribose units from NAD⫹ to nuclear proteins, resulting in a depletion of the intracellular NAD⫹ and ATP pools, which impairs glycolysis and mitochondrial respiration leading cellular dysfunction.175 PARP-172 and polyol pathway–induced

FIGURE 23–6 Suggested pathogenesis of endothelial dysfunction.

reduction in NADPH results in reduced endothelial nitric oxide synthase (eNOS) activity, an observation that has consistently been found in diabetic nerves.84,112,200 NADPH depletion also results in a reduction in GSH and escalating oxidative stress. PARP is implicated in the process of pro-inflammatory gene expression.176 NF-␬B activation, intracellular adhesion molecule-1 (ICAM-1), tumor necrosis factor-␣ (TNF-␣), and nitric oxide synthase (NOS) have been implicated in the pathogenesis of diabetic endothelial dysfunction.5,120,149 Support for this notion derives from the observation that, in macrophages from PARP-deficient animals, there is a reduction in the activation of NF-␬B, with subsequent suppression of pro-inflammatory gene expression, ICAM-1,199 TNF-␣,144 and inducible NOS176 in PARP-deficient cells. Treatment with the PARP inhibitors reversed endothelial dysfunction.172 The pathogenesis of ROS-mediated endothelial dysfunction is shown in Figure 23–6. Hyperglycemia induces ROS, especially superoxide anion. ROS, especially hydroxyl and peroxynitrite radicals, cause DNA strand breaks and PARP generation. PARP activation depletes NAD⫹, NADPH, and ATP, resulting directly and indirectly (by reducing eNOS) in endothelial dysfunction. Hyperglycemia and PARP activate NF-␬B, resulting in pro-inflammatory mediators and endothelial dysfunction. The net effect of PARP activation, therefore, results in energy depletion, activation of adhesion molecules, an inflammatory response, NO depletion and endoneurial ischemia/hypoxia, and escalating oxidative stress.

Molecular Pathogenesis of Sensory Neuropathy: Pivotal Role of the Mitochondrion Diabetic neuropathy disproportionately affects sensory nerve fibers. The most consistent findings in chronic EDN in our hands is conduction slowing of distal sensory fibers.126 The most convincing pathologic alterations are in the nerve roots, with prominent myelin and axonal changes that develop in chronic (ⱖ6 months) experimental diabetes.158,178 These alterations in ventral root can occur in rats older than 12 months,92 although the changes in diabetes are more advanced and occur earlier; the changes in dorsal root are specific to the disorder. These findings, coupled with the recognition of the pivotal role of oxidative stress in diabetes and the high density of mitochondria in DRG neurons, led to the hypothesis that much of the sensory neuropathy is due to the effect of oxidative injury to the DRG and in particular the mitochondrion. The suggested pathogenesis is shown in Figure 23–7. Hyperglycemia results in oxidative stress by numerous pathways described earlier. Of particular importance to mitochondrial function is the reduction in GSH, reduction in NADPH, and increase in NADH. The generation of

Oxidative Stress and Excitatory Neurotoxins in Neuropathy

519

depletion of energy substrates blocks necrosis and commits the cell to apoptosis.71 We evaluated 8-hydroxydeoxyguanosine and caspase-3 expression using immunohistochemistry and terminal deoxynucleotidyl transferase–mediated dUTP nick end-labeling (TUNEL) staining in L5 DRGs of streptozotocin diabetic rats. Duration of diabetes was 1 month, 3 months, and 12 months. Oxidative injury began early and was sustained, with 12% of neurons showing TUNEL positivity at 12 months. Morphometric analysis showed a modest reduction of neurons, with a selective loss of 43% of the largest neurons.88

Molecular Pathogenesis of Schwann Cell Injury

FIGURE 23–7 Suggested pathogenesis of sensory neuropathy.

ROS, especially superoxide, and peroxynitrite results in single-strand breaks of DNA. DRG neuronal labeling with 8-hydroxydeoxyguanosine, indicative of oxidative DNA injury to the neuron, is consistently seen in chronic EDN (Fig. 23–8). Pivotal to the pathogenesis to apoptosis of DRG neurons is the release of cytochrome c from outer mitochondrial membrane. The egress of cytochrome c requires the opening of mitochondrial pores. There is good evidence for this event. Mitochondrial membrane potential falls in the hyperglycemic mitochondrion173 associated with swelling of the organelle. The precise mechanism of pore opening is not known. The most likely mechanisms are a combination of an alteration in the balance between the protective and pro-apoptotic components of the Bcl family and a reduction in GSH.70 Cytochrome c release has been demonstrated,173 and this binds to Apaf 1, which then recruits procaspase-9 (aptosome complex).59 This oligomerization and autoactivation of procaspase-9 to caspase-9 is followed by proteolytic cleavage of procaspase-3 (and other executioner procaspases), activating the executioner caspase-3. It is consistently increased in experimental diabetes.45,156 Caspase-3 will activate PARP with several important consequences. It promotes the inflammatory response by increasing NF-␬B,59 depletes energy substrates, and starts the cycles of escalating oxidative injury. The net effect is apoptosis of DRG neurons (see Fig. 23–7). The

Limited information is available on this subject. Most of the studies on the Schwann cell have been subsidiary to studies on DRG neurons. Changes such as chromatin condensation and vacuolar mitochondrial changes have been described.37,146 These observations of Schwann cells in DRGs have been supplemented by in vitro studies. Overexpression of Bcl-XL, or insulin-like growth factor, signaling via phosphatidylinositol 3-kinase, was reported to protect Schwann cells from glucose-mediated apoptosis in vitro.37 Nukada and colleagues have studied morphologic alterations that occur in Schwann cells subjected to reperfusion following ischemic injury to nerve.140 Additional studies have been done on the unusual susceptibility of diabetic Schwann cells to ischemia and reperfusion injury. For convenience, these studies are described under Neuropathology in the next section.

OXIDATIVE STRESS RELATED TO NERVE ISCHEMIA Nerve Ischemia There have been special problems in the study of peripheral nerve ischemia. Unlike brain and heart, in which ischemic effects are readily produced, peripheral nerve is relatively resistant because of its low energy needs and extensive anastomoses.105 Thus early attempts to produce experimental ischemia were largely unsuccessful. Models of ischemia have, however, been subsequently developed using ligation or infusion of microemboli of appropriate size to occlude microvessels.42,62,93,146 To study reperfusion injury, we used the model of ligation of supplying arteries (and collaterals) for a specified duration, followed by release of ligatures for reperfusion.117,163 This model reliably permits the separate evaluation of the effects of ischemia and those of reperfusion. The effects of ischemia and reperfusion have been studied on the blood-nerve barrier (BNB) and then on endoneurial contents.

520

Neurobiology of the Peripheral Nervous System

Diabetic

TUNEL

8-OHdG

Control

FIGURE 23–8 L5 dorsal root ganglion neuron of chronic experimental diabetic neuropathy (duration 12 months) shows labeling with 8-hydroxydeoxyguanosine (top) and TUNEL positivity (bottom). Age-matched control neurons are negative. See Color Plate

The Blood-Nerve Barrier Ischemia results in depletion of energy metabolites, an increase in tissue reducing equivalents,104,128 and an increase in tissue xanthine.69 Ischemia activates a calciumdependent protease that converts cytosolic xanthine dehydrogenase to xanthine oxidase6 in capillary endothelium.67 The above-mentioned conditions create an oxygen free radical–generating system. Furthermore, reperfusion accelerates the formation of ROS by introducing O2 into a system primed and generating ROS.113 ROS regularly increase microvascular permeability,36,94 and disruption of the BNB is an important first step in oxidative injury to nerve. Reperfusion-mediated BNB breakdown was clearly shown in direct measurements of permeability of rat sciatic nerve. We found that ischemia alone for 1 hour did not cause BNB alteration, but reperfusion resulted in reduced reflow and an increase in the permeability coefficient, indicating disruption of the interface.163 We measured both nerve blood flow and the permeability–surface area product. The permeability–surface area product may be increased as a result of an increase in permeability or an increase in the surface area (recruitment of capillaries). In this study, nerve blood flow was found to be persistently reduced with reperfusion. Therefore, the progressive increase in the permeability–surface area product must

be due to an increase in permeability coefficient. Reduced reflow is due to swelling of endothelial cells and pericytes with adhesion of leukocytes,139 resulting from activation of adhesion molecules, especially ICAM-1.140 Adrenergic Contribution Ischemia results in the generation of ROS from several sources. There is release of norepinephrine from sympathetic nerve terminals.12 Catecholamine release results in the increased synthesis and release of prostacyclin metabolites49,52 by ␣ receptor–mediated and calcium/calmodulindependent mechanisms. Calmodulin is known to activate phospholipase A2,192 resulting in a breakdown of membrane phospholipids and activation of the arachidonic acid cascade,15 generating leukotrienes (further damaging endothelial cells) and prostaglandins. Prostacyclin, localized in vascular endothelial cells,106 is the major vasodilator and inhibitor of platelet aggregation described, acting by stimulating platelet adenylate cyclase.51 Prostacyclin synthetase is rich in vascular endothelial cells119,187 and inhibited by lipid endoperoxides.118 The ratio of prostacyclin to thromboxane A2 is considered to be important in the maintenance of vascular tone118 and is reduced in ischemia-reperfusion and experimental diabetes.101

Oxidative Stress and Excitatory Neurotoxins in Neuropathy

Endoneurial Contents The effects of oxidative stress depend on a balance among the severity of oxidative stress, pro-oxidant status, and free radical defenses. Free radical defenses include the enzymatic scavengers, especially SOD, glutathione peroxidase, glutathione reductase, GSH, ascorbate, and ␣-tocopherol. We and others have demonstrated changes in all three areas in ischemic peripheral nerve. An increase in free radical generation occurs in ischemia-reperfusion, indicated by the breakdown of the blood-nerve interface,163 endoneurial edema, increase in hydroperoxides, and ischemic fiber degeneration.127 Altered pro-oxidant status occurs with an increase in polyunsaturated fatty acids, increasing arachidonic acid.196 Free radical defenses in nerve are selectively reduced. GSH and glutathione-containing enzymatic scavengers are only about 10% that of brain.153 With ischemia, superoxide anion is converted to H2O2, but its further decomposition, mediated by glutathione peroxidase,26 may be compromised if the low content of this enzyme is further reduced. 4-HNE, an aldehydic product of membrane lipid peroxidation, is especially pertinent in that ischemia, by increasing arachidonic acid, provides substrate for HNE,80 which mediates neuronal apoptosis.95 GSH is an especially sensitive indicator of ongoing oxidative stress.126 Similarly, Wagner et al.183 reported that the frequency of fiber degeneration is inversely proportional to GSH. ␣-Lipoic acid, a powerful lipophilic ROS scavenger,145 normalized the perfusion and conduction deficits in experimental diabetes126 and neuroprotected nerve subjected to ischemia-reperfusion.115 Lipid peroxidation generates HNE, which selectively inhibits ␣-ketoglutarate and pyruvate dehydrogenases and depletes these two multienzyme complexes of ␣-lipoic acid,65 rendering treatment with this antioxidant particularly relevant.

Reperfusion Injury The effects of ischemia in several tissues are amplified during reperfusion, a phenomenon referred to as reperfusion injury.57,113 Applied to peripheral nerves, there are a number of important observations. Peripheral nerve has low energy requirements and extensive collateral flow. One implication is its resistance to both ischemia and reperfusion injury. We demonstrated that reperfusion injury does occur in nerve, but only if the ischemia is near total and lasts at least 60 minutes.34,163 Ischemia followed by reperfusion results in disruption of the BNB and the activation of an immediate response involving adrenergic and protease activation, followed later by ROS associated with the inflammatory response and cytokines. The role of norepinephrine has been described earlier, and the inflammatory and cytokine responses are described later. Reperfusion following ischemia results in protease activation and a massive increase in intracellular calcium,25 leading to calcium-mediated activation of phospholipases and the production of free fatty acids and lysophospholipids.

521

There is some indirect evidence that similar ischemic mechanisms are operative in peripheral nerve. Calcium ionophore will cause vesicular disruption of myelin161,169 and axonal degeneration,160 and nerve reconnection in a calcium-free medium resulted in improved functional recovery. 35 The vasoconstriction and microvascular ischemia/hypoxia in disorders such as diabetes have been in part ascribed to perturbations of prostaglandins and ROS generation. Lipid hydroperoxides are increased and inhibit prostacyclin synthetase activity, resulting in a reduced prostacyclin:thromboxane ratio, and vasoconstriction and platelet aggregation.198 Footprints of Oxidative Injury Further support for the role of ROS comes from measurement of footprints of oxidative stress. We evaluated the effect of ischemia caused by the ligation of the supplying arteries to sciatic and tibial nerve for 3 hours, followed by reperfusion.127 Reperfusion resulted in an increase in nerve lipid hydroperoxides, greatest at 3 hours, followed by a gradual decline over the next month. Nerve edema and ischemic fiber degeneration consistently became more severe with reperfusion, indicating that oxidative stress impairs the BNB, resulting in edema, and causes ischemic fiber degeneration. Reduced reperfusion was greatest over distal sciatic nerve and mid-tibial nerve at day 7. The most ischemic segment (mid-tibial) of non-reperfused ischemic nerves (duration of ischemia, 3 hours) underwent both edema and ischemic fiber degeneration that was as pronounced as those of other segments after reperfusion, and underwent a smaller increase with reperfusion, suggesting that ischemia alone can also cause ischemic fiber degeneration and edema. The type of fiber degeneration was that of axonal degeneration. These indices indicate the presence of oxidative stress but do not indicate oxidative injury. Hence, recent focus has shifted to immunohistochemical indices that provide cellular localization and indicate DNA damage. These findings are detailed next.

Neuropathology The neuropathology of ischemic fiber degeneration has been well described.86,127,137 We have related the severity of nerve ischemic fiber degeneration to that of ischemic stress (dose of emboli or number of vessels ligated).86 The severity of fiber degeneration regresses linearly with the severity of ischemia (reduction in nerve blood flow).86 Ischemic demyelination is well recognized, but its frequency and pathophysiology have been largely unknown. Its presence has been emphasized by Nukada and colleagues,137,139 especially in the context of ischemia-reperfusion, with a perivascular distribution and in association with endoneurial edema, intramyelinic edema, and endothelial swelling.139 It has also been noted by others.62,127 Nukada and Dyck138 provided convincing evidence, based on a serial-section reconstruction of 55 single myelinated fibers showing ischemic fiber

522

Neurobiology of the Peripheral Nervous System

degeneration, that demyelination can be secondary to changes in axonal caliber. However, Schwann cell and myelin pathology, with intramyelinic edema, suggests a separate effect of oxidative injury associated with ischemia, and especially with ischemia followed by reperfusion, on the Schwann cell.139 Ischemic demyelination takes on greater importance with evidence of ischemia-reperfusion–induced Schwann cell oxidative injury, activated caspase-3 expression, and apoptosis. This mechanism of injury is especially pronounced in the chronically diabetic nerve, in which oxidative stress is already increased. Hyperglycemic nerves, although resistant to ischemic conduction failure,104 are, if ischemic stress is maintained, excessively susceptible to ischemic fiber degeneration.135,136 The resistance to ischemic conduction failure is due to a combination of low metabolic rate and increased energy substrate stores, so that anaerobic metabolism could provide sufficient ATP to maintain conduction.104 In an elegant study of rat peripheral nerve, intracellular H⫹ and calcium were measured.183 Nerve was rendered hypoxic for 3 to 6 hours in a 25- or 5-mM glucose medium. These authors found that hyperglycemia resulted in intracellular acidosis and an increase in intracellular glucose. Similar intracellular calcium levels were found in normal and highglucose media. The investigators concluded that ischemia increased intracellular calcium, and hyperglycemia compounded the ischemic insult with intracellular acidosis. A high-glucose medium causes apoptosis of DRG neurons; in vivo confirmation has been reported.155 Nukada made the observation that diabetic nerve is unusually susceptible to ischemia, but only after a duration of diabetes of at least 4 months.135,136 This mechanism is especially pronounced if diabetic nerves are subjected to ischemia-reperfusion. Figure 23–9 shows results in rats with EDN (duration of diabetes, 4 months), and age- and gender-matched controls, subjected to nerve ischemia for 3 hours followed by reperfusion. There was a lack of tibial nerve pathology in controls subjected to ischemic stress, whereas corresponding diabetic nerve had severe axonal pathology. The lack of caspase-3 activation in controls but prominent Schwann cell caspase-3 activation in diabetic nerve indicates that the Schwann cell is committed to the efferent limb of the apoptotic pathway. Finally, there was prominent Schwann cell expression of 8-hydroxydeoxyguanosine and 4-HNE in diabetic Schwann cells, indicative of oxidative DNA damage. These observations taken together lead to the appreciation that ischemic demyelination is different pathophysiologically from ischemic fiber degeneration. There is some evidence that demyelination occurs secondary to the inflammatory response with injury from cytokines and ROS in concert with axon-glia interactions, including changes in axonal caliber. Some of these changes are described next. Histochemical detection of oxidative stress in ischemic nerve was undertaken by Anderson et al.2 These workers labeled carbonyl compounds by applying naphthoic acid hydrazide (NAH) and Fast Blue B (FBB) and studied sciatic,

tibial, and peroneal nerves. NAH-FBB reactivity was not seen in ischemic nerve but appeared in vessels of nerves subjected to reperfusion following ischemia. Positively stained epi-, peri-, and endoneurial vessels were invariably observed after 2 hours of reperfusion at all levels examined. Pretreatment with ␣-tocopherol prevented NAH-FBB staining. In a subsequent study, He et al.56 measured carbonyl formation using a sensitive enzymelinked immunosorbent assay. Protein carbonyl content was unaffected by ischemia alone, but increased by 55% after 12 to 18 hours of reperfusion, correlating with the onset of nerve pathology. Morphometry in endoneurial vessels showed an increase in endothelial cell area that was prevented in ␣-tocopherol–treated reperfused nerves. Ischemia-reperfusion results in ICAM-1 expression on endothelial cells, followed by endoneurial edema and the inflammatory response with demyelination.140

Inflammatory Response Recently Nukada et al.140 studied the inflammatory response following total ischemia of 5 hours’ duration and varying durations of reperfusion, using immunohistochemistry, light, and electron microscopy. They evaluated the role of acute inflammation in the development of demyelination in reperfused rat sciatic, tibial, and peroneal nerves. They studied the time course of expression of ICAM-1, polymorphonuclear neutrophils, and macrophage migration (Fig. 23–10). After 18 hours of reperfusion, there was maximal ICAM-1 expression on endoneurial vessels, and polymorphonuclear neutrophil accumulation was then prominent, reaching a peak 24 hours after reperfusion (Fig. 23–10). Endoneurial mononuclear macrophages increased nearly fourfold after 48 to 72 hours of reperfusion. Macrophages were observed invading Schwann cells and myelinated lamellae, with associated demyelination. They concluded that the macrophage was most closely associated with demyelination after reperfusion.

Role of Cytokines Associated with the inflammatory response is the activation of cytokines, and much of the oxidative injury is due to the role of cytokines. Recently, we demonstrated that ischemia induced by ischemia-reperfusion114,127 results in an increase in gene expression of TNF-␣, interleukin (IL)–1␤, IL-6, and IL-10 (Fig. 23–11). In the studies of ischemia-reperfusion of nerve, we found that TNF-␣ and IL-1␤ expression occurred in ischemic-reperfused nerves but not following ischemia alone. Furthermore, their expression required ischemia of sufficient severity to cause ischemic fiber degeneration and demyelination. These cytokines, derived mainly from macrophages, reflect a hierarchy with TNF-␣ and IL-1 assuming particular importance.46 Cytokine message and protein expression in rat peripheral nerve is reported to be

Oxidative Stress and Excitatory Neurotoxins in Neuropathy

Control

523

Diabetic

FIGURE 23–9 Experimental diabetic neuropathy model (duration 12 months) subjected to moderate nerve ischemia for 3 hours followed by 7 days of reperfusion. Note the lack of tibial nerve pathology in controls subjected to the same ischemic stress (top left), whereas corresponding diabetic nerve has severe axonal pathology. Top right, Lack of caspase-3 activation in controls but prominent Schwann cell caspase-3 activation in diabetic nerve. Bottom, Prominent Schwann cell expression of 8-hydroxydeoxyguanosine (left) and 4-hydroxynonenal (right) in diabetic Schwann cell, indicative of oxidative damage. See Color Plate

concordant (e.g., Wagner and Myers185). Some cytokines, especially TNF-␣, will activate adhesion molecules and leukocyte adherence.121 These cytokines are functionally interactive but have specific functions. For instance, IL-10 is neuroprotective.63 Administering a single dose of IL-10 at the site of a chronic constriction injury to nerve attenuated the inflammatory response and hyperalgesia.183 In IL-10–treated nerves, macrophage recruitment and cell profiles immunoreactive for TNF-␣ were also reduced. IL-6 is reported to be deleterious but is also, at least in part, neuroprotective in cerebral ischemia.98 Subperineurial administration of TNF-␣ will result in both demyelination

and ischemic fiber degeneration, and hyperalgesia.184 Schwann cells play an important role in the inflammatory response; they express major histocompatibility complex (MHC) class I and II, TNF-␣, and IL-6 mRNA.124

Genetic Susceptibility There has been no attention to the importance of genetic susceptibility. Ischemic susceptibility may be dependent on MHC class of the mouse, which influences host inflammatory response. MHC class II region and CD4⫹ T cells may synergize with other genes to preferentially produce high

524

Neurobiology of the Peripheral Nervous System

release of cytochrome c,91 which binds to Apaf-1, triggering activation of caspase-9,167 which activates downstream caspases, especially caspase-3, the principal effector in the apoptotic pathway in neurons and support cells.79 This hypothesis takes on special importance with the availability of potent antioxidants, anticaspases, and anticytokine treatment. If the inflammatory response is important, the severity of demyelination should be greater with ischemiareperfusion than ischemia alone.

Neuroprotection of Ischemia-Reperfusion Injury

FIGURE 23–10 Time course of densities (mean ⫾ 6 SD) of intercellular adhesion molecule-1 (ICAM-1)–positive vessels (top), polymorphonuclear neutrophils expressed by HIS48 (middle), and mononuclear macrophages visualized by 1C7 (bottom) in the endoneurium of ischemic-reperfused rat sciatic nerve at the thigh level.

levels of pro-inflammatory cytokines such as TNF-␣.142 Differences in genetic background with different cytokine responses could result in different susceptibilities to ischemia-reperfusion–related oxidative injury.

Molecular Pathogenesis On the efferent limb, caspases have been demonstrated to be important in delayed neuronal death in cerebral ischemia.24 Ischemia is proposed as being especially damaging to mitochondria-rich cells such as Schwann cells, causing mitochondrial membrane depolarization with

Mitsui et al.116 undertook a detailed study of hypothermic neuroprotection of ischemia-reperfusion injury of nerve. They compared an ischemia duration of 3 versus 5 hours and temperatures of 37°, 32°, and 28° C. Electrophysiologic, behavioral, and histologic end points (edema grade and ischemic fiber degeneration grade) were evaluated. The groups treated at 37° C underwent marked fiber degeneration, associated with a reduction in action potential and impairment in behavioral score. The groups treated at 28° C (for both 3 and 5 hours) showed significantly less (P ⫽ .01; analysis of variance, Bonferoni post hoc test) reperfusion injury for all indices (behavioral score, electrophysiology, and neuropathology), and the groups treated at 32° C had scores intermediate between the groups treated at 36° to 37° C and at 28° C. Figure 23–12 shows the graded neuroprotection as a function of degree of hypothermia. Our results showed that cooling the limbs dramatically protects the peripheral nerve from ischemia-reperfusion injury. Antioxidant pretreatment will prevent histochemical evidence of oxidative injury. ␣-Tocopherol pretreatment resulted in less carbonyl staining (as an index of oxidative injury) and the area of endothelial cells was also reduced, indicating amelioration of endothelial swelling.2 Pretreatment with the xanthine oxidase inhibitor allopurinol prevented this protein carbonyl formation.56 In a more formal study, Mitsui et al.115 evaluated whether racemic ␣-lipoic acid, a potent antioxidant, would protect peripheral nerve from reperfusion injury. ␣-Lipoic acid was given intraperitoneally daily for 3 days both pre- and postsurgery. Distal sensory conduction (amplitude of sensory action potential and sensory conduction velocity of digital nerve) was significantly improved in the 3-hour ischemia group treated with ␣-lipoic acid (P ⬍ .05) but not if the duration of ischemia was longer (5-hour ischemia). Figure 23–13 shows the effects of lipoic acid–mediated neuroprotection.

N-METHYL-D-ASPARTATE AND GLUTAMATE RECEPTORS Most of the interest in N-methyl-D-aspartate (NMDA) receptors has revolved around their contribution to the pain state. NMDA receptors are present in several sites.

525

Oxidative Stress and Excitatory Neurotoxins in Neuropathy

A: TNF-α 3

Sciatic Nerve

Arbitrary units

Tibial Nerve

β

2

β

α α

α

1

α

0 SHAM

3I

3I3R

3I12R

3I24R

24I

3I48R

3I7dR

B: IL1-β 30

Sciatic Nerve Tibial Nerve

20

γ

10

α

α

α

β

α

0 SHAM

3I

3I3R

3I12R

3I24R

24I

3I48R

3I7dR

FIGURE 23–11 Time course of messenger RNA (mRNA) expression of tumor necrosis factor-␣ (TNF-␣) and interleukin-1␤ (IL-1␤). mRNA expression of TNF-␣ peaked at 24 hours of reperfusion and remained elevated even at 7 days. In contrast, IL-1␤ peaked at 12 hours of reperfusion. Tibial versus sciatic nerves: *, P ⬍ .05; ***, P ⬍ .001; tibial nerves versus sham: ␣, P ⬍ .05; ␤, P ⬍ .01; ␥, P ⬍ .001. (From Mitsui, Y., Okamoto, K., Martin, D. P., et al.: The expression of proinflammatory cytokine mRNA in the sciatic-tibial nerve of ischemia-reperfusion injury. Brain Res. 844:192, 1999, with permission.)

These include the nerve root entry zone, primary afferent endings, and sympathetic efferents.23 These receptors appear to be vitally involved in the pathophysiology of painfulness.38,194 Three closely interlinked sites of sensitization (nerve axon microenvironment, DRG, and nerve root entry zone) may together be operative in producing hyperalgesia. There is interest in the effect of inflammation on glutamate receptors. Carlton and Coggeshall23 demonstrated that the number of peripheral primary afferent axons expressing NMDA, ␣-aminohydroxy-5-methyl4-isoxazolepropionic acid (AMPA), or kainate inotropic glutamate receptors were significantly increased during complete Freund’s adjuvant–induced inflammation of the hind paw. This suggests an increased role for glutamate in the pain that accompanies inflammation. Glutamate may also regulate sympathetic efferents locally because inotropic receptors are expressed by peripheral postganglionic sympathetic fibers.22 In a subsequent paper, Carlton and Coggeshall23 demonstrated that, following inflammation, there was a 2-fold increase in postgan-

glionic adrenergic axons expressing NMDA receptors and a 10-fold increase in axons expressing AMPA or kainate receptors. These data suggest that postganglionic activity may be enhanced by glutamate receptor activation during inflammation. Increased activity in postganglionic fibers could lead to an increased release of norepinephrine and other substances in postganglionic efferents, such as prostaglandins, which in turn could enhance nociceptor activity. This change in glutamate receptor organization offers a possible site of pharmacologic intervention for the maladaptive symptoms that often arise following peripheral inflammation. The chronic constriction model results in increased norepinephrine, neuropeptide Y (NPY), and calcitonin gene–related peptide (CGRP) at the nerve lesion. Nerve blood flow is increased within the lesion but reduced at the lesion edge by more than 50%, a deficit that is sufficient to cause fiber degeneration.86,166 This increase in CGRP in the lesion is associated with reduced pain threshold. Second, we found a large increase of NPY and CGRP by radioimmunoassay and immunocytochemistry

526

Neurobiology of the Peripheral Nervous System

37˚C

32˚C

28˚C

FIGURE 23–12 Representative transverse sections of the mid-tibial nerve after 3 hours of ischemia-reperfusion. Most fibers degenerated when treated at 37° C; fewer degenerated fibers were seen after treatment at 32° C. Most fibers were normal after treatment at 28° C. Bar: 100 ␮m. (From Mitsui, Y., Schmelzer, J. D., Zollman, P. J., et al.: Hypothermic neuroprotection of peripheral nerve of rats from ischaemia-reperfusion injury. Brain 122:161, 1999, with permission.)

Isch

Isch- LA

FIGURE 23–13 Light microscopic findings of transverse sections of the mid-tibial nerve after 3 hours of ischemia followed by 1-week reperfusion. Degenerated fibers were predominant for the nontreatment group, and a small number of degenerated fibers were seen in the lipoic acid group. Bar: 100 ␮m. Isch ⫽ ischemiareperfusion; Isch-LA ⫽ ischemia-reperfusion ⫹ lipoic acid treatment. (From Mitsui, Y., Schmelzer, J. D., Zollman, P. J., et al.: Alphalipoic acid provides neuroprotection from ischemia-reperfusion injury of peripheral nerve. J. Neurol. Sci. 163:11, 1999, with permission.)

Oxidative Stress and Excitatory Neurotoxins in Neuropathy

in ipsilateral L5 DRGs but not in sympathetic ganglia. For instance, central sensitization requires peripheral inputs and can be prevented if peripheral input is blocked or absent.38 Rats with a chronic constriction injury to the sciatic nerve have been found to have small- to medium-sized, pyknotic, and hyperchromatic neurons (“dark neurons”) in spinal dorsal horn laminae I through III.125 These authors proposed that these hyperchromatic neurons are produced by an excitotoxic insult involving NMDA receptor activation subsequent to ectopic nociceptor discharge, and that at least some dark neurons are inhibitory interneurons whose functional impairment or death contributes to a central state of hyperexcitability that underlies neuropathic hyperalgesia and allodynia. NMDA antagonists will prevent or reverse allodynia.168,177 Pain models include the chronic constriction nerve trunk model of Bennett and Xie,9 the root constriction model of Kim and Chung,87 and the paw formalin injection model. A sustained afferent barrage from small nerve fibers is associated with spinal release of glutamate, prostaglandins, and aspartate,107 and results in a hyperalgesic state. There are close interactions among spinal NMDA, neurokinin-1 (NK), and ␣ adrenoreceptors. There is also NO modulation and mediation by the spinal release of cyclooxygenase products.195 The hyperalgesic state can be prevented by antinociceptive doses of cyclooxygenase inhibitors,108 the ␣2 adrenoceptor agonist clonidine, or NMDA antagonists (such as MK-801). Intrathecal clonidine may act to diminish sympathetic outflow, whereas MK-801 blocks the NMDA receptor. The two separate mechanisms may account for the powerful synergy observed by Lee and Yaksh.97 This hyperalgesic component appears to be initiated by the activation of a spinal NMDA receptor that, through the generation of NO, leads to the observed augmented processing of afferent input and the associated hyperalgesic component of the subsequent pain behavior.107 Hyperalgesia is prevented by pretreatment with agents known to block afferent input (local anesthetics) or C-fiber transmitter release (opiates) or to act at one of several links to block a complex spinal cascade involving the NMDA receptor, NOS, and cyclooxygenase.193 Neuropathic animals that have thermal hyperalgesia had an ipsilateral decrease in substance P (SP) staining density without an accompanying change in CGRP staining density. MK-801–treated animals showed a dose-dependent attenuation of the thermal hyperalgesia, and the expected ipsilateral decrease in SP was prevented. MK-801 treatment in naive rats caused a global increase in both SP and CGRP staining in the dorsal horn. The results suggest a functional interaction between excitatory amino acids (EAAs) and SP, with activation of NMDA receptors mediating depletion of SP in neuropathic animals. It is suggested that SP-containing interneurons are a target of the EAAs in the dorsal horn.47

527

Although NO-NMDA interactions can result in hyperalgesia, NO also produces a dose-dependent antinociceptive response in diabetic mice.74 The antinociceptive effects were significantly antagonized by subcutaneous administration of naltrindole, a selective ␦ opioid receptor antagonist, suggesting the involvement of activation of ␦ opioid receptors. The hyperalgesic state in experimental diabetic rats is associated with increased spinal release,76 reduced spinal levels of SP, and a significant increase in the number of binding sites for SP in dorsal spinal cord.75 RP-67580, a specific tachykinin NK1 receptor antagonist, relieves chronic hyperalgesia in diabetic rats.31

REFERENCES 1. Ahlskog, J. E., Uitti, R. J., Low, P. A., et al.: No evidence for systemic oxidant stress in Parkinson’s or Alzheimer’s disease. Mov. Disord. 10:566, 1995. 2. Anderson, G. M., Nukada, H., and McMorran, P. D.: Carbonyl histochemistry in rat reperfusion nerve injury. Brain Res. 772:156, 1997. 3. Babiy, A. V., Gebicki, J. M., Sullivan, D. R., and Willey, K.: Increased oxidizability of plasma lipoproteins in diabetic patients can be decreased by probucol therapy and is not due to glycation. Biochem. Pharmacol. 43:995, 1992. 4. Baker, M. S., and Gebicki, J. M.: The effect of pH on the conversion of superoxide to hydroxyl free radicals. Arch. Biochem. Biophys. 234:258, 1984. 5. Barouch, F. C., Miyamoto, K., Allport, J. R., et al.: Integrin-mediated neutrophil adhesion and retinal leukostasis in diabetes. Invest. Ophthalmol. Vis. Sci. 41:1153, 2000. 6. Battelli, M. G., Corte, E. D., and Stirpe, F.: Xanthine oxidase type d (dehydrogenase) in the intestine and other organs of the rat. Biochem. J. 126:747, 1972. 7. Baynes, J. W.: Role of oxidative stress in the development of complications in diabetes. Diabetes 40:405, 1991. 8. Benarroch, E. E., Opfer-Gehrking, T. L., and Low, P. A.: Use of the photoplethysmographic technique to analyze the Valsalva maneuver in normal man. Muscle Nerve 14:1165, 1991. 9. Bennett, G. J. and Xie, Y. K.: A peripheral mononeuropathy in rat that produces disorders of pain sensation like those seen in man. Pain 33:87, 1988. 10. Bhimji, S., Godin, D. V., and McNeill, J. H.: Insulin reversal of biochemical changes in hearts of diabetic rats. Am. J. Physiol. 251:H670, 1986. 11. Boveris, A., Cardenas, E., and Stoppani, A. O.: Role of ubiquinone in the mitochondrial generation of hydrogen peroxide. Biochem. J. 156:435, 1976. 12. Boveris, A., and Chance, B.: The mitochondrial generation of hydrogen peroxide: general properties and effect of hyperbaric oxygen. Biochem. J. 134:707, 1973. 13. Bravenboer, B., Kappelle, A. C., Hamers, F. P., et al.: Potential use of glutathione for the prevention and treatment of diabetic neuropathy in the streptozotocin-induced diabetic rat. Diabetologia 35:813, 1992.

528

Neurobiology of the Peripheral Nervous System

14. Bucala, R., Tracey, K. J., and Cerami, A.: Advanced glycosylation products quench nitric oxide and mediate defective endothelium-dependent vasodilatation in experimental diabetes. J. Clin. Invest. 87:432, 1991. 15. Burton, N. K., and Aherne, G. W.: Sensitive measurement of glutathione using isocratic high-performance liquid chromatography with fluorescence detection. J. Chromatogr. 382:253, 1986. 16. Busse, E., Zimmer, G., Schopohl, B., and Kornhuber, B.: Influence of alpha-lipoic acid on intracellular glutathione in vitro and in vivo. Arzneimittelforschung 42:829, 1992. 17. Cameron, N. E., and Cotter, M. A.: Neurovascular dysfunction in diabetic rats: potential contribution of autoxidation and free radicals examined using transition metal chelating agents. J. Clin. Invest. 96:1159, 1995. 18. Cameron, N. E., Cotter, M. A., Archibald, V., et al.: Antioxidant and pro-oxidant effects on nerve conduction velocity, endoneurial blood flow and oxygen tension in non-diabetic and streptozotocin-diabetic rats. Diabetologia 37:449, 1994. 19. Cameron, N. E., Cotter, M. A., Jack, A. M., et al.: Protein kinase C effects on nerve function, perfusion, Na( ⫹ ), K( ⫹ )ATPase activity and glutathione content in diabetic rats. Diabetologia 42:1120, 1999. 20. Cameron, N. E., Cotter, M. A., and Low, P. A.: Nerve blood flow in early experimental diabetes in rats: relation to conduction deficits. Am. J. Physiol. 261:E1, 1991. 21. Cameron, N. E., Cotter, M. A., and Maxfield, E. K.: Antioxidant treatment prevents the development of peripheral nerve dysfunction in streptozotocin-diabetic rats. Diabetologia 36:299, 1993. 22. Carlton, S. M., Chung, K., Ding, Z., and Coggeshall, R. E.: Glutamate receptors on postganglionic sympathetic axons. Neuroscience 83:601, 1998. 23. Carlton, S. M., and Coggeshall, R. E.: Inflammation-induced changes in peripheral glutamate receptor populations. Brain Res. 820:63, 1999. 24. Chen, J., Nagayama, T., Jin, K., et al.: Induction of caspase3-like protease may mediate delayed neuronal death in the hippocampus after transient cerebral ischemia. J. Neurosci. 18:4914, 1998. 25. Cheung, J. Y., Bonventre, J. V., Malis, C. D., and Leaf, A.: Calcium and ischemic injury. N. Engl. J. Med. 314:1670, 1986. 26. Cohen, G., and Hochstein, P.: Glutathione peroxidase: the primary agent for the elimination of hydrogen peroxide in erythrocytes. Biochemistry 2:1420, 1963. 27. Collier, A., Leach, J. P., McLellan, A., et al.: Plasma endothelin-like immunoreactivity levels in IDM patients with microalbuminuria. Diabetes Care 15:1038, 1992. 28. Cotter, M. A., and Cameron, N. E.: Correction of impaired sciatic nerve perfusion by desferrioxamine in diabetic rats. Diabetologia 38:898, 1995. 29. Cotter, M. A., and Cameron, N. E.: Neuroprotective effects of carvedilol in diabetic rats: prevention of defective peripheral nerve perfusion and conduction velocity. Naunyn Schmiedebergs Arch. Pharmacol. 351:630, 1995. 30. Cotter, M. A., Love, A., Watt, M. J., et al.: Effects of natural free radical scavengers on peripheral nerve and neurovascular function in diabetic rats. Diabetologia 38:1285, 1995.

31. Courteix, C., Lavarenne, J., and Eschalier, A.: RP-67580, a specific tachykinin NK1 receptor antagonist, relieves chronic hyperalgesia in diabetic rats. Eur. J. Pharmacol. 241:267, 1993. 32. Crouch, R., Kimsey, G., Priest, D. G., et al.: Effect of streptozotocin on erythrocyte and retinal superoxide dismutase. Diabetologia 15:53, 1978. 33. Day, T. J., Lagerlund, T. D., and Low, P. A.: Analysis of H2 clearance curves used to measure blood flow in rat sciatic nerve. J. Physiol. (Lond.) 414:35, 1989. 34. Day, T. J., Schmelzer, J. D., and Low, P. A.: Aortic occlusion and reperfusion and conduction, blood flow, and the bloodnerve barrier of rat sciatic nerve. Exp. Neurol. 103:173, 1989. 35. de Medinaceli, L., Wyatt, R. J., and Freed, W. J.: Peripheral nerve reconnection: mechanical, thermal, and ionic conditions that promote the return of function. Exp. Neurol. 81:469, 1983. 36. Del Maestro, R. F., Bjork, J., and Arfors, K. E.: Increase in microvascular permeability induced by enzymatically generated free radicals. II. Role of superoxide anion radical, hydrogen peroxide, and hydroxyl radical. Microvasc. Res. 22:255, 1981. 37. Delaney, C. L., Russell, J. W., Cheng, H. L., and Feldman, E. L.: Insulin-like growth factor-I and over-expression of Bcl-xL prevent glucose-mediated apoptosis in Schwann cells. J. Neuropathol. Exp. Neurol. 60:147, 2001. 38. Dickenson, A. H.: NMDA receptor antagonists as analgesics. In Fields, H. L., and Liebeskind, J. C. (eds.): Progress in Pain Research and Management, Vol. 1. Seattle, IASP Press, p. 173, 1994. 39. Dietrich, R. B., and Bradley, W. G. Jr.: Iron accumulation in the basal ganglia following severe ischemic-anoxic insults in children. Radiology 168:203, 1988. 40. Dimpfel, W., Spuler, M., Pierau, F. K., and Ulrich, H.: Thioctic acid induces dose-dependent sprouting of neurites in cultured rat neuroblastoma cells. Dev. Pharmacol. Ther. 14:193, 1990. 41. Duchen, L. W., Anjorin, A., Watkins, P. J., and Mackay, J. D.: Pathology of autonomic neuropathy in diabetes mellitus. Ann. Intern. Med. 92:301, 1980. 42. Dyck, P. J., Karnes, J., O’Brien, P., et al.: Spatial pattern of nerve fiber abnormality indicative of pathologic mechanisms. Am. J. Pathol. 117:225, 1984. 43. Dyck, P. J., Low, P. A., Windebank, A. J., et al.: Plasma exchange in polyneuropathy associated with monoclonal gammopathy of undetermined significance. N. Engl. J. Med. 325:1482, 1991. 44. Dyck, P. J. B., Norell, J. E., and Dyck, P. J.: Microvasculitis and ischemia in diabetic lumbosacral radiculoplexus neuropathy. Neurology 53:2113, 1999. 45. Feldman, E. L., Russell, J. W., Sullivan, K. A., and Golovoy, D.: New insights into the pathogenesis of diabetic neuropathy. Curr. Opin. Neurol. 12:553, 1999. 46. Firestein, G. S., and Zvaifler, N. J.: Anticytokine therapy in rheumatoid arthritis. N. Engl. J. Med. 337:195, 1997. 47. Garrison, C. J., Dougherty, P. M., and Carlton, S. M.: Quantitative analysis of substance P and calcitonin generelated peptide immunohistochemical staining in the dorsal horn of neuropathic MK-801-treated rats. Brain Res. 607:205, 1993.

Oxidative Stress and Excitatory Neurotoxins in Neuropathy 48. Gaudet, R. J., and Levine, L.: Transient cerebral ischemia and brain prostaglandins. Biochem. Biophys. Res. Commun. 86:893, 1979. 49. Gilmore, N., Vane, J. R., and Wyllie, J. H.: Prostaglandins released by the spleen. Nature 218:1135, 1968. 50. Godin, D. V., Wohaieb, S. A., Garnett, M. E., and Goumeniouk, A. D.: Antioxidant enzyme alterations in experimental and clinical diabetes. Mol. Cell. Biochem. 84:223, 1988. 51. Gorman, R. R., Bunting, S., and Miller, O. V.: Modulation of human platelet adenylate cyclase by prostacyclin (PGX). Prostaglandins 13:377, 1977. 52. Greenberg, R.: The neuronal origin of prostaglandin released from the rabbit portal vein in response to electrical stimulation. Br. J. Pharmacol. 63:79, 1978. 53. Greene, D. A., Cao, X., Van Huysen, C., and Obrosova, I.: Effects of DL-alpha-lipoic acid (LA) on diabetic nerve function, bioenergetics and antioxidant defense. In Abstracts of the Workshop on Oxidative Stress in Diabetes and Its Complications, Santa Barbara, CA, February 5–8, p. 80, 1998. 54. Greene, D. A., Stevens, M. J., Obrosova, I., and Feldman, E. L.: Glucose-induced oxidative stress and programmed cell death in diabetic neuropathy. Eur. J. Pharmacol. 375:217, 1999. 55. Greene, R. M., Winkelmann, R. K., Opfer-Gehrking, T. L., and Low, P. A.: Sweating patterns in atopic dermatitis patients. Arch. Dermatol. Res. 281:373, 1989. 56. He, K., Nukada, H., McMorran, P. D., and Murphy, M. P.: Protein carbonyl formation and tyrosine nitration as markers of oxidative damage during ischaemia-reperfusion injury to rat sciatic nerve. Neuroscience 94:909, 1999. 57. Hearse, D. J., Manning, A. S., Downey, J. M., and Yellon, D. M.: Xanthine oxidase: a critical mediator of myocardial injury during ischemia and reperfusion? Acta Physiol. Scand. Suppl. 548:65, 1986. 58. Hellweg, R., and Hartung, H. D.: Endogenous levels of nerve growth factor (NGF) are altered in experimental diabetes mellitus: a possible role for NGF in the pathogenesis of diabetic neuropathy. J. Neurosci. Res. 26:258, 1990. 59. Hengartner, M. O.: The biochemistry of apoptosis. Nature 407:770, 2000. 60. Herbst, U., Toborek, M., Kaiser, S., et al.: 4-Hydroxynonenal induces dysfunction and apoptosis of cultured endothelial cells. J. Cell. Physiol. 181:295, 1999. 61. Hermenegildo, C., Raya, A., Roma, J., and Romero, F. J.: Decreased glutathione peroxidase activity in sciatic nerve of alloxan-induced diabetic mice and its correlation with blood glucose levels. Neurochem. Res. 18:893, 1993. 62. Hess, K., Eames, R. A., Darveniza, P., and Gilliatt, R. W.: Acute ischaemic neuropathy in the rabbit. J. Neurol. Sci. 44:19, 1979. 63. Ho, A. S., and Moore, K. W.: Interleukin-10 and its receptor. Ther. Immunol. 1:173, 1994. 64. Hofmann, M. A., Schiekofer, S., Isermann, B., et al.: Peripheral blood mononuclear cells isolated from patients with diabetic nephropathy show increased activation of the oxidative-stress sensitive transcription factor NF-kappaB. Diabetologia 42:222, 1999. 65. Humphries, K. M., and Szweda, L. I.: Selective inactivation of alpha-ketoglutarate dehydrogenase and pyruvate

66.

67.

68.

69. 70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

529

dehydrogenase: reaction of lipoic acid with 4-hydroxy2-nonenal. Biochemistry 37:15835, 1998. Hunt, J. V., and Wolff, S. P.: Oxidative glycation and free radical production: a causal mechanism of diabetic complications. Free Radic. Res. Commun. 12–13:115, 1991. Jarasch, E. D., Grund, C., Bruder, G., et al.: Localization of xanthine oxidase in mammary-gland epithelium and capillary endothelium. Cell 25:67, 1981. Jennings, P. E., Chirico, S., Lunec, J., and Barnett, A. H.: Vitamin C metabolites and microangiopathy in diabetes mellitus. Diabetes Res. 6:151, 1987. Jennings, R. B., and Reimer, K. A.: Lethal myocardial ischemic injury. Am. J. Pathol. 102:241, 1981. Ji, C., Amarnath, V., Pietenpol, J. A., and Marnett, L. J.: 4-Hydroxynonenal induces apoptosis via caspase-3 activation and cytochrome c release. Chem. Res. Toxicol. 14:1090, 2001. Jiang, C., Wang, Z., Ganther, H., and Lu, J.: Caspases as key executors of methyl selenium-induced apoptosis (anoikis) of DU-145 prostate cancer cells. Cancer Res. 61:3062, 2001. Kaji, H., Kurasaki, M., Ito, K., et al.: Increased lipoperoxide value and glutathione peroxidase activity in blood plasma of type 2 (non-insulin dependent) diabetic women. Klin. Wochenschr. 63:765, 1985. Kalyanaraman, B., Darley-Usmar, V. M., Wood, J., et al.: Synergistic interaction between the probucol phenoxyl radical and ascorbic acid in inhibiting the oxidation of low density lipoprotein. J. Biol. Chem. 267:6789, 1992. Kamei, J., Iwamoto, Y., Misawa, M., et al.: Antinociceptive effect of L-arginine in diabetic mice. Eur. J. Pharmacol. 254:113, 1994. Kamei, J., Ogawa, M., and Kasuya, Y.: Development of supersensitivity to substance P in the spinal cord of the streptozotocin-induced diabetic rats. Pharmacol. Biochem. Behav. 35:473, 1990. Kamei, J., Ogawa, Y., Ohhashi, Y., and Kasuya, Y.: Alterations in the potassium-evoked release of substance P from the spinal cord of streptozotocin-induced diabetic rats in vitro. Gen. Pharmacol. 22:1093, 1991. Karasu, C., Dewhurst, M., Stevens, E. J., and Tomlinson, D. R.: Effects of anti-oxidant treatment on sciatic nerve dysfunction in streptozotocin-diabetic rats: comparison with essential fatty acids. Diabetologia 38:129, 1995. Karpen, C. W., Cataland, S., O’Dorisio, T. M., and Panganamala, R. V.: Interrelation of platelet vitamin E and thromboxane synthesis in type I diabetes mellitus. Diabetes 33:239, 1984. Keane, R. W., Srinivasan, A., Foster, L. M., et al.: Activation of CPP32 during apoptosis of neurons and astrocytes. J. Neurosci. Res. 48:168, 1997. Keller, J. N., Hanni, K. B., and Markesbery, W. R.: 4-Hydroxynonenal increases neuronal susceptibility to oxidative stress. J. Neurosci. Res. 58:823, 1999. Keller, J. N., Mark, R. J., Bruce, A. J., et al.: 4-Hydroxynonenal, an aldehydic product of membrane lipid peroxidation, impairs glutamate transport and mitochondrial function in synaptosomes. Neuroscience 80:685, 1997. Keller, J. N., and Mattson, M. P.: Roles of lipid peroxidation in modulation of cellular signaling pathways, cell dysfunction, and death in the nervous system. Rev. Neurosci. 9:105–116, 1998.

530

Neurobiology of the Peripheral Nervous System

83. Khechai, F., Ollivier, V., Bridey, F., et al.: Effect of advanced glycation end product-modified albumin on tissue factor expression by monocytes: role of oxidant stress and protein tyrosine kinase activation. Arterioscler. Thromb. Vasc. Biol. 17:2885–2890, 1997. 84. Kihara, M., and Low, P. A.: Impaired vasoreactivity to nitric oxide in experimental diabetic neuropathy. Exp. Neurol. 132:180–185, 1995. 85. Kihara, M., Weerasuriya, A., and Low, P. A.: Endoneurial blood flow in rat sciatic nerve during development. J. Physiol. (Lond.) 439:351–360, 1991. 86. Kihara, M., Zollman, P. J., Schmelzer, J. D., and Low, P. A.: The influence of dose of microspheres on nerve blood flow, electrophysiology, and fiber degeneration of rat peripheral nerve. Muscle Nerve 16:1383, 1993. 87. Kim, S. H., and Chung, J. M.: An experimental model for peripheral neuropathy produced by segmental spinal nerve ligation in the rat. Pain 50:355, 1992. 88. Kishi, M., Tanabe, J., Schmelzer, J. D., and Low, P. A.: Morphometry of dorsal root ganglion in chronic experimental diabetic neuropathy. Diabetes 51:819, 2002. 89. Kishi, Y., Nickander, K. K., Schmelzer, J. D., and Low, P. A.: Gene expression of antioxidant enzymes in experimental diabetic neuropathy. J. Peripher. Nerv. Syst. 5:3, 2000. 90. Kishi, Y., Schmelzer, J. D., Yao, J. K., et al.: ␣-Lipoic acid: effect on glucose uptake, sorbitol pathway, and energy metabolism in experimental diabetic neuropathy. Diabetes 48:2045, 1999. 91. Kluck, R. M., Bossy-Wetzel, E., Green, D. R., and Newmeyer, D. D.: The release of cytochrome c from mitochondria: a primary site for Bcl-2 regulation of apoptosis. Science 275:1132, 1997. 92. Knox, C. A., Kokmen, E., and Dyck, P. J.: Morphometric alteration of rat myelinated fibers with aging. J. Neuropathol. Exp. Neurol. 48:119, 1989. 93. Korthals, J. K., and Wisniewski, H. M.: Peripheral nerve ischemia. Part 1. Experimental model. J. Neurol. Sci. 24:65, 1975. 94. Korthuis, R. J., Granger, D. N., Townsley, M. I., and Taylor, A. E.: The role of oxygen-derived free radicals in ischemiainduced increases in canine skeletal muscle vascular permeability. Circ. Res. 57:599, 1985. 95. Kruman, I., Bruce-Keller, A. J., Bredesen, D., et al.: Evidence that 4-hydroxynonenal mediates oxidative stress-induced neuronal apoptosis. J. Neurosci. 17:5089, 1997. 96. Lee, A. Y., and Chung, S. S.: Contributions of polyol pathway to oxidative stress in diabetic cataract. FASEB J. 13:23, 1999. 97. Lee, Y. W., and Yaksh, T. L.: Analysis of drug interaction between intrathecal clonidine and MK-801 in peripheral neuropathic pain rat model. Anesthesiology 82:741, 1995. 98. Loddick, S. A., Turnbull, A. V., and Rothwell, N. J.: Cerebral interleukin-6 is neuroprotective during permanent focal cerebral ischemia in the rat. J. Cereb. Blood Flow Metab. 18:176, 1998. 99. Loven, D. P., and Oberley, L. W.: Free radicals, insulin action and diabetes. In Oberley, L. W. (ed.) Superoxide Dismutase, Vol. III. Disease States. Boca Raton, FL, CRC Press, p. 151, 1985.

100. Loven, D. P., Schedl, H. P., Oberley, L. W., et al.: Superoxide dismutase activity in the intestine of the streptozotocindiabetic rat. Endocrinology 111:737, 1982. 101. Low, P. A., Lagerlund, T. D., and McManis, P. G.: Nerve blood flow and oxygen delivery in normal, diabetic, and ischemic neuropathy. Int. Rev. Neurobiol. 31:355, 1989. 102. Low, P. A., and Nickander, K. K.: Oxygen free radical effects in sciatic nerve in experimental diabetes. Diabetes 40:873, 1991. 103. Low, P. A., Schmelzer, J. D., Ward, K. K., et al.: Effect of hyperbaric oxygenation on normal and chronic streptozotocin diabetic peripheral nerves. Exp. Neurol. 99:201, 1988. 104. Low, P. A., Ward, K., Schmelzer, J. D., and Brimijoin, S.: Ischemic conduction failure and energy metabolism in experimental diabetic neuropathy. Am. J. Physiol. 248:E457, 1985. 105. Lundborg, G.: Structure and function of the intraneural microvessels as related to trauma, edema formation, and nerve function. J. Bone Joint Surg. Am. 57:938, 1975. 106. MacIntyre, D. E., Pearson, J. D., and Gordon, J. L.: Localisation and stimulation of prostacyclin production in vascular cells. Nature 271:549, 1978. 107. Malmberg, A. B., and Yaksh, T. L.: Spinal nitric oxide synthesis inhibition blocks NMDA-induced thermal hyperalgesia and produces antinociception in the formalin test in rats. Pain 54:291, 1993. 108. Malmberg, A. B., and Yaksh, T. L.: Cyclooxygenase inhibition and the spinal release of prostaglandin E2 and amino acids evoked by paw formalin injection: a microdialysis study in unanesthetized rats. J. Neurosci. 15:2768, 1995. 109. Martinez-Blasco, A., Bosch-Morell, F., Trenor, C., and Romero, F. J.: Experimental diabetic neuropathy: role of oxidative stress and mechanisms involved. Biofactors 8:41, 1998. 110. Mateo, M. C. M., Bustamante, J. B., and Cantalapiedra, M. A. G.: Serum zinc, copper and insulin in diabetes mellitus. Biomedicine 29:56, 1978. 111. Matkovics, B., Varga, S. I., Szabo, L., and Witas, H.: The effect of diabetes on the activities of the peroxide metabolism enzymes. Horm. Metab. Res. 14:77, 1982. 112. Maxfield, E. K., Cameron, N. E., and Cotter, M. A.: Effects of diabetes on reactivity of sciatic vasa nervorum in rats. J. Diabetes Complications 11:47, 1997. 113. McCord, J. M.: Oxygen-derived free radicals in postischemic tissue injury. N. Engl. J. Med. 312:159, 1985. 114. Mitsui, Y., Okamoto, K., Martin, D. P., et al.: The expression of proinflammatory cytokine mRNA in the sciatic-tibial nerve of ischemia-reperfusion injury. Brain Res. 844:192, 1999. 115. Mitsui, Y., Schmelzer, J. D., Zollman, P. J., et al.: Alpha-lipoic acid provides neuroprotection from ischemia-reperfusion injury of peripheral nerve. J. Neurol. Sci. 163:11, 1999. 116. Mitsui, Y., Schmelzer, J. D., Zollman, P. J., et al.: Hypothermic neuroprotection of peripheral nerve of rats from ischaemiareperfusion injury. Brain 122:161, 1999. 117. Mitsui, Y., Schmelzer, J. D., Zollman, P. J., et al.: Hypothermic neuroprotection of peripheral nerve of rats from ischemiareperfusion injury: intraischemic vs. reperfusion hypothermia. Brain Res. 827:63, 1999. 118. Moncada, S.: Eighth Gaddum Memorial Lecture, University of London Institute of Education, December 1980. Biological importance of prostacyclin. Br. J. Pharmacol. 76:3, 1982.

Oxidative Stress and Excitatory Neurotoxins in Neuropathy 119. Moncada, S., Herman, A. G., Higgs, E. A., and Vane, J. R.: Differential formation of prostacyclin (PGX or PGI2) by layers of the arterial wall: an explanation for the antithrombotic properties of vascular endothelium. Thromb. Res. 11:323, 1977. 120. Morigi, M., Angioletti, S., Imberti, B., et al.: Leukocyteendothelial interaction is augmented by high glucose concentrations and hyperglycemia in a NF-kB-dependent fashion. J. Clin. Invest. 101:1905, 1998. 121. Munoz-Fernandez, M. A., and Fresno, M.: The role of tumour necrosis factor, interleukin 6, interferon-gamma and inducible nitric oxide synthase in the development and pathology of the nervous system. Prog. Neurobiol. 56:307, 1998. 122. Murakami, K., Kondo, T., Ohtsuka, Y., et al.: Impairment of glutathione metabolism in erythrocytes from patients with diabetes mellitus. Metabolism 38:753, 1989. 123. Murase, K., Hattori, A., Kohno, M., and Hayashi, K.: Stimulation of nerve growth factor synthesis/secretion in mouse astroglial cells by coenzymes. Biochem. Mol. Biol. Int. 30:615, 1993. 124. Murwani, R., Hodgkinson, S., and Armati, P.: Tumor necrosis factor alpha and interleukin-6 mRNA expression in neonatal Lewis rat Schwann cells and a neonatal rat Schwann cell line following interferon gamma stimulation. J. Neuroimmunol. 71:65, 1996. 125. Nachemson, A. K., and Bennett, G. J.: Does pain damage spinal cord neurons? Transsynaptic degeneration in rat following a surgical incision. Neurosci. Lett. 162:78, 1993. 126. Nagamatsu, M., Nickander, K. K., Schmelzer, J. D., et al.: Lipoic acid improves nerve blood flow, reduces oxidative stress, and improves distal nerve conduction in experimental diabetic neuropathy. Diabetes Care 18:1160, 1995. 127. Nagamatsu, M., Schmelzer, J. D., Zollman, P. J., et al.: Ischemic reperfusion causes lipid peroxidation and fiber degeneration. Muscle Nerve 19:37, 1996. 128. Neely, J. R., and Feuvray, D.: Metabolic products and myocardial ischemia. Am. J. Pathol. 102:282, 1981. 129. Nickander, K. K., McPhee, B. R., Low, P. A., and Tritschler, H. J.: Alpha-lipoic acid: antioxidant potency against lipid peroxidation of neural tissue in vitro and implications for diabetic neuropathy. Free Radic. Biol. Med. 21:631, 1996. 130. Nickander, K. K., Schmelzer, J. D., Rohwer, D. A., and Low, P. A.: Effect of ␣-tocopherol deficiency on indices of oxidative stress in normal and diabetic peripheral nerve. J. Neurol. Sci. 126:6, 1994. 131. Niki, E.: Antioxidants in relation to lipid peroxidation. Chem. Phys. Lipids 44:227, 1987. 132. Nishimura, C., and Kuriyama, K.: Alteration of lipid peroxide and endogenous antioxidant contents of retina of streptozotocin-induced diabetic rats: effect of vitamin A administration. Jpn. J. Pharmacol. 37:365, 1985. 133. Nistico, G., Cirolo, M. R., Fiskin, K., et al.: NGF restores decrease in catalase activity and increases superoxide dismutase and glutathione peroxidase activity in the brain of aged rats. Free Radic. Biol. Med. 12:177, 1992. 134. Noto, R., Alicata, R., and Sfogliano, L. A.: A study of cupremia in a group of elderly diabetics. Acta Diabetol. Lat. 20:81, 1983. 135. Nukada, H.: Increased susceptibility to ischemic damage in streptozocin-diabetic nerve. Diabetes 35:1058, 1986.

531

136. Nukada, H.: Mild ischaemia causes severe pathological changes in experimental diabetic nerve. Muscle Nerve 15:1116, 1992. 137. Nukada, H., and Dyck, P. J.: Microsphere embolization of nerve capillaries and fiber degeneration. Am. J. Pathol. 115:275, 1984. 138. Nukada, H., and Dyck, P. J.: Acute ischemia causes axonal stasis, swelling, attenuation, and secondary demyelination. Ann. Neurol. 22:311, 1987. 139. Nukada, H., and McMorran, P. D.: Perivascular demyelination and intramyelinic oedema in reperfusion nerve injury. J. Anat. 185:259, 1994. 140. Nukada, H., McMorran, P. D., and Shimizu, J.: Acute inflammatory demyelination in reperfusion nerve injury. Ann. Neurol. 47:71, 2000. 141. Obrosova, I. G., Van Huysen, C., Fathallah, L., et al.: Evaluation of alpha(1)-adrenoceptor antagonist on diabetesinduced changes in peripheral nerve function, metabolism, and antioxidative defense. FASEB J. 14:1548, 2000. 142. Oksenberg, J. R.: Immunogenetics and heterogeneity in multiple sclerosis. Ann. Neurol. 40:557, 1996. 143. Olanow, C. W.: An introduction to the free radical hypothesis in Parkinson’s disease. Ann. Neurol. 32:S2, 1992. 144. Oliver, F. J., Menissier-de Murcia, J., Nacci, C., et al.: Resistance to endotoxic shock as a consequence of defective NF-kappaB activation in poly (ADP-ribose) polymerase-1 deficient mice. EMBO J. 18:4446, 1999. 145. Packer, L., Witt, E. H., and Tritschler, H. J.: alpha-Lipoic acid as a biological antioxidant. Free Radic. Biol. Med. 19:227, 1995. 146. Parry, G. J., and Brown, M. J.: Selective fiber vulnerability in acute ischemic neuropathy. Ann. Neurol. 11:147, 1982. 147. Pascoe, M. K., Low, P. A., Windebank, A. J., and Litchy, W. J.: Subacute diabetic proximal neuropathy. Mayo Clin. Proc. 72:1123, 1997. 148. Pedersen, W. A., Cashman, N. R., and Mattson, M. P.: The lipid peroxidation product 4-hydroxynonenal impairs glutamate and glucose transport and choline acetyltransferase activity in NSC-19 motor neuron cells. Exp. Neurol. 155:1, 1999. 149. Pieper, G. M., and Riaz-ul-Haq, G.: Activation of nuclear factor-kappaB in cultured endothelial cells by increased glucose concentration: prevention by calphostin C. J. Cardiovasc. Pharmacol. 30:528, 1997. 150. Pritchard, K. A. Jr., Patel, S. T., Karpen, C. W., et al.: Triglyceride-lowering effect of dietary vitamin E in streptozocin-induced diabetic rats: increased lipoprotein lipase activity in livers of diabetic rats fed high dietary vitamin E. Diabetes 35:278, 1986. 151. Pryor, W. A.: Oxy-radicals and related species: their formation, lifetimes, and reactions. Annu. Rev. Physiol. 48:657, 1986. 152. Rehncrona, S., Westerberg, E., Akesson, B., and Siesjo, B. K.: Brain cortical fatty acids and phospholipids during and following complete and severe incomplete ischemia. J. Neurochem. 38:84, 1982. 153. Romero, F. J., Monsalve, E., Hermenegildo, C., et al.: Oxygen toxicity in the nervous tissue: comparison of the antioxidant defense of rat brain and sciatic nerve. Neurochem. Res. 16:157, 1991.

532

Neurobiology of the Peripheral Nervous System

154. Romero, F. J., Segura-Aguilar, J., Monsalve, E., et al.: Antioxidant and glutathione-related enzymatic activities in rat sciatic nerve. Neurotoxicol. Teratol. 12:603, 1990. 155. Russell, J. W., Hermann, D. N., Sullivan, K. A., and Feldman, E. L.: Hyperglycemia induces programmed cell death in sensory neurons in vivo and in vitro. Endocr. Soc. Abstr. P1–475:218, 1998. 156. Russell, J. W., Sullivan, K. A., Windebank, A. J., et al.: Neurons undergo apoptosis in animal and cell culture models of diabetes. Neurobiol. Dis. 6:347, 1999. 157. Sampath, D., Jackson, G. R., Werrbach-Perez, K., and Perez-Polo, J. R.: Effects of nerve growth factor on glutathione peroxidase and catalase on PC12 cells. J. Neurochem. 62:2476, 1994. 158. Sasaki, H., Schmelzer, J. D., Zollman, P. J., and Low, P. A.: Neuropathology and blood flow of nerve, spinal roots and dorsal root ganglia in longstanding diabetic rats. Acta Neuropathol. (Berl.) 93:118, 1997. 159. Sato, Y., Hotta, N., Sakamoto, N., et al.: Lipid peroxide level in plasma of diabetic patients. Biochem. Med. 21:104, 1979. 160. Schlaepfer, W. W.: Structural alterations of peripheral nerve induced by the calcium ionophore A23187. Brain Res. 136:1, 1977. 161. Schlaepfer, W. W.: Vesicular disruption of myelin simulated by exposure of nerve to calcium ionophore. Nature 265:734, 1977. 162. Schmelzer, J. D., and Low, P. A.: The effect of hyperbaric oxygenation and hypoxia on the blood-nerve barrier. Brain Res. 473:321, 1988. 163. Schmelzer, J. D., Zochodne, D. W., and Low, P. A.: Ischemic and reperfusion injury of rat peripheral nerve. Proc. Natl. Acad. Sci. U. S. A. 86:1639, 1989. 164. Schmidt, A. M., Hori, O., Chen, J. X., et al.: Advanced glycation endproducts interacting with their endothelial receptor induce expression of vascular cell adhesion molecule-1 (VCAM-1) in cultured human endothelial cells and in mice: a potential mechanism for the accelerated vasculopathy of diabetes. J. Clin. Invest. 96:1395, 1995. 165. Shupeck, M., Ward, K. K., Schmelzer, J. D., and Low, P. A.: Comparison of nerve regeneration in vascularized and conventional grafts: nerve electrophysiology, norepinephrine, prostacyclin, malondialdehyde, and the blood-nerve barrier. Brain Res. 493:225, 1989. 166. Sladky, J. T., Greenberg, J. H., and Brown, M. J.: Regional perfusion in normal and ischemic rat sciatic nerves. Ann. Neurol. 17:191, 1985. 167. Slee, E. A., Harte, M. T., Kluck, R. M., et al.: Ordering the cytochrome c-initiated caspase cascade: hierarchical activation of caspases-2, -3, -6, -7, -8, and -10 in a caspase-9-dependent manner. J. Cell Biol. 144:281, 1999. 168. Smith, G. D., Wiseman, J., Harrison, S. M., et al.: Pre treatment with MK-801, a non-competitive NMDA antagonist, prevents development of mechanical hyperalgesia in a rat model of chronic neuropathy, but not in a model of chronic inflammation. Neurosci. Lett. 165:79, 1994. 169. Smith, K. J., Hall, S. M., and Schauf, C. L.: Vesicular demyelination induced by raised intracellular calcium. J. Neurol. Sci. 71:19, 1985.

170. Sodhi, N., Camilleri, M., Camoriano, J. K., et al.: Autonomic function and motility in intestinal pseudoobstruction caused by paraneoplastic syndrome. Dig. Dis. Sci. 34:1937, 1989. 171. Som, S., Basu, S., Mukherjee, D., et al.: Ascorbic acid metabolism in diabetes mellitus. Metabolism 30:572, 1981. 172. Soriano, F. G., Virag, L., Jagtap, P., et al.: Diabetic endothelial dysfunction: the role of poly(ADP-ribose) polymerase activation. Nat. Med. 7:108, 2001. 173. Srinivasan, S., Stevens, M., and Wiley, J. W.: Diabetic peripheral neuropathy: evidence for apoptosis and associated mitochondrial dysfunction. Diabetes 49:1932, 2000. 174. Stevens, M. J., Lattimer, S. A., Kamijo, M., et al.: Osmoticallyinduced nerve taurine depletion and the compatible osmolyte hypothesis in experimental diabetic neuropathy in the rat. Diabetologia 36:608, 1993. 175. Szabo, C., Cuzzocrea, S., Zingarelli, B., et al.: Endothelial dysfunction in a rat model of endotoxic shock: importance of the activation of poly (ADP-ribose) synthetase by peroxynitrite. J. Clin. Invest. 100:723, 1997. 176. Szabo, C., Virag, L., Cuzzocrea, S., et al.: Protection against peroxynitrite-induced fibroblast injury and arthritis development by inhibition of poly(ADP-ribose) synthase. Proc. Natl. Acad. Sci. U. S. A. 95:3867, 1998. 177. Tal, M., and Bennett, G. J.: Dextrorphan relieves neuropathic heat-evoked hyperalgesia. Neurosci. Lett. 151:107, 1993. 178. Tamura, E., and Parry, G. J.: Severe radicular pathology in rats with longstanding diabetes. J. Neurol. Sci. 127:29, 1994. 179. Toyokuni, S., Miyake, N., Hiai, H., et al.: The monoclonal antibody specific for the 4-hydroxy-2-nonenal histidine adduct. FEBS Lett. 359:189, 1995. 180. Tuck, R. R., Schmelzer, J. D., and Low, P. A.: Endoneurial blood flow and oxygen tension in the sciatic nerves of rats with experimental diabetic neuropathy. Brain 107:935, 1984. 181. Turrens, J. F., and Boveris, A.: Generation of superoxide anion by the NADH dehydrogenase of bovine heart mitochondria. Biochem. J. 191:421, 1980. 182. Wachtler, J., Mayer, C., Rucker, F., and Grafe, P.: Glucose availability alters ischaemia-induced changes in intracellular pH and calcium of isolated rat spinal roots. Brain Res. 725:30, 1996. 183. Wagner, R., Janjigian, M., and Myers, R. R.: Antiinflammatory interleukin-10 therapy in CCI neuropathy decreases thermal hyperalgesia, macrophage recruitment, and endoneurial TNF-alpha expression. Pain 74:35, 1998. 184. Wagner, R., and Myers, R. R.: Endoneurial injection of TNF-alpha produces neuropathic pain behaviors. Neuroreport 7:2897, 1996. 185. Wagner, R., and Myers, R. R.: Schwann cells produce tumor necrosis factor alpha: expression in injured and non-injured nerves. Neuroscience 73:625, 1996. 186. Weening, R. S., Wever, R., and Roos, D.: Quantitative aspects of the production of superoxide radicals by phagocytizing human granulocytes. J. Lab. Clin. Med. 85:245, 1975. 187. Weksler, B. B., Marcus, A. J., and Jaffe, E. A.: Synthesis of prostaglandin I2 (prostacyclin) by cultured human and bovine endothelial cells. Proc. Natl. Acad. Sci. U. S. A. 74:3922, 1977. 188. Wohaieb, S. A., and Godin, D. V.: Alterations in free radical tissue-defence mechanisms in streptozotocin-induced diabetes in the rat. Diabetes 36:1014, 1987.

Oxidative Stress and Excitatory Neurotoxins in Neuropathy 189. Wohaieb, S. A., and Godin, D. V.: Alterations in tissue antioxidant systems in the spontaneously diabetic (BB Wistar) rat. Can. J. Physiol. Pharmacol. 65:2191, 1987. 190. Wolff, S. P.: Diabetes mellitus and free radicals. Br. Med. Bull. 49:642, 1993. 191. Wolff, S. P., Jiang, Z. Y., and Hunt, J. V.: Protein glycation and oxidative stress in diabetes mellitus and ageing. Free Radic. Biol. Med. 10:339, 1991. 192. Wong, P. Y., and Cheung, W. Y.: Calmodulin stimulates human platelet phospholipase A2. Biochem. Biophys. Res. Commun. 90:473, 1979. 193. Yaksh, T. L.: The spinal pharmacology of facilitation of afferent processing evoked by high-threshold afferent input of the postinjury pain state. Curr. Opin. Neurol. Neurosurg. 6:250, 1993. 194. Yaksh, T. L., Chaplan, S. R., and Malmberg, A. B.: Future directions in the pharmacological management of hyperalgesic and allodynic pain states: the NMDA receptor. NIDA Res. Monogr. 147:84, 1995. 195. Yaksh, T. L., and Malmberg, A. B.: Spinal actions of NSAIDs in blocking spinally mediated hyperalgesia: the role of cyclooxygenase products. Agents Actions Suppl. 41:89, 1993.

533

196. Yao, J. K., and Low, P. A.: Improvement of endoneurial lipid abnormalities in experimental diabetic neuropathy by oxygen modification. Brain Res. 362:362, 1986. 197. Yoshida, K., Kirokawa, J., Tagami, S., et al.: Weakened cellular scavenging activity against oxidative stress in diabetes mellitus: regulation of glutathione synthesis and efflux. Diabetologia 38:201, 1995. 198. Ziboh, V. A., Maruta, H., Lord, J., et al.: Increased biosynthesis of thromboxane A2 by diabetic platelets. Eur. J. Clin. Invest. 9:223, 1979. 199. Zingarelli, B., Salzman, A. L., and Szabo, C.: Genetic disruption of poly (ADP-ribose) synthetase inhibits the expression of P-selectin and intercellular adhesion molecule-1 in myocardial ischemia/reperfusion injury. Circ. Res. 83:85, 1998. 200. Zochodne, D. W., Verge, V. M., Cheng, C., et al.: Nitric oxide synthase activity and expression in experimental diabetic neuropathy. J. Neuropathol. Exp. Neurol. 59:798, 2000. 201. Zollman, P. J., Awad, O., Schmelzer, J. D., and Low, P. A.: Effect of ischemia and reperfusion in vivo on energy metabolism of rat sciatic-tibial and caudal nerves. Exp. Neurol. 114:315, 1991.

24 Transgenic Models of Nerve Degeneration RUDOLF MARTINI

Transgenic Models Related to Myelin Sheath Organization and Maintenance Mice Homozygously Deficient for P0 Mice Heterozygously Deficient for P0 Mice Overexpressing P0 Mice Deficient in PMP22 Mice and Rats Overexpressing PMP22 Mice Deficient in the Tunnel Protein Cx32 Mice Deficient in Periaxin and Dystroglycan

Mice Deficient in ␤1 and ␤4 Integrins Mice Deficient in MAG Transgenic Models Primarily Related to the Axon Mice Deficient in the Complex Gangliosides GD1a and GT1b Mice Deficient in the Cell Adhesion Molecule L1 Mice Deficient in NF Genes Transgenic Models Related to the Organization of the Node of Ranvier

As a result of its relatively simple organization, the peripheral nerve is an excellent microcosmos to study axon-glia interactions, myelination, and disease mechanisms in myelin disorders. The principle neural partners, the axons and the Schwann cells, interact in a specific way and mutually influence each other. For instance, axonal properties trigger Schwann cell survival, differentiation, and myelination, whereas Schwann cell characteristics modulate axonal features such as axonal diameter, axonal transport, and survival. One instrumental approach to dissect these mechanisms is to establish appropriate co-culture techniques. Among many topics,17 such techniques were helpful in the resolution of Schwann cell behavior during myelination,18 in the investigation of myelin-axon communication,134 and in the investigation of the role of extracellular matrix components,39,40 cell adhesion molecules,45,46,88,89,136 and neurotrophic factors.23 In the intact organism, the influence of axons on Schwann cell phenotype was monitored by surgical cross-anastomosis of different nerve types. Larger caliber axons from sciatic nerve were allowed to grow into bundles of Remak fibers of the sympathetic system in which the sciatic axons induced myelination, reflecting the dependency of Schwann cell fate on axons.4,5,131 Thus it was

Mice Deficient in Caspr Mice Deficient in Contactin and CD9 Mice Lacking Galactocerebroside and Sulfatide Transgenic Models with Peripheral Neuropathy Not Directly Related to Neural Genes Mice Deficient in Porphobilinogen Deaminase

concluded that it is predominantly the axon that determines whether a Schwann cell forms myelin. The reciprocal influence—the effect of Schwann cells on axonal properties—has also been investigated by transplantation studies. Such experiments revealed that mutant Schwann cells can modify axonal caliber, phosphorylation of neurofilaments, and slow axonal transport.3,29,102,103 Another and more modern method to understand axon–Schwann cell interaction and myelination was to create mutants deficient in or overexpressing distinct myelin components. By such approaches, the function of several molecular players involved in myelin formation could be identified in the intact organism.75,79,97,109,132,143 In addition, the effect of impaired myelination on axon structure could be analyzed.76 Another lesson was that, in some myelin-related mutants, phenotypic alterations were unexpectedly mild. This suggests that nontarget molecules might have compensated for the roles of the inactivated genes.20–22,77,85,86 A very challenging perspective was generated when some mutants could be viewed as models for distinct genetically mediated myelin disorders. This made possible the study of pathogenesis not only in the context of axon–Schwann cell interaction, but also in relation to 535

536

Neurobiology of the Peripheral Nervous System

implications for the entire organism. For instance, the aforementioned consequences of demyelination on axonal fate and target organs, the effect of nerve length on severity of nerve pathology, and the initially unexpected diseasemodifying impact of the immune system are very important issues that have to be considered in the context of understanding pathogenesis and disease treatment.75,143 To understand the cross-talk between axons and Schwann cells, it was also of interest to generate mutants related to axonal genes. Therefore, a mutant deficient in complex gangliosides, the putative receptor molecules for myelin-associated glycoprotein (MAG), was created, resulting in a pathologic phenotype similar to that seen in MAG-deficient mice.120 Moreover, mice deficient in distinct neurofilament genes displayed not only the expected axon-related abnormalities, but also secondary defects related to myelination.61 A recent track in peripheral nervous system (PNS) research is focusing on the molecular organization of the node of Ranvier and its neighboring domains, the paranode and the juxtaparanode.54,92,95a,112 Mice deficient in two closely associated axonal components of the paranode show a very similar phenotype with disrupted paranodal organization.12,16 Interestingly, a mutant deficient in nearly ubiquitous myelin lipids developed almost identical paranodal abnormalities.32,60 This chapter gives an introduction into the phenotype of various engineered mutants related to the PNS, with the aim to provide an understanding of axon-glia interactions and the consequences when these recognition processes are disturbed in the intact organism. Furthermore, it is demonstrated that some of the mutants are instrumental in understanding pathogenesis of inherited peripheral neuropathies, which is an important prerequisite for developing novel treatment strategies for the still untreatable disorders.

TRANSGENIC MODELS RELATED TO MYELIN SHEATH ORGANIZATION AND MAINTENANCE Mice Homozygously Deficient for P0 P0 (also designated myelin protein zero) is a 30-kDa adhesion molecule belonging to the immunoglobulin (Ig) superfamily. Based on in vitro studies, it has been speculated to mediate compaction of myelin as a result of its homophilic adhesion properties (reviewed by Martini and Schachner79), leading to the close apposition of the extracellular aspects of the spiraling Schwann cell membrane forming the intraperiod line. This model received strong support from the determination of the three-dimensional structure of the extracellular domain of P0 by x-ray crystallography.118 The intracellular domain of P0 contains predominantly basic residues, which have been suggested to

interact with negatively charged phospholipids of the adjacent cytoplasmic parts of the Schwann cell membrane, leading to the formation of the major dense line.30,63,68 Interestingly, mutations in the P0 gene affecting the intracellular domain of the protein lead to reduced extracellular adhesion.135,139 P0: The Major Mediator of Myelin Compaction in Peripheral Nerves Mice homozygously deficient in P0 have been generated by targeted disruption of the gene in embryonic stem cells. This was the first engineered mouse mutant deficient in an adhesion molecule. Corroborating previous in vitro experiments, myelin compaction was substantially affected in the absence of P0 (Fig. 24–1A)51 (see Martini and Schachner 79 for review). In addition, the abnormal myelin sheaths were not stable but rather prone to degenerate, as reflected by a steady increase of demyelinated axons in mutants older than 4 weeks.79 Although myelin compaction was impaired, major dense lines were still preserved in approximately 60% of the axon–Schwann cell units of 4-week-old mutants (see Fig. 24–1A). This was unexpected, because formation of the major dense line has been speculated to be mediated by the intracellular part of the molecule.30 Based on the fact that, in the central nervous system (CNS), myelin basic protein (MBP) is a mediator of major dense line formation,99 it was proposed that this protein acts similarly in peripheral nerves of P0-deficient mice. Indeed, immunoelectron microscopy revealed expression of MBP at those sites of abnormal P0-deficient myelin that contained major dense lines.79 The hypothesis was then tested by crossbreeding P0-deficient mice with a spontaneous mutant deficient in MBP, the shiverer mouse. These MBP-deficient mice show almost normal peripheral nerves (i.e., in the presence of P0), but severely impaired CNS myelin with missing major dense lines.99 The double mutants deficient in both P0 and MBP showed a complete absence of major dense lines in peripheral nerves (Fig. 24–1B)77 thus proving that P0 and MBP can fulfill interchangeable roles in the PNS. Based on the fact that MBP can partially compensate for the loss of P0, the hypothesis of whether other adhesion molecules upregulated in peripheral nerves of P0-deficient mice51,138 had functional significance was tested. In analogy to the experiment with MBP-deficient mutants, P0deficient mice were crossbred with other mutants deficient in molecules that are upregulated in the P0 single mutant. Double-mutant mice deficient in P0 and either MAG, neural cell adhesion molecule (NCAM), or peripheral myelin protein 22 (kDa) (PMP22), revealed only subtle changes in myelin phenotype when compared with P0 single mutants.21,22 Thus P0 is the pivotal mediator of myelin compaction in the PNS, sharing some functional properties for intracellular compaction with MBP.

Transgenic Models of Nerve Degeneration

537

B

A

FIGURE 24–1 Myelin sheath in femoral nerve of a P0⫺⫺ mouse (A) and of a P0⫺⫺/MBP⫺⫺ (shiverer) double mutant (B). Note that, in the absence of P0 alone, most Schwann cells form an abnormal myelin sheath comprising both noncompacted and partially compacted myelin. Myelin sheaths of the double mutants are completely uncompacted. Bars: 1 ␮m.

P0-Deficient Mice: A Genetic Model for Myelin-Related Axonopathy Electrophysiologic investigations in P0-deficient mice revealed features indicative not only of abnormal myelination, but also of axonal impairment, because the amplitudes of compound muscle action potentials (CMAPs) were strongly reduced.80,145 Morphometric analysis of peripheral nerves of P0 mutants revealed a significant reduction of axonal diameters in proximal regions of femoral and in facial nerves. In addition, an increase in neurofilament (NF) density was detected.48 Even more striking was the loss of approximately 75% of the distal portions of myelin-competent axons and features indicative of Wallerian degeneration in the toes of the myelin mutants.48 Immunolabeling of footpads with antibodies to cytokeratin 20 revealed a 75% loss of Merkel cells, suggesting that survival of these cells is dependent on the presence or maintenance of their innervating myelinated axons.48 A similar loss of axons was found in the plantar nerve of P0-deficient mice.106 Strikingly, the most prominent rate of axon loss was found within the first 3 months (Fig. 24–2). To investigate whether myelin-related axonal loss at least partially shares the mechanisms underlying Wallerian degeneration, P0-deficient mice were crossbred with the spontaneous mutant C57BL / Wlds, which typically shows protection from Wallerian degeneration as a result of fusion of the genes for ubiquitination factor E4B and nicotinamide mononucleotide adenylyltransferase.71 The double mutants showed a significantly delayed myelin-related axonal loss,106 and retrograde

labeling of plantar nerves revealed that especially motor axons were preserved by the Wlds mutation (see Fig. 24–2). Most interestingly, the surviving axons appeared functionally active because both the amplitude of CMAP and the muscle strength were significantly increased.106 Thus myelin-related axonal loss is a process similar to Wallerian degeneration and can be delayed by introducing the Wlds gene.

Mice Heterozygously Deficient for P0 In contrast to mice homozygously deficient for P0, the reduction of P0 dosage by 50% in heterozygous mutants does not strongly impair myelination. Rather, myelin can form almost normally for approximately 4 months (Fig. 24–3A and B). Then a progressive demyelinating neuropathy is detectable in motor but not sensory nerves.74,80,105,122 Typical pathologic features are demyelinated axons, unusually thin myelin reflecting incomplete remyelination, and supernumerary Schwann cells reminiscent of onion bulbs in human neuropathies (Fig. 24–3C). In line with the demyelinating phenotype was the prolonged F-wave latency of the CMAP of these mice.80 Partial Mediation of Demyelination by Immune Cells A striking and initially unexpected feature was the occurrence of CD8-positive T lymphocytes and macrophages in the endoneurium of these mutants.116,122 The number of these immune cells increased with time and progressing demyelinating neuropathy, suggesting that they might be

538

Neurobiology of the Peripheral Nervous System

A

PO++/WT PO--/WT

1800 ***

1600

***

*

*

Number of axons

1400 1200 1000 800 600 400 200 0 1 month 1.5 months 3 months 12 months n= 3

4

2

3

3

4

Number of labelled motoneurons

B

2

2

PO++/Wlds PO++/WT PO--/WT PO--/Wlds-hom/het

250 200

* *

150 100 50 0 n=

3

2

4

4

FIGURE 24–2 Progressive axonal loss in plantar nerves of P0⫺⫺ mice (A) and delay of plantar motor axon degeneration resulting from cross-breeding of P0⫺⫺ mice with Wlds mice (B). A, Note substantial axonal loss in plantar nerve, particularly at younger ages. Axon numbers have been determined on ultrathin sections using electron microscopy. B, Number of plantar motor axons is reduced in 3-month-old P0⫺⫺ mice. At this age, motor axon loss is not vivsible in P0⫺⫺/ Wlds mutants. Values have been obtained by counting spinal motor neurons after retrograde labeling of plantar nerves using fluorogold.

functionally related to the disorder. To investigate the functional roles of these cells, the myelin mutants were crossbred with mice deficient in mature T and B lymphocytes (i.e., recombinant activating gene 1 [RAG-1]–deficient mice).84 Histologically, the double mutants showed a less severe myelin degeneration in the absence of lymphocytes (Fig. 24–4). This improvement in myelin maintenance was reflected by an amelioration of nerve conduction properties116 and was experimentally reversible, because P0⫹⫺/RAG-1⫺⫺ mutants showed an aggravated phenotype when reconstituted with bone marrow from wild-type mice (Fig. 24–4),81 proving that T lymphocytes are involved in the primarily genetically mediated demyelination.

The role of macrophages in demyelination was also investigated in P0 mutants. Electron microscopy and immunoelectron microscopy revealed that, in the P0 mutants, some macrophages had entered the endoneurial tubes, contacting either still morphologically intact myelin or demyelinated axons (Fig. 24–5A and B). Because this apposition of macrophages with endoneurial tubes was highly suggestive for an involvement in degeneration and resembled macrophage-myelin interaction in GuillainBarré syndrome,9,57,67 myelin mutants were crossbred with spontaneous mutants deficient in macrophage colonystimulating factor (M-CSF), hence displaying impaired macrophage activation. In the P0-deficient mutants also deficient in M-CSF, the numbers of macrophages were not elevated in the demyelinating nerves. The demyelinating phenotype was much less severe than in the P0 single mutants, proving that macrophages are functionally involved in genetic demyelination (Fig. 24–5C and D).19 These examples clearly demonstrate that inherited demyelination can be mediated and modulated by the immune system. The mechanisms underlying this involvement of the immune system are not yet understood, but our observations may have putative impact on future treatment strategies for inherited demyelination using immunomodulators. Sensory Nerve Preservation from Demyelination but Impairment of Function on Mechanical and Thermal Stimuli Another striking feature of mice heterozygously deficient in P0 is the relative preservation of sensory nerves and dorsal roots.74,75,80,105,122 This is reflected by the absence of supernumerary Schwann cells and other features indicative of demyelination. Moreover, neither T lymphocytes nor macrophages are elevated in the sensory nerves.19,116 Unexpectedly, a behavioral test revealed that some sensory functions were slightly impaired, as reflected by raised withdrawal thresholds to mechanical and thermal stimuli, whereas behavioral signs of a painful neuropathy were not detectable.105 On electron microscopy of longitudinal sections of sensory nerves, many nodes of Ranvier were abnormally formed and displayed enlarged nodal gaps with poorly developed nodal Schwann cell microvilli.105 These alterations might be causally linked to the sensory deficits in the absence of profound myelin degeneration in the sensory nerves of the mutants. It is not yet known how reduction in gene dosage of P0 leads to structural alteration of nodes of Ranvier. In summary, mice homo- and heterozygously deficient in P0 have been instrumental in the analysis of the functional roles of distinct myelin components. These mutants also revealed that loss of an important myelin component leads to many secondary reactions by the Schwann cells, such as upregulation of nontarget myelin components with partially compensatory functions. In addition, abnormal myelin sheaths are not stable but are prone to degenerate.

Transgenic Models of Nerve Degeneration

539

B

A

C FIGURE 24–3 Myelinated axons of 10-day-old wild-type (A), 10-day-old P0⫹⫺ (B), and 8-month-old P0⫹⫺ mice (C). Note that in young P0⫹⫺ mice myelination is normal (compare A with B). In adult mice, features indicative of demyelination, such as thin myelin and supernumerary Schwann cells (arrows), are striking (C). Bars: 1 ␮m.

Moreover, the axonal partners are influenced by the loss of the myelin protein, as reflected by a reduction in diameter and eventually degeneration. A striking observation was the involvement of the immune system in the degener-

ation of unstable myelin sheaths as a secondary reaction to reduction in gene dosage. Thus the cellular, subcellular, and molecular consequences of P0 disruption are not confined to the targeted organelle (the myelin) or the

540

Neurobiology of the Peripheral Nervous System

A

B

C

D

FIGURE 24–4 Semithin sections of femoral quadriceps nerves of P0⫹⫺ mice (A), of P0⫹⫺/RAG-1⫺⫺ mice (B), of P0⫹⫺/RAG-1⫺⫺ mice that received bone marrow from wild-type mice (C), and of P0⫹⫺/RAG-1⫺⫺ mice that received bone marrow from RAG-1 mice (D). Note that the presence of immune-competent wild-type bone marrow (A and C) leads to demyelination in P0⫹⫺ mice. All mice were investigated at the age of 1 year. Bars: 20 ␮m.

target cell (the Schwann cell), but involve the neuronal partners and even cells of the immune system. Most interestingly, both features are not confined to P0 mutants, because axonal injury and the involvement of immune cells have also been described in another myelin mutant, the connexin 32 (Cx32)–deficient mouse64 (see below). These common pathologic features of different myelin mutants are of particular interest when the mutants are viewed as animal models for inherited neuropathies.

Mice Overexpressing P0 Based on the fact that other myelin components, such as PMP22 and proteolipid protein (PLP), cause myelinopathies when expressed at unusually high levels,132,143 transgenic mice overexpressing P0 have been generated by introducing additional gene copies. The vector used consisted of the whole P0 gene of the mouse, including 6 kb of the promoter, all exons and introns, and the polyadenylation site.43,137,142

Transgenic Models of Nerve Degeneration

A

B

C

D FIGURE 24–5 A and B, Immunoelectron microscopic localization of macrophages in peripheral nerves of 6-month-old P0⫹⫺ mice. Note that F4/80-positive macrophages have penetrated the Schwann cell basal laminae. A, A macrophage contacts morphologically intact myelin (ruptures in myelin are the consequence of mild fixation necessary for immunoelectron microscopy). B, Myelin is already phagocytosed and the macrophage directly contacts the demyelinated axon. Arrows indicate electron-dense immunoreaction product identifying macrophages. S ⫽ Schwann cell. C and D, Ventral spinal roots of P0⫹⫺ mice (C) and of P0⫹⫺ mice additionally deficient in M-CSF showing an impaired macrophage activation (D). Note that demyelinating phenotype in the M-CSF–deficient myelin mutants is much less severe than in the P0 single mutants. Bars: 1 ␮m. (A and B); 5 ␮m (C and D).

541

542

Neurobiology of the Peripheral Nervous System

The transgenic mice that contained extra copies of P0 showed a dosis-dependent dysmyelinating neuropathy. Mice with mild overexpression showed a transient perinatal formation of too-thin myelin; mice with stronger elevation of P0 showed an arrest of myelination at the stage of the 1:1 ratio between Schwann cells and myelin-competent axons.137,142 Most interestingly, very high expression led to an unexpected impaired sorting of larger caliber axons (Fig. 24–6).137,142 To prove that the dysmyelination is the result of the gene dosage and not the result of a structural effect of the transgene P0 protein, the overexpressing mice were crossbred to P0⫺ mice. When crossbreeding of the transgenic mice and the knockouts was designed in a way that the gene dosage was brought to wild-type levels, the phenotype could be corrected.137 To get insight into the molecular mechanisms that underlie the impaired myelination when P0 is overexpressed, immunoelectron microscopy in peripheral nerves of mice highly overexpressing P0 has been performed. The most striking finding was that those fibers arrested in myelination showed a misdirected trafficking of P0 with abaxonal P0 expression. Moreover, P0 was unusually strongly expressed at the axon–Schwann cell interface and at the apposing Schwann cell membranes of the mesaxon. This led to an abnormally tight membrane apposition reminiscent of that seen in compacted myelin where P0 is the mediator of compaction. Ectopic expression of P0 may thus

A

block spiraling of the myelinating Schwann cells as a result of its adhesive properties in membranes that should be able to be highly mobile. Another mechanism that might exist in overexpressing mice is a disturbed stoichiometry in the expression of myelin components. Indeed, D’Urso and collaborators found evidence for the possibility that P0 and PMP22 interact with each other.34 However, it is not yet known how the misbalanced relationship of the two components led to dysmyelination.

Mice Deficient in PMP22 PMP22 was the first identified culprit in inherited peripheral neuropathies. A 1.5-megabase DNA duplication of chromosome 17p11.2-p12 containing the PMP22 gene has been shown to be the cause of the most frequent form of Charcot-Marie-Tooth (CMT) disease, CMT1A.125 The reciprocal event, the heterozygous deletion of the same chromosomal region as that causing CMT1A, leads to another inherited disorder, hereditary neuropathy with liability to pressure palsies, when duplicated.125 In Chapter 66, the use of conventional and conditional transgenic mice and rats overexpressing PMP22 and heterozygous knockout mice as models for the respective disorders is thoroughly discussed. This chapter briefly deals with mice homozygously deficient in

B FIGURE 24–6 Myelination in 10-day-old transgenic mutants overexpressing P0. Note fasciculating large-caliber axons surrounded by processes of a nonmyelinating Schwann cell (left side) (A) and a myelinating Schwann cell contacting a supernumerary larger caliber axon (asterisk) at the site of the outer mesaxon (B). This situation and fasciculating larger caliber axons reflect deficient sorting of myelin-competent axons. Bars: 1 ␮m.

Transgenic Models of Nerve Degeneration

PMP22, trying to get insight into some functional roles of the molecule. PMP22 is an integral membrane glycoprotein mainly expressed in compact myelin in peripheral nerves. Initially, PMP22 was belived to be a tetraspan membrane molecule, but recent studies suggest that alternative arrangements of the intramembrane domains are conceivable.127 There are several lines of evidence that the molecule is involved in myelin formation and maintenance, but a simplistic model such as cell adhesion, as discussed for P0, is presently not available. Biochemical approaches revealed that PMP22 and P0 can be copurified from peripheral nerve myelin, suggesting that they form complexes in the myelin membrane.34 In vitro experiments suggest a function in Schwann cell growth and differentiation,146 but mice lacking PMP22 or overexpressing the gene show a normal number of Schwann cells perinatally, and altered (increased) numbers of Schwann cells are detected only at later developmental stages.108 In the absence of PMP22, myelination is severely affected. A significant delay in myelination of most profiles, as well as some fibers with hypermyelinated myelin sheaths, are the striking abnormalities visible in peripheral nerves of 4-day-old knockout mice.2 At approximately 3 weeks of age, myelination proceeded in an unusual way in that myelin tomacula formed around almost every myelincompetent axon, as revealed by single fiber preparations. Typically, most myelin tomacula were found at paranodal aspects. In 10-week-old mice, the majority of the tomacula had disappeared in favor of features indicative of demyelination, such as demyelinated axons and supernumerary Schwann cell profiles reminiscent of onion bulbs. Thus the PMP22-deficient myelin sheaths appear to be unstable and prone to degenerate.2 Another striking feature was the involvement of axonal properties in the absence of PMP22. Mice homozygously deficient for PMP22 showed a significant reduction of axonal diameters. Moreover, axonal loss in both lumbar ventral roots and quadriceps nerves was found. In the lumbar roots, approximately 25% of the axons degenerated, whereas in more distal regions of the femoral nerve, axonal loss was even more pronounced, with 40% of axons degenerating.107 Thus, similarly to P0-deficient mice, PMP22 knockout mice show a distally pronounced axonal loss in peripheral nerves. Taken together, these observations suggest multiple functions of PMP22. The molecule might be involved in the turning of Schwann cell loops, the determination of myelin thickness, and in the maintenance of axon–Schwann cell integrity. How these functions are achieved and how other myelin components are involved, such as complex formation with P0, is not yet known.

Mice and Rats Overexpressing PMP22 Transgenic mice and rats have been generated to mimic the PMP22 duplication that is the culprit of the majority of

543

CMT1A cases.58,59,72,93,117 As in transgenic mice overexpressing P0, there was a clear positive correlation between the number of the extra gene copies and the severity of the neuropathy. A more detailed description is given in Chapter 66.

Mice Deficient in the Tunnel Protein Cx32 The discovery Cx32 as a component of peripheral nerve myelin was initially unexpected and originated from the linkage of the X-chromosomal dominant form of Charcot-MarieTooth disorder (CTMX) to mutations in Cx32.11 Detailed immuncytologic investigations113 and the demonstration of functional connexin channels in teased nerve fibers8 established the molecule as a true component of noncompacted membranes of the Schwann cell–related myelin sheath. Cx32 is thought to form so-called reflexive junctions connecting cytoplasmic domains (paranodal loops, SchmidtLanterman incisures) of the myelin spiral, thus forming a radial rapid pathway for ions and small molecules from the periaxonal collar to the Schwann cell body and vice versa.8,110 Several mutations causing CMTX have been investigated in vitro with respect to their influence on channel expression, channel properties, and trafficking of the mutant and wildtype proteins in the cell.1,100 This chapter focuses on the pathologic changes seen in null mutants with the aim to try to understand the functional role of the molecule in myelination and myelin maintenance. In addition, evidence is provided that, as is the case with reduced P0 expression, Cx32 deficiency leads to an activation of immune cells that fosters demyelination and axonopathic changes. Similar to mice heterozygously deficient in P0 or homozygously deficient in MAG (see above and subsequent paragraphs), Cx32-knockout mice show initially normal myelin formation followed by demyelination from postnatal month 4 onward.5a The features indicative of demyelination, such as thinly remyelinated axons and supernumerary Schwann cells, are very similar to those in P0⫹⫺ mice. Another similarity to these mutants is that predominantly motor nerves are affected by the mutations. However, a few features are unique to Cx32 deficiency. A characteristic hallmark is abnormally enlarged periaxonal collars5a (Fig. 24–7A). Based on the fact that periaxonal Schwann cell cytoplasm is the most distant cytoplasmic domain from the Schwann cell body, it is plausible to assume that this aspect suffers most from an interruption of the rapid pathway resulting from Cx32 deficiency. In this contex, it is interesting that absence of Cx32 strongly accelerates demyelination in P0⫹⫺ mice, which could reflect the dependency of myelin in P0⫹⫺ mice on radial pathways.86 Another specific feature of Cx32-deficient mice is axonal damage and resprouting. How this damage is mediated is not known, but it is worthwhile to mention in this context that neurons from Cx32-deficient mice are particularly vulnerable when injured.87

544

Neurobiology of the Peripheral Nervous System

Based on the fact that Cx32 is linked to the X chromosome, Scherer and colleagues115 investigated whether Cx32 is subject to X-inactivation resulting in demyelination of a subpopulation of Schwann cells. In a first step, labeling of teased fibers from Cx32⫹⫺ female mice was perfomed with Cx32-specific antibodies. Indeed, a mosaic labeling was found in that some Schwann cells displayed Cx32 immunoreactivity at their typical compartments, while other myelinating Schwann cells remained entirely negative. In a second step, pathologic changes of teased fibers from heterozygous females were compared with fibers from wild-type mice. In these preparations Cx32⫹⫺ mice, but not their wild-type littermates, showed pathologic features. These investigations suggest that the mild demyelinating neuropathy in heterozygous females is the result of lack of Cx32 in individual Schwann cells.115 Based on the observation that T lymphocytes and macrophages foster demyelination in mice heterozygously deficient in P0, the same hypothesis was investigated in homo-or hemizygous Cx32-deficient mice. Increased numbers of T lymphocytes and macrophages could be detected in demyelinating nerves of the Cx32 mutants. In addition, macrophages were found in apposition to degenerating myelin, reminiscent of a macrophage-mediated demyelinating neuropathy.65 Crossbreeding of Cx32-deficient mice with RAG-1⫺⫺ mice led not only to reduced numbers of endoneurial macrophages, but also to a substantial mitigation of features indicative of myelin degeneration and axonopathic changes (Fig. 24–7B to E).64 However, hallmarks for Cx32 deficiency, such as enlarged periaxonal Schwann cell collars, were not reduced (see Fig. 24–7D and E).64 Thus the immune system appears to be a possibly widespread modulator of the severity of pathologic changes in myelin mutants. This observation might have an important impact on treatment strategies of the corresponding human disorders.

hypermyelinated fibers showing tomacula-like morphology at 6 weeks of age (Fig. 24–8).53 As with the unstable tomacula in PMP22-deficient mice, substantial degeneration of myelin occurred at 6 months of age, as reflected by production of supernumerary Schwann cells and thinly myelinated fibers. The demyelinating features were accompanied by impaired conduction properties of peripheral nerves. Another feature is impaired remyelination of regrowing axons following peripheral nerve injury.133 Thus the progression of the disease, with initially normal myelination followed by degenerative changes, is comparable to that in MAG⫺⫺ and Cx32⫺⫺ mice and mutants heterozygously deficient in P0. However, there is a striking difference between the aforementioned myelin mutants and the periaxin mutants. Whereas in MAG⫺⫺, Cx32⫺⫺, and heterozygous P0 mutants the sensory nerves were mostly preserved from degeneration, these nerves were highly affected in the periaxin mutants. Moreover, in periaxin mutants, mechanical allodynia and thermal hyperalgesia were detectable, as revealed by decreased retraction threshold on mechanical stimuli and shortened retraction latency on heat stimuli, respectively.53 This behavior was in marked contrast to that in heterozygous

Mice Deficient in Periaxin and Dystroglycan Periaxin is a Schwann cell membrane–related and cytoskeleton-associated protein. Two proteins of different sizes (147 kDa [L-periaxin] and 16 kDa [S-periaxin]) are encoded by the corresponding gene, and each form contains a PDZ (postsynaptic density protein 95, Drosophila discs large tumor suppressor, zonula occludens-1) motif, suggesting an involvement in cell-cell contact and cell signaling.35,121 The name of the protein is derived from its early localization within the periaxonal Schwann cell loop in young mice.52 Studies focusing on later stages revealed that the expression of the protein shifts from the periaxonal to the abaxonal compartment of the Schwann cell, where it is predominantly expressed in its large splice form.114 To evaluate the functional roles of periaxin, the periaxin gene was inactivated by homologous recombination in embryonic stem cells. The homozygous null mutants displayed almost normal myelin formation, with a few

A FIGURE 24–7 Pathologic features in Cx32-deficient mice. A, Electron micrograph of a myelinated axon in the femoral quadriceps nerve of a 13-month-old Cx32⫺⫺ mouse. Note the thin myelin sheath, supernumerary Schwann cells (arrows), and typically enlarged periaxonal collar (arrowheads). Figure continued on opposite page

Transgenic Models of Nerve Degeneration

545

C

B

D

E

FIGURE 24–7 Continued B through E, Light (B and C) and electron (D and E) microscopy of ventral spinal roots of 13-month-old Cx32⫺⫺ mice with intact immune system (B and D) and of Cx32⫺⫺ littermates deficient in RAG-1 lacking mature lymphocytes (C and E). Note that degenerative features, such as thin or missing myelin (asterisk) and axonopathic changes (double asterisks), are strongly reduced in myelin mutants deficient in RAG-1. However, the typical hallmarks for Cx32 deficiency, such as enlarged periaxonal collars (arrowheads in D and E), are not influenced by immune deficiency. BV, blood vessel. Bars: 1 ␮m (A, D, and E); 10 ␮m (B and C).

P0-deficient mice, which showed the opposite behavior as reflected by raised withdrawal thresholds to mechanical and thermal stimuli, possibly as a result of malformation of nodes of Ranvier.105

The demyelinating phenotype of periaxin-deficient mice clearly demonstrates that the protein is a nonredundant component of the myelinating Schwann cell in stabilizing the myelinated axon–Schwann cell unit. This stabilization

546

Neurobiology of the Peripheral Nervous System

B

A

FIGURE 24–8 Electron microscopy of a sciatic nerve of a periaxin-deficient knockout mouse. A, A tomaculum-like profile is visible. B, Features of demyelination, remyelination, and possibly axonal resprouting (arrows) reflecting instability in myelin and axon maintenance in the absence of periaxin. Note abundant supernumerary Schwann cells forming onion bulbs. Bars: 2 ␮m. (Courtesy of Drs. Diane Sherman and Peter Brophy.)

is most probably mediated by the interaction of periaxin with dystroglycan-dystrophin–related protein 2 (DRP2), which in turn forms a complex with ␣␤-dystroglycan that interacts with laminin of the Schwann cell basal lamina.121 Thus periaxin and DRP2 form an essential link between the extracellular matrix of the Schwann cell and the Schwann cell itself, stabilizing the mature myelin sheath. In line with this pivotal function is the finding that mutations in the periaxin gene lead to distinct forms of inherited peripheral neuropathies in humans.13,55 In a recent study, the Schwann cell–specific deletion of dystroglycan using the Cre-loxP system led to misfolding of peripheral myelin as similarly observed in periaxin-deficient mice.104 The abnormally folded myelin was related to the strong downregulation of DRP2, an alteration that has been similarly described for the periaxin mutants. In addition, nodal sodium channels were significantly more weakly expressed than in wild-type mice, and mild ultrastructural changes in the organization of nodal microvilli were described, suggesting a role of dystroglycan in the cytoarchitecture of nodes of Ranvier.

Mice Deficient in ␤1 and ␤4 Integrins Other molecules linking the Schwann cell surface and the Schwann cell basal lamina are integrin heterodimers such as ␣6 /␤1 and ␣6 /␤4, which bind to laminin98,111 (see also

Chapter 19). Using the Cre-loxP system, Feltri and colleagues44 inactivated ␤1 integrin exclusively in Schwann cells, which was necessary to circumvent early embryonic death noted to occur in conventional ␤1 integrin knockout mice.41,124 The Schwann cell–specific gene inactivation of ␤1 integrin led to a severe dysmyelinating phenotype with impaired axonal sorting.44 Although regulated similarly to myelin proteins,36,42 ␤4 integrin appears to be redundant during myelination. ␤4 integrin–deficient mice that die perinatally show myelin formation in peripheral nerves similar to that of wild-type mice.47 In addition, Schwann cells of dorsal root ganglion explants from these mutants develop a capacity for myelination in vitro similar to that in explants from wildtype mice.47

Mice Deficient in MAG MAG is a transmembrane glycoprotein with five Ig-like extracellular domains, a transmembrane domain, and an intracellular portion. As a result of alternative splicing, two isoforms of 67 kDa (S-MAG) and 72 kDa (L-MAG) have been found in the deglycosylated stage that also differ in their intracellular domains.66,130 The major form in the PNS is S-MAG. This glycoprotein is a typical adhesion molecule with plasma membrane binding partners.96,140 In terms of axon–Schwann cell interactions, knowledge of the

Transgenic Models of Nerve Degeneration

axonal partner(s) might be very interesting, but these have not yet been identified. However, there is accumulating evidence that sialic acid–containing glycoproteins and complex gangliosides are good candidates for axonal MAG receptors.28,140 Thus MAG can also been viewed as a member of the siglecs.62 As opposed to P0 and PMP22, MAG is a constituent not of compact, but of noncompacted myelin. As a result of its early expression during myelination and its strategic location at the axon–Schwann cell interface,78,128,129 the molecule has been thought to play pivotal roles during myelination. However, mice deficient in MAG showed, as opposed to the CNS,109 an unexpected normal myelination in the PNS (Fig. 24–9A).69,85 Interestingly, the initially normally shaped axon–Schwann cell units were not stable in the absence of MAG. At 6 to 8 months of age, features indicative of axon and myelin disruption were visible (Fig. 24–9B) and predominantly associated with motor nerves rather than with sensory nerves.20,49,141 Reminiscent of Wallerian degeneration, myelin ovoids and upregulation of tenascin C were typical alterations in nerves of aging MAG-deficient mice.20,49 Moreover, tomacula-like structures were found predominantly at paranodal aspects of motor nerve fibers.20,141 In these regions, axonal diameters appeared strongly reduced. Detailed morphometric analysis in MAG-deficient mice revealed a dysregulation of axonal calibers already present at 3 months of age. This

A

547

reduction in axonal diameter was accompanied by an increase in NF density and by a reduction of phosphorylation of NF in the mutants.141 One interpretation of these alterations was that the too-thick myelin resulting from tomacula formation was a result of axonal shrinkage and that MAG regulates axonal caliber. However, formal proof is still missing that reduced axonal diameters are directly caused by MAG deficiency rather than by constriction of the axons resulting from poorly controlled growth of MAG-deficient myelin. Indeed, paranodal tomacula-like structures and myelin folding with reduced axonal caliber have been described in other myelin mutants, including PMP22- and periaxin-deficient mice (see above). It is therefore possible that the highly organized paranodal structures, including the complicated axon–Schwann cell apposition in the form of the paranodal network,50 are particularly susceptible to molecular alterations leading to abnormal myelin folding and axonal constriction. In summary, studies in mice deficient in MAG revealed that, in motor nerves, the myelin-related axon–Schwann cell unit cannot persist in the absence of MAG. For myelin formation, however, MAG appears to be dispensable. Based on the observation that the related cell adhesion molecule NCAM is slightly upregulated periaxonally in MAG-deficient mice, it was tempting to speculate that NCAM functionally compensates for the lack of MAG during myelin formation. Therefore, mice doubly deficient

B FIGURE 24–9 A, Normal myelin sheath of a 4-week-old MAG⫺⫺ mouse demonstrating that myelin formation is possible in the absence of MAG. B, Axon–Schwann cell units are unstable in the absence of MAG. This profile reflects axon degeneration in an 8-month-old MAG⫺⫺ mouse. Bars: 1 ␮m.

548

Neurobiology of the Peripheral Nervous System

in MAG and NCAM have been generated by crossbreeding the respective single mutants. In the absence of both adhesion molecules, myelin formation was still normal.20 However, tomacula formation and degeneration of axon–Schwann cell units was protracted by several months.20 Thus NCAM is a compensatory molecule in the absence of MAG, preventing axon–Schwann cell units from degenerating for a limited time period.

TRANSGENIC MODELS PRIMARILY RELATED TO THE AXON Mice Deficient in the Complex Gangliosides GD1a and GT1b Based on the observation that MAG-transfected COS cells adhere to microwell plates to which the gangliosides GD1a and GT1b have been adsorbed, it was proposed that GD1a and GT1b are putative axonal receptors for MAG.140 The finding that the GT1b-binding tetanus toxin C decorates the axolemma was further support for this hypothesis.119 If this view is correct, one would expect that mice deficient in GD1a and GT1b would develop neuropathologic changes similar to those of MAG-deficient mice. Therefore, mice deficient in the complex ganglioside synthesizing enzyme UDP-N-acetyl-D-galactosamine:GM3 /GD3 N-acetyl-Dgalactoaminyltransferase have been generated.120 In 12- to 16-week-old GD1a- and GT1b-deficient mutants, the pathologic changes were indeed very similar to the alterations seen in MAG-deficient mutants. In the PNS, particularly striking features were axonal collapse, myelin destruction, production of supernumerary Schwann cells, and the presence of axonal sprouts. In addition, behavioral and electrophysiologic features were abnormal.24 Based on the electron micrographs presented by Sheikh et al.,120 it appears, however, that the degenerative features are of earlier onset than in the MAG-deficient mutants. This may indicate that GD1a and GT1b are additional ligands for other Schwann cell–related molecules with functions similar to those of MAG.

Mice Deficient in the Cell Adhesion Molecule L1 L1 is an adhesion molecule of the Ig superfamily. It is expressed by immature axons and Schwann cells and is downregulated to undetectable levels when myelination starts with the turning of Schwann cell loops around segregated, prospective myelinated axons.78 Although the molecule is expressed both by neurons and Schwann cells, the corresponding null mutant is regarded as axon related, because the major phenotype in peripheral nerves appears to be caused predominantly by axonal defects (see below).

L1-deficient mice suffer from many severe abnormalities of the nervous system, including enlarged ventricles, misguiding of axons of the corticospinal tract, and abnormal organization of peripheral nonmyelinated fibers where L1 is constitutively expressed in wild-type mice.26,27 In wild-type mice, the latter structures consist of thin axonal profiles, most of which are ensheathed and separated from one another by slender Schwann cell processes. In the mutants, however, nonmyelinating Schwann cells formed not only such ensheathing processes, but also additional processes not associated with axons and protruding into the endoneurial space (Fig. 24–10). Another abnormality was a strong reduction of ensheathment in some nonmyelinating axon–Schwann cell units leading to extensive fasciculation of individual axonal profiles without intervening Schwann cell processes. In addition, the nonmyelinating Schwann cells in the mutants were associated with a much lower number of axonal profiles as compared with the wild types.27 In an elegant series of experiments, Haney and colleagues56 tested the possiblity that the disturbed axon ensheathment and the reduced number of axons is the result of the lack of L1 on the Schwann cell site, the axon site, or both. For this purpose, sciatic nerve stumps from wild-type mice were sutured into transected nerves from L1-deficient mice and vice versa. The principle finding was that axonal rather than Schwann cell L1 is required for correct ensheathment and axon survival.56

Mice Deficient in NF Genes Neurofilaments are composed of three subunits (NF-L, NF-M, and NF-H), each of which is the product of a separate gene.61 The NF-H component, with its frequent lysine-serine-proline repeats, provides potential phosphorylation sites that contribute to the side arms of neurofilaments, but NF-M can be phosphorylated as well.61 Highly phosphorylated side arms of NFs are believed to be repelled from the filamentous cores by negative charges, resulting in an increase in space between the NFs that leads to an increase in axon caliber. Deletion of the NF-L gene resulted in severe axonal atrophy and impaired maturation of regenerating axons, most probably the result of the fact that, in the absence of NF-L, NF-M and NF-H cannot form filaments.61,144 Deficiency in NF-H leads to a surprisingly modest reduction in axonal diameter, particularly in the C57/B16 mouse strains, whereas in 129J strain the axon size reduction was stronger. NF-M deficiency showed a more robust and uniform decrease in axonal diameters. Surprisingly, all mutants do not show a significant defect in the organization of the nervous system in general. However, a significant delay in myelination of regenerating axons has been documented in NF-L–deficient mice. In addition, NF-L and NF-M mutants appear to show a decrease in axon number of 10% to 20% in peripheral

Transgenic Models of Nerve Degeneration

A

549

B FIGURE 24–10 Nonmyelinated fibers in saphenous nerves of wild-type (A) and L1–deficient knockout (B) mice. Note abnormal Schwann cell processes of nonmyelinating Schwann cells protruding into the endoneurial space of the mutant nerves. This feature probably reflects loss of thin-caliber axons in the mutants. In the wild-type mice, thin-caliber axons are well ensheathed by the nonmyelinating Schwann cells and abnormal Schwann cell processes are not visible. Bars: 1 ␮m.

nerves. In a study focusing on axonal changes in aging NF-L, NF-M, and NF-H knockout mutants, Elder and colleagues38 showed a progressing axonal atrophy in lumbar roots of NF-M and NF-M/H mutants. Substantial axonal loss in aged animals, however, was not seen at the level of the roots. Investigation of more distal aspects of the peripheral nerves, where axonal loss is often more pronounced, has not been carried out. This aspect is of particular interest because mutations in NF-L have been shown to cause an axonal form of CMT with substantial axonal loss in lower nerves.83

TRANSGENIC MODELS RELATED TO THE ORGANIZATION OF THE NODE OF RANVIER In recent years, substantial progress in the knowledge of the sophisticated molecular and cytologic architecture of the node of Ranvier has been achieved.6,7,92,95a,112 In the peripheral nerve, the nodal membrane is contacted by Schwann cell microvilli. The molecular characteristics of the nodal membrane are the clustered voltage-gated Na⫹ channels and the cell adhesion molecules Nr-CAM and

neurofascin 186, which form cytoskeleton-linked complexes via ankyrin G with the Na⫹ channels. In addition, tenascin C and the proteoglycan NG2 characterize the extracellular matrix milieu of the nodal gap.6,7,54,73,92 A recent study showed that inactivation of the dystroglycan gene leads to mild abnormalities in the nodal architecture (see above). Also of particular interest is the paranodal zone, which is characterized by the paranodal or lateral loops that abut the axolemma and cause the typical “scalloping” of the axon membrane (see Chapter 19). A characteristic subcellular element is the septate junction–like transverse bands that form a partially sealing barrier between the nodal gap and the periaxonal space of the juxtaparanode and the internode. Recent studies have focused on the molecular composition of these bands and their putative functional roles. Two groups have independently identified a novel axonal component of these bands initially designated contactin-associated protein (Caspr) or Paranodin.37,82,91 As a result of the homology to the septate junction–related PDZ-related protein neurexin-IV in Drosophila,90 the molecule has also been designated NCP1 (Neurexin-Caspr-Paranodin12). Caspr2 is another axolemmal component, but found at the juxtaparanodal region together with the K⫹ channels Kv1.1 and Kv1.2.94

550

Neurobiology of the Peripheral Nervous System

During development, Caspr2 and the K⫹ channels are first expressed at the paranodal site and then shift to their final destination when Caspr is upregulated.95

Mice Deficient in Caspr The functional role of the paranodal component Caspr has recently been investigated in the corresponding knockout mutant.12 In these mice, paranodal regions of peripheral nerves lack transverse bands and the scalloping of the axonal membrane is mostly absent. Occasionally, protrusion of nodal microvilli below the loops was observed.12 Another abnormality was the absence of the putative cisrelated binding partner of Caspr, contactin, in peripheral nerves, whereas contactin was abnormally distributed in the CNS. Na⫹ channels are mainly organized similar to their organization in wild-type mice, whereas K⫹ channels, which are normally localized at the juxtaparanodal regions, directly neighbored or even partially overlapped with the nodal Na⫹ channels in the mutants. This finding may argue for a role of Caspr in sorting or separating nodal/paranodal ion channels.95 The abnormal morphologic and molecular organization of the paranode of Casprdeficient mutants was paralleled by reduced conduction velocities.12 Although Caspr deficiency produced a clearly altered paranodal organization, barely any degenerative features were detectable. This may be the result of the limited age the mutants reach, possibly because of abnormal cerebellar development (M. Bhat, personal communication, 2001).

Mice Deficient in Contactin and CD9 Contactin, the putative cis-related binding partner of Caspr, has also been inactivated in mice.10 Similar to Caspr mutants, the neurologic phenotype is strong and limits lifespan to approximately 3 months. This is most probably the result of abnormal cerebellar development, as has been suggested for the Caspr mutants.10 More detailed investigations on the nodal domains in peripheral nerves revealed alterations very similar to those seen in Caspr mutants.16 Striking features were the loss of paranodal transverse bands, the abnormally increased space between paranodal loops and the axolemma, and the loss of axolemmal scalloping. An additional finding was the absence of Caspr at paranodal regions but an accumulation in neuronal somata, indicating that presence of contactin is necessary for a correct trafficking of Caspr, possibly as molecular complexes.101 The putative paranodal trans-related binding partner of contactin, neurofascin 155,126 was expressed at the paranodes, but its detectability was more variable than in the wild-type mice. A more striking feature was the abnormal expression of shaker-like K⫹ channels. Similarly to Caspr

mutants, they directly neighbored or even partially overlapped with the nodal Na⫹ channels. Again, these morphologic and molecular alterations in the mutants led to a substantially impaired nerve conduction. A similar, yet less constant, phenotype has recently been described for mice deficient in the glial paranodal component CD9.60a

Mice Lacking Galactocerebroside and Sulfatide Lack of transverse band structures and mislocalization of K⫹ channels is not confined to the Caspr and contactin mutants. Mice deficient in the UDP-galactose:ceramide galactosyltransferase (CGT), which are unable to synthesize galactocerebroside and sulfatide, show a very similar disorganization of paranodal structures.32,33 In analogy to the Caspr and contactin mutants, this was accompanied by a lack of the paranodal proteins Caspr and contactin, and neurofascin 155 was only occasionally detectable and then diffusely distributed in the region of the paranodal Schwann cell loops.33,95 The unchanged strong expression of this glial molecule in Schmidt-Lanterman incisures clearly reflected that the downregulation of neurofascin 155 in the mutants was paranode specific.95 Similarly, as seen in the contactin- and Caspr-deficient mutants, the K⫹ channels and the Caspr2 protein remained paranodal, where the molecules are normally only transiently expressed during development, rather than being translocated to the juxtaparanodal compartment as is the case in adult wild-type mice,95 possibly as a result of the lack of Caspr in paranodal domains. Initially, altered electrophysiologic parameters in the PNS, such as reduced conduction velocities and CMAP amplitudes, have been interpreted as reduced insulation properties of the compact myelin sheaths lacking galactocerebroside or sulfatide.14,15,25 However, it is plausible that these alterations are at least partially caused by the abnormal organization of the paranodes. In contrast to mice deficient in Caspr and contactin, the CGT-deficient mice showed a severe dysmyelination in the ventral white matter of the spinal cord, while the dorsal columns were preserved.25,31 The reason for this locally restricted myelin deficiency is not known but certainly reflects impaired axon-glia interaction in the CNS.97 It is possible that the myelin deficiency in the ventral columns is related to the role of galactocerebroside and sulfatide in the incorporation of PLP into lipid rafts of the oligodendrocyte membrane so that the CGT-deficient mice suffer from a missorting of PLP123 (see also Chapter 19). Recently, it has been found that inactivation of the galactosylceramide sulfotransferase gene (GST), causing deficiency of sulfatide, leads to a very similar disorganization of the nodal region, including paranodal clustering of K⫹ channels and diffuse localization of Caspr.60 Figure 24–11 summarizes the molecular changes in node-related mutants.

551

Transgenic Models of Nerve Degeneration

A

WILD TYPE MICE Perinodal astrocyte

Juxtaparanode

Microvilli

Nav1.6 ankG

CNS/Oligodendrocytes

B

NF155

PNS/Schwann cells

CGT, GST, CANTACTIN, CASPR-I-MICE Perinodal astrocyte

FIGURE 24–11 Synoptic view of the consequences for cellular and molecular organization of the nodal complex resulting from the absence of CGT, GST, contactin, and Caspr. For comparison, consequences for myelinated fibers of the CNS (left side) are also indicated, reflecting principally similar functional roles of the molecules in PNS and CNS. (From Scherer, S. S., and Arroyo, E. J.: Recent progress on the molecular organization of myelinated axons. J. Peripher. Nerv. Syst. 7:1, 2002, with permission.)

Incisure

Kv1.1 Kv1.2 Kvβ2 Caspr2

Node

Internode

Basal lamina

Internode

Microvilli

Caspr 1.1 Nav1.6 1.2 Juxtaparanode para- Node node ank G

CNS/Oligodendrocytes

TRANSGENIC MODELS WITH PERIPHERAL NEUROPATHY NOT DIRECTLY RELATED TO NEURAL GENES Mice Deficient in Porphobilinogen Deaminase The mutant deficient in the heme biosynthesis enzyme porphobilinogen deaminase has been created to obtain an animal model for acute intermittent porphyria manifesting in a neurologic syndrome that includes autonomic neuropathy, vomiting, hypertension, tachycardia, and motor weakness associated with a peripheral neuropathy.70 In this context, it was of interest that the mouse developed an axonopathy in

NF155

Basal lamina

1.1 Caspr 1.2 para- Juxtaparanode node

Incisure

Internode

PNS/Schwann cells

motor but not sensory nerves. Particularly, the fibers larger than 8 ␮m degenerated from postnatal month 6 onward, resulting in a decrease of 90% by the age of 17 months.70 The pathologic features resembled hallmarks of Wallerian degeneration, including disruption of axonal cytoskeleton, contorted myelin sheaths devoid of axons, and the formation of Schwann cell profiles reminiscent of bands of Büngner. These degenerative events were reflected by a significant decrease of sciatic nerve–related CMAP amplitude in small foot muscles, whereas features indicative of primary demyelination were mostly absent or only mildly expressed. Major electromyographic changes were a severe and chronic neurogenic pattern with decreased motor unit action potential recruitment, increased motor unit action potential, and polyphasia.70

552

Neurobiology of the Peripheral Nervous System

The mechanisms leading to axonal damage in the mutants are still unresolved. However, based on the observation that axonal loss in these mice occurs at almost normal levels of ␦-aminolevulinic acid, one can exclude this neurotoxic heme precursor as a cause for the neuropathy, although it is typically found in plasma and urine of human patients. Rather, a decrease in heme proteins such as mitochondrial cytochromes could damage the motor neurons and/or their processes, leading to neuropathy.70

ACKNOWLEDGMENTS The author’s laboratory is supported by the Deutsche Forschungsgemeinschaft (SFB 581, special research program on “Microglia”), the Gemeinnützige Hertie-Stiftung, the State of Bavaria, and local research funds of the University of Würzburg. I am grateful to Klaus V. Toyka for stimulating discussions and support, and to all the members of the laboratory and external colleagues for their valuable contributions during scientific collaboration. I am particularly grateful to Steve Scherer, Diane Sherman, and Peter Brophy for supplying their figures.

REFERENCES 1. Abrams, C. K., Oh, S., Ri, Y., and Bargiello, T. A.: Mutations in connexin 32: the molecular and biophysical bases for the X-linked form of Charcot-Marie-Tooth disease. Brain Res. Rev. 32:203, 2000. 2. Adlkofer, K., Martini, R., Aguzzi, A., et al.: Hypermyelination and demyelinating peripheral neuropathy in Pmp22deficient mice. Nat. Genet. 11:274, 1995. 3. Aguayo, A. J., Attiwell, M., Trecarten, J., et al.: Abnormal myelination in transplanted trembler mouse Schwann cells. Nature 265:73, 1977. 4. Aguayo, A. J., Charron, L., and Bray, G. M.: Potential of Schwann cells from unmyelinated nerves to produce myelin: a quantitative ultrastructural and radiographic study. J. Neurocytol. 5:565, 1976. 5. Aguayo, A. J., Epps, J., Charron, L., and Bray, G. M.: Multipotentiality of Schwann cells in cross-anastomosed and grafted myelinated and unmyelinated nerves: quantitative microscopy and radioautography. Brain Res. 104:1, 1976. 5a. Anzini, P., Neuberg, D. H. H., Schachner, M., et al.: Structural abnormalities and deficient maintenance of peripheral nerve myelin in mice lacking the gap junction protein connexin 32. J. Neurosci. 17:4545, 1997. 6. Arroyo, E. J., and Scherer, S. S.: On the molecular architecture of myelinated fibers. Histochem. Cell Biol. 113:1, 2000. 7. Arroyo, E. J., Xu, Y.-T., Zhou, L., et al.: Myelinating Schwann cells determine the internodal localization of Kv1.1, Kv1.2, Kv␤2, and Caspr. J. Neurocytol. 28:333, 1999. 8. Balice-Gordon, R. J., Bone, L. J., and Scherer, S. S.: Functional gap junctions in the Schwann cell myelin sheath. J. Cell Biol. 142:1095, 1998.

9. Ballin, R. H., and Thomas, P. K.: Electron microscope observations on demyelination and remyelination in experimental allergic neuritis. I. Demyelination. J. Neurol. Sci. 8:1, 1969. 10. Berglund, E. O., Murai, K. K., Fredette, B., et al.: Ataxia and abnormal cerebellar microorganization in mice with ablated contactin gene expression. Neuron 24:739, 1999. 11. Bergoffen, J., Scherer, S. S., Wang, S., et al.: Connexin mutations in X-linked Charcot-Marie-Tooth disease. Science 262:2039, 1993. 12. Bhat, M. A., Rios, J. C., Lu, Y., et al.: Axon-glia interactions and the domain organization of myelinated axons requires neurexin IV/Caspr/Paranodin. Neuron 30:369, 2001. 13. Boerkoel, C., Takashima, H., Stankiewicz, P., et al.: Periaxin mutations cause recessive Dejerine-Sottas neuropathy. Am. J. Hum. Genet. 68:325, 2001. 14. Bosio, A., Binczek, E., Haupt, W. F., and Stoffel, W.: Composition and biophysical properties of myelin lipid define the neurological defects in galactocerebroside- and sulfatide-deficient mice. J. Neurochem. 70:308, 1998. 15. Bosio, A., Binczek, E., and Stoffel, W.: Functional breakdown of the lipid bilayer of the myelin membrane in central and peripheral nervous system by disrupted galactocerebroside synthesis. Proc. Natl. Acad. Sci. U. S. A. 93:13280, 1996. 16. Boyle, M. E., Berglund, E. O., Murai, K. K., et al.: Contactin orchestrates assembly of the septate-like junctions at the paranode in myelinated peripheral nerve. Neuron 30:385, 2001. 17. Bunge, R. P.: Expanding roles for the Schwann cell: ensheathment, myelination, trophism and regeneration. Curr. Opin. Neurobiol. 3:805, 1993. 18. Bunge, R. P., Bunge, M. B., and Bates, M.: Movements of the Schwann cell nucleus implicate progression of the inner (axon-related) Schwann cell process during myelination. J. Cell Biol. 109:273, 1989. 19. Carenini, S., Mäurer, M., Werner, A., et al.: The role of macrophages in demyelinating peripheral nervous system of mice heterozygously deficient in P0. J. Cell Biol. 152:301, 2001. 20. Carenini, S., Montag, D., Cremer, H., et al.: Absence of myelin-associated glycoprotein (MAG) and the neural cell adhesion molecule (N-CAM) interferes with the maintenance, but not with the formation of peripheral myelin. Cell Tissue Res. 287:3, 1997. 21. Carenini, S., Montag, D., Schachner, M., and Martini, R.: Subtle roles of neural cell adhesion molecule and myelinassociated glycoprotein during Schwann cell spiralling in P0-deficient mice. Glia 27:203, 1999. 22. Carenini, S., Neuberg, D., Schachner, M., et al.: Localization and functional roles of PMP22 in peripheral nerves of P0-deficient mice. Glia 28:256, 1999. 23. Chan, J. R., Cosgaya, J. M., Wu, Y. J., and Shooter, E. M.: Neurotrophins are key mediators of the myelination program in the peripheral nervous system. Proc. Natl. Acad. Sci. U. S. A. 98:14661, 2001. 24. Chiavegatto, S., Sun, J., Nelson, R. J., and Schnaar, R. L.: A functional role for complex gangliosides: motor deficits in GM2/GD2 synthase knockout mice. Exp. Neurol. 166:227, 2000.

Transgenic Models of Nerve Degeneration 25. Coetzee, T., Fujita, N., Dupree, J., et al.: Myelination in the absence of galactocerebroside and sulfatide: normal structure with abnormal function and regional instability. Cell 86:209, 1996. 26. Cohen, N. R., Taylor, J. S., Scott, L. B., et al.: Errors in corticospinal axon guidance in mice lacking the neural cell adhesion molecule L1. Curr. Biol. 8:26, 1998. 27. Dahme, M., Bartsch, U., Martini, R., et al.: Disruption of the mouse L1 gene leads to malformations of the nervous system. Nat. Genet. 17:346, 1997. 28. De Bellard, M. E., and Filbin, M. T.: Myelin-associated glycoprotein, MAG, selectively binds several neuronal proteins. J. Neurosci. Res. 56:213, 1999. 29. de Waegh, S. M., Lee, V. M. Y., and Brady, S. T.: Local modulation of neurofilament phosphorylation, axonal caliber, and slow axonal transport by myelinating Schwann cells. Cell 68:451, 1992. 30. Ding, Y., and Brunden, K. R.: The cytoplasmic domain of myelin glycoprotein P0 interacts with negatively charged phospholipid bilayers. J. Biol. Chem. 269:10764, 1994. 31. Dupree, J. L., Coetzee, T., Suzuki, K., and Popko, B.: Myelin abnormalities in mice deficient in galactocerebroside and sulfatide. J. Neurocytol. 27:649, 1998. 32. Dupree, J. L., Girault, J.-A., and Popko, B.: Axo-glial interactions regulate the localization of axonal paranodal proteins. J. Cell Biol. 147:1145, 1999. 33. Dupree, J. L., and Popko, B.: Genetic dissection of myelin galactolipid function. J. Neurocytol. 28:271, 1999. 34. D’Urso, D., Ehrhardt, P., and Müller, H. W.: Peripheral myelin protein 22 and protein zero: a novel association in peripheral nervous system myelin. J. Neurosci. 19:3396, 1999. 35. Dytrych, L., Sherman, D. L., Gillespie, C. S., and Brophy, P. J.: Two PDZ domain proteins encoded by the murine periaxin gene are the result of alternative intron retention and are differentially targeted in Schwann cells. J. Biol. Chem. 273:5794, 1998. 36. Einheber, S., Milner, T. A., Giancotti, F., and Salzer, J. L.: Axonal regulation of Schwann cell integrin expression suggests a role for alpha6 beta4 in myelination. J. Cell Biol. 123:1223, 1993. 37. Einheber, S., Zanazzi, G., Ching, W., et al.: The axonal membrane protein Caspr, a homologue of neurexin IV, is a component of the septate-like paranodal junctions that assemble during myelination. J. Cell Biol. 139:1495, 1997. 38. Elder, G. A., Friedrich, V. L. Jr., Margita, A., and Lazzarini, L. A.: Age-related atrophy of motor axons in mice deficient in the mid-sized neurofilament subunit. J. Cell Biol. 146:181, 1999. 39. Eldridge, C. F., Bunge, M. B., and Bunge, R. P.: Differentiation of axon-related Schwann cells in vitro: II. Control of myelin formation by basal lamina. J. Neurosci. 9:625, 1989. 40. Eldridge, C. F., Bunge, M. B., Bunge, R. P., and Wood, P. M.: Differentation of axon-related Schwann cells in vitro. I. Ascorbic acid regulates basal lamina assembly and myelin formation. J. Cell Biol. 105:1023, 1987. 41. Fassler, R., and Meyer, M.: Consequences of lack of beta 1 integrin gene expression in mice. Genes Dev. 9:1896, 1995.

553

42. Feltri, L., Scherer, S. S., Nemni, R., et al.: ␤4 Integrin expression in myelinating Schwann cells is polarized, developmentally regulated and axonally dependent. Development 120:1287, 1994. 43. Feltri, M. L., D’Antonio, M., Quattrini, A., et al.: A novel P0 glycoprotein transgene activates expression of lacZ in myelinforming Schwann cells. Eur. J. Neurol. 11:1577, 1999. 44. Feltri, M. L., Graus Porta, D., Previtali, S. C., et al.: Conditional disruption of beta 1 integrin in Schwann cells impedes interactions with axons. J. Cell Biol. 156:199, 2002. 45. Fernandez-Valle, C., Fregien, N., Wood, P. M., and Bunge, M. B.: Expression of the protein zero myelin gene in axon-related Schwann cells is linked to basal lamina formation. Development 119:867, 1993. 46. Fernandez-Valle, C., Gwynn, L., Wood, P. W., et al.: Anti␤1 integrin antibody inhibits Schwann cell myelination. J. Neurobiol. 25:1207, 1994. 47. Frei, R., Dowling, J., Carenini, S., et al.: Myelin formation by Schwann cells in the absence of ␤4 integrin. Glia 27:269, 1999. 48. Frei, R., Mötzing, S., Kinkelin, I., et al.: Loss of distal axons and sensory Merkel cells and features indicative of muscle denervation in hindlimbs of P0-deficient mice. J. Neurosci. 19:6058, 1999. 49. Fruttiger, M., Montag, D., Schachner, M., and Martini, R.: Crucial role for the myelin-associated glycoprotein in the maintenance of axon-myelin integrity. Eur. J. Neurosci. 7:511, 1995. 50. Gatzinsky, K. P., Persson, G. H., and Berthold, C. H.: Removal of retrogradely transported material from rat lumbosacral alpha-motor axons by paranodal axonSchwann cell networks. Glia 20:115, 1997. 51. Giese, K. P., Martini, R., Lemke, G., et al.: Mouse P0 gene disruption leads to hypomyelination, abnormal expression of recognition molecules, and degeneration of myelin and axons. Cell 71:565, 1992. 52. Gillespie, C. S., Sherman, D. L., Blair, G. E., and Brophy, P. J.: Periaxin, a novel protein of myelinating Schwan cells with a possible role in axon ensheathment. Neuron 12:497, 1994. 53. Gillespie, C. S., Sherman, D. L., Fleetwood-Walker, S. M., et al.: Peripheral demyelination and neuropathic pain behavior in periaxin-deficient mice. Neuron 26:523, 2000. 54. Girault, J. A., and Peles, E.: Development of nodes of Ranvier. Curr. Opin. Neurobiol. 12:476, 2002. 55. Guilbot, A., Williams, A., Ravise, N., et al.: A mutation in periaxin is responsible for CMT4F, an autosomal recessive form of Charcot-Marie-Tooth disease. Hum. Mol. Genet. 10:415, 2001. 56. Haney, C. A., Sahenk, Z., Li, C., et al.: Heterophilic binding of L1 on unmyelinated sensory axons mediates Schwann cell adhesion and is required for axonal survival. J. Cell Biol. 146:1173, 1999. 57. Ho, T. W., McKhann, G. M., and Griffin, J. W.: Human autoimmune neuropathies. Annu. Rev. Neurosci. 21:187, 1998. 58. Huxley, C., Passage, E., Manson, A., et al.: Construction of a mouse model of Charcot-Marie-Tooth disease type 1A by pronuclear injection of human YAC DNA. Hum. Mol. Genet. 5:563, 1996.

554

Neurobiology of the Peripheral Nervous System

59. Huxley, C., Passage, E., Robertson, A. M., et al.: Correlation between varying levels and the degree of demyelination and reduction in nerve conduction velocity in transgenic mice. Hum. Mol. Genet. 7:449, 1998. 60. Ishibashi, T., Dupree, J. L., Ikenaka, K., et al.: A myelin galactolipid, sulfatide, is essential for maintenance of ion channels on myelinated axon but not essential for initial cluster formation. J. Neurosci. 22:6507, 2002. 60a. Ishibashi, T., Ding, L., Ikenaka, K., et al.: Tetraspanin protein CD9 is a novel paranodal component regulating paranodal junctional formation. J. Neurosci. 24:96, 2004. 61. Julien, J.-P.: Neurofilament functions in health and disease. Curr. Opin. Neurobiol. 9:554, 1999. 62. Kelm, S., Schauer, R., and Crocker, P. R.: The sialoadhesins—a family of sialic acid-dependent cellular recognition molecules within the immunoglobulin superfamily. Glycoconj. J. 13:913, 1996. 63. Kirschner, D. A., and Ganser, A. L.: Compact myelin exists in the absence of basic protein in the shiverer mutant mouse. Nature 283:207, 1908. 64. Kobsar, I., Berghoff, M., Samsam, M., et al.: Preserved myelin integrity and reduced axonopathy in connexin32 deficient mice lacking the recombination activation gene1. Brain 126:804, 2003. 65. Kobsar, I., Mäurer, M., Ott, T., and Martini, R.: Macrophage-related demyelination in peripheral nerves of mice deficient in the gap junction protein connexin 32. Neurosci. Lett. 320:17, 2002. 66. Lai, C., Brow, M. A., Nave, K. A., et al.: Two forms of 1B236/myelin-associated glycoprotein, a cell adhesion molecule for postnatal neural development, are produced by alternative splicing. Proc. Natl. Acad. Sci. U. S. A. 84:4337, 1987. 67. Lampert, P. W.: Mechanism of demyelination in experimental allergic neuritis. Lab. Invest. 20:127, 1969. 68. Lemke, G., Lamar, E., and Patterson, J.: Isolation and analysis of the gene encoding peripheral myelin protein zero. Neuron 1:73, 1988. 69. Li, C., Tropak, M. B., Gerial, R., et al.: Myelination in the absence of myelin-associated glycoprotein. Nature 369:747, 1994. 70. Lindberg, R. L. P., Martini, R., Baumgartner, M., et al.: Motor neuropathy in porphobilinogen deaminase-deficient mice imitiates the peripheral neuropathy of human acute porphyria. J. Clin Invest. 103:1127, 1999. 71. Mack, T. G. A., Reiner, M., Beurowski, B., et al.: Wallerian degeneration of injured axons and synapses is delayed by a Ube4b/Nmnat chimeric gene. Nat. Neurosci. 4:1199, 2001. 72. Magyar, J. P., Martini, R., Ruelicke, T., et al.: Impaired differentiation of Schwann cells in transgenic mice with increased PMP22 gene dosage. J. Neurosci. 16:5351, 1996. 73. Martin, S., Levine, A. K., Chen, Z. J., et al.: Deposition of the NG2 proteoglycan at nodes of Ranvier in the peripheral nervous system. J. Neurosci. 21:8119, 2001. 74. Martini, R.: Animal models for inherited peripheral neuropathies. J. Anat. 191:321, 1997. 75. Martini, R.: Animal models for inherited peripheral neuropathies: chances to find treatment strategies? J. Neurosci. Res. 61:244, 2000.

76. Martini, R.: The effect of myelinating Schwann cells on axons. Muscle Nerve 24:456, 2001. 77. Martini, R., Mohajeri, M. H., Kasper, S., et al.: Mice doubly deficient in the genes for P0 and myelin basic protein show that both proteins contribute to the formation of the major dense line in peripheral nerve myelin. J. Neurosci. 15:4488, 1995. 78. Martini, R., and Schachner, M.: Immunoelectron microscopic localization of neural cell adhesion molecules (L1, N-CAM, and MAG) and their shared carbohydrate epitope and myelin basic protein in developing sciatic nerve. J. Cell Biol. 103:2439, 1986. 79. Martini, R., and Schachner, M.: Molecular bases of myelin formation as revealed by investigations on mice deficient in glial cell surface molecules. Glia 19:298, 1997. 80. Martini, R., Zielasek, J., Toyka, K. V., et al.: Protein zero (P0)-deficient mice show myelin degeneration in peripheral nerves characteristic of inherited human neuropathies. Nat. Genet. 11:281, 1995. 81. Mäurer, M., Schmid, C. D., Bootz, F., et al.: Bone marrow transfer from wild type mice reverts the beneficial effect of genetically-mediated immune deficiency in myelin mutants. Mol. Cell. Neurosci. 17:1094, 2001. 82. Menegoz, M., Gaspar, P., Le Bert, M., et al.: Paranodin, a glycoprotein of neuronal paranodal membranes. Neuron 19:319, 1997. 83. Mersiyanova, I. V., Perepelov, A. V., Polyakov, A. V., et al.: A new variant of Charcot-Marie-Tooth disease type 2 is probably the result of a mutation in the neurofilamentlight gene. Am. J. Hum. Genet. 67:37, 2000. 84. Mombaerts, P., Iacomini, J., Johnson, R. S., et al.: RAG-1deficient mice have no mature B and T lymphocytes. Cell 68:869, 1992. 85. Montag, D., Giese, K. P., Bartsch, U., et al.: Mice deficient for the myelin-associated glycoprotein show subtle abnormalities in myelin. Neuron 13:229, 1994. 86. Neuberg, D. H.-H., Carenini, S., Schachner, M., et al.: Accelerated demyelination of peripheral nerves in mice deficient in connexin 32 and protein zero. J. Neurosci. Res. 53:542, 1998. 87. Oguro, K., Jover, T., Tanaka, H., et al.: Global ischemiainduced increases in the gap junctional proteins connexin 32 (Cx32) and Cx36 in hippocampus and enhanced vulnerability of Cx32 knock-out mice. J. Neurosci. 21:7534, 2001. 88. Owens, G. C., Boyd, C. J., Bunge, R. P., and Salzer, J. L.: Expression of recombinant myelin-associated glycoprotein in primary Schwann cells promotes the initial investment of axons by myelinating Schwann cells. J. Cell Biol. 111:1171, 1990. 89. Owens, G. C., and Bunge, R. P.: Schwann cells infected with a recombinant retrovirus expressing myelin-associated glycoprotein antisense RNA do not form myelin. Neuron 7:565, 1991. 90. Peles, E., Joho, K., Plowman, G., and Schlessinger, J.: Close similarity between Drosophila neurexin IV and mammalian Caspr protein suggests a conserved mechanism for cellular interactions. Cell 88:745, 1997. 91. Peles, E., Nativ, M., Lustig, M., et al.: Identification of a novel contactin-associated transmembrane receptor with

Transgenic Models of Nerve Degeneration

92. 93.

94.

95.

95a.

96.

97. 98.

99.

100.

101.

102.

103.

104.

105.

106.

107.

multiple domains implicated in protein-protein interactions. EMBO J. 16:978, 1997. Peles, E., and Salzer, J. L.: Molecular domains of myelinated axons. Curr. Opin. Neurobiol. 10:558, 2000. Perea, J., Robertson, A., Tolmachova, T., et al.: Induced myelination and demyelination in a conditional mouse model of Charcot-Marie-Tooth disease type 1A. Hum. Mol. Genet. 10:1007, 2001. Poliak, S., Gollan, L., Martinez, R., et al.: Caspr2, a new member of the neurexin superfamily, is localized at juxtaparanodes of myelinated axons and associates with K⫹ channels. Neuron 24:1037, 1999. Poliak, S., Gollan, L., Salomon, D., et al.: Localization of Caspr2 in myelinated nerves depends on axon-glia interactions and the generation of barriers along the axon. J. Neurosci. 21:7568, 2001. Poliak, S., and Peles, E.: The local differentiation of myelinated axons at nodes of Ranvier. Nat. Rev. Neurosci. 4:968, 2003. Poltorak, M., Sadoul, R., Keilhauer, G., et al.: Myelinassociated glycoprotein, a member of the L2/HNK-1 family of neural cell adhesion molecules, is involved in neuronoligodendrocyte and oligodendrocyte-oligodendrocyte interaction. J. Cell Biol. 105:1893, 1987. Popko, B.: Myelin galactolipids: mediators of axon-glial interactions? Glia 29:149, 2000. Previtali, S. C., Feltri, M. L., Archelos, J. J., et al.: Role of integrins in the peripheral nervous system. Prog. Neurobiol. 64:35, 2001. Privat, A., Jacque, C., Bourre, J. M., et al.: Absence of the major dense line in the myelin of the mutant mouse “shiverer.” Neurosci. Lett. 12:107, 1979. Ressot, C., and Bruzzone, R.: Connexin channels in Schwann cells and the development of the X-linked form of Charcot-Marie-Tooth disease. Brain Res. Rev. 32:192, 2000. Rios, J. C., Melendez-Vasquez, C. V., Einheber, S., et al.: Contactin-associated protein (Caspr) and contactin form a complex that is targeted to the paranodal junctions during myelination. J. Neurosci. 20:8354, 2000. Sahenk, Z., and Chen, L.: Abnormalities in the axonal cytoskeleton induced by a connexin32 mutation in nerve xenografts. J. Neurosci. Res. 51:174, 1998. Sahenk, Z., Chen, L., and Mendell, J. R.: Effects of PMP22 duplication and deletions on the axonal cytoskeleton. Ann. Neurol. 45:16, 1999. Saito, F., Moore, S. A., Barresi, R., et al.: Unique role of dystroglycan in peripheral nerve myelination, nodal structure, and sodium channel stabilization. Neuron 38:747, 2003. Samsam, M., Frei, R., Marziniak, M., et al.: Impaired sensory function in heterozygous P0 knockout mice is associated with nodal changes in sensory nerves. J. Neurosci. Res. 67:167, 2002. Samsam, M., Mi, W., Wessig, C., et al.: The wlds mutation delays robust loss of motor and sensory axons in a genetic model for myelin-related axonopathy. J. Neurosci. 23:2833, 2003. Sancho, S., Magyar, J. P., Aguzzi, A., and Suter, U.: Distal axonopathy in peripheral nerves of PMP22-mutant mice. Brain 122:1563, 1999.

555

108. Sancho, S., Young, P., and Suter, U.: Regulation of Schwann cell proliferation and apoptosis in PMP22-deficient mice and mouse models of Charcot-Marie-Tooth disease type 1A. Brain 124:2177, 2001. 109. Schachner, M., and Bartsch, U.: Multiple functions of the myelin-associated glycoprotein MAG (siglec-4a) in formation and maintenance of myelin. Glia 29:154, 2000. 110. Scherer, S. S.: Molecular genetics of demyelination: new wrinkles on an old membrane. Neuron 18:13, 1997. 111. Scherer, S. S.: Myelination: some receptors required. J. Cell Biol. 156:13, 2002. 112. Scherer, S. S., and Arroyo, E. J.: Recent progress on the molecular organization of myelinated axons. J. Peripher. Nerv. Syst. 7:1, 2002. 113. Scherer, S. S., Deschênes, S. M., Xu, Y. T., et al.: Connexin32 is a myelin-related protein in the PNS and CNS. J. Neurosci. 15:8281, 1995. 114. Scherer, S. S., Xu, Y. T., Bannerman, P. G. C., et al.: Periaxin expression in myelinating Schwann cells: modulation by axon-glial interactions and polarized localization during development. Development 121:4265, 1995. 115. Scherer, S. S., Xu, Y.-T., Nelles, E., et al.: Connexin32-null mice develop demyelinating peripheral neuropathy. Glia 24:8, 1998. 116. Schmid, C. D., Stienekemeier, M., Oehen, S., et al.: Immune deficiency in mouse models for inherited peripheral neuropathies leads to improved myelin maintenance. J. Neurosci. 20:729, 2000. 117. Sereda, M., Griffiths, I., Pühlhofer, A., et al.: A transgenic rat model of Charcot-Marie-Tooth disease. Neuron 16:1049, 1996. 118. Shapiro, L., Doyle, J. P., Hensley, P., et al.: Crystal structure of the extracellular domain from P0, the major structural protein of peripheral nerve myelin. Neuron 17:435, 1996. 119. Sheikh, K. A., Deerinck, T. J., Ellisman, M. H., and Griffin, J. W.: The distribution of ganglioside-like moieties in peripheral nerves. Brain 122:449, 1999. 120. Sheikh, K. A., Sun, J., Liu, Y., et al.: Mice lacking complex gangliosides develop Wallerian degeneration and myelination defects. Proc. Natl. Acad. Sci. U. S. A. 96:7532, 1999. 121. Sherman, D. L., Fabrizi, C., Gillespie, C. S., and Brophy, P. J.: Specific disruption of a Schwann cell dystrophin-related protein complex in a demyelinating neuropathy. Neuron 30:677, 2001. 122. Shy, M. E., Arroyo, E., Sladky, J., et al.: Heterozygous P0 knock-out mice develop a peripheral neuropathy that resembles chronic inflammatory demyelinating polyneuropathy. J. Neuropathol. Exp. Neurol. 56:811, 1997. 123. Simons, M., Kramer, E. M., Thiele, C., et al.: Assembly of myelin by association of proteolipid protein with cholesterol- and galactosylceramide-rich membrane domains. J. Cell Biol. 151:143, 2000. 124. Stephens, L. E., Sutherland, A. E., Klimanskaya, I. V., et al.: Deletion of beta 1 integrins in mice results in inner cell mass failure and peri-implantation lethality. Genes Dev. 9:1883, 1995. 125. Suter, U., and Snipes, G. J.: Biology and genetics of hereditary motor and sensory neuropathies. Annu. Rev. Neurosci. 18:45, 1995.

556

Neurobiology of the Peripheral Nervous System

126. Tait, S., Gunn-Moore, F., Collinson, J. M., et al.: An oligodendrocyte adhesion molecule at the site of assembly of the paranodal axo-glial junction. J. Cell Biol. 150:657, 2000. 127. Taylor, V., Zgraggen, C., Naef, R., and Suter, U.: Membrane topology of peripheral myelin protein 22. J. Neurosci. Res. 62:15, 2001. 128. Trapp, B. D., O’Connell, M. F., and Andrews, S. B.: Ultrastructural immunolocalization of MAG and P0 proteins in cryosections of peripheral nerve. J. Cell Biol. 135:228a, 1986. 129. Trapp, B. D., and Quarles, R. H.: Presence of the myelinassociated glycoprotein correlates with alterations in the periodicity of peripheral myelin. J. Cell Biol. 92:877, 1982. 130. Tropak, M. B., Johnson, P. W., Dunn, R. J., and Roder, J. C.: Differential splicing of MAG transcripts during CNS and PNS development. Brain Res. 464:143, 1988. 131. Weinberg, H. J., and Spencer, P. S.: Studies on the control of myelinogenesis. I. Myelination of regenerating axons after entry into a foreign unmyelinated nerve. J. Neurocytol. 4:395, 1975. 132. Werner, H., Jung, M., Klugmann, M., et al.: Mouse models of myelin disease. Brain Pathol. 8:771, 1998. 133. Williams, A. C., and Brophy, P. J.: The function of the periaxin gene during nerve repair in a model of CMT4F. J. Anat. 200:323, 2002. 134. Windebank, A. J., Wood, P., Bunge, R. P., and Dyck, P. J.: Myelination determines the caliber of dorsal root ganglions in culture. J. Neurosci. 56:1563, 1985. 135. Wong, M. H., and Filbin, M. T.: Dominant-negative effect on adhesion by myelin P0 protein truncated in its cytoplasmic domain. J. Cell Biol. 134:1531, 1996. 136. Wood, P. M., Schachner, M., and Bunge, R. P.: Inhibition of Schwann cell myelination in vitro by antibody to the L1 adhesion molecule. J. Neurosci. 111:3635, 1990.

137. Wrabetz, L., Feltri, M. L., Quattrini, A., et al.: P0 glycoprotein overexpression causes congenital hypomyelination of peripheral nerves. J. Cell Biol. 148:1021, 2000. 138. Xu, W., Menichella, D., Jiang, H., et al.: Absence of P0 leads to the dysregulation of myelin gene expression and myelin morphogenesis. J. Neurosci. Res. 60:714, 2000. 139. Xu, W., Shy, M., Kamholz, J., et al.: Mutations in the cytoplasmic domain of P0 reveal a role for PKC-mediated phosphorylation in adhesion and myelination. J. Cell Biol. 155:439, 2001. 140. Yang, L. J. S., Zeller, C. B., Shaper, N. L., et al.: Gangliosides are neural ligands for myelin-associated glycoprotein. Proc. Natl. Acad. Sci. U. S. A. 93:814, 1996. 141. Yin, X., Crawford, T. O., Griffin, J. W., et al.: Myelinassociated glycoprotein is a myelin signal that modulates the caliber of myelinated axons. J. Neurosci. 18:1953, 1998. 142. Yin, X., Kidd, G. J., Wrabetz, L., et al.: Schwann cell myelination requires timely and precise targeting of P0 protein. J. Cell Biol. 148:1009, 2000. 143. Young, P., and Suter, U.: Disease mechanisms and potential therapeutic strategies in Charcot-Marie-Tooth disease. Brain Res. Rev. 36:213, 2001. 144. Zhu, Q., Couillard-Despres, S., and Julien, J.-P.: Delayed maturation of regenerating myelinated axons in mice lacking neurofilaments. Exp. Neurol. 148:299, 1997. 145. Zielasek, J., Martini, R., and Toyka, K. V.: Functional abnormalities in P0-deficient mice resemble human hereditary neuropathies linked to P0 gene mutations. Muscle Nerve 19:946, 1996. 146. Zoidl, G., Blass-Kampmann, S., D’Urso, D., et al.: Retroviral-mediated gene transfer of the peripheral myelin protein PMP22 in Schwann cells: modulation of cell growth. EMBO J. 14:1122, 1995.

25 Introduction to Immune Reactions in the Peripheral Nervous System HANS-PETER HARTUNG, BERND C. KIESEIER, RALF GOLD, GUIDO STOLL, AND KLAUS V. TOYKA

Immunocompetent Cellular Components of the PNS Macrophages T Lymphocytes

B Lymphocytes Mast Cells Schwann Cells Tolerance and Autoimmunity

The peripheral nervous system (PNS) has traditionally been considered as “immunologically privileged.” This view has undergone revision within recent years, and nowadays differences from other organs appear to be more quantitative than qualitative. The PNS is separated from the external environment by the blood-nerve barrier (BNB), which does restrict access of immune cells and soluble mediators to a certain degree; however, this restriction is not complete, either anatomically (it is absent or relatively deficient at the roots, in the ganglia, and in the motor terminals) or functionally. The PNS, as are most organs, is subject to immune surveillance. Activated T lymphocytes can cross the BNB irrespective of their antigen specificity, patrolling the PNS scanning for non–self-antigens, and abundant antigen-presenting cells (APCs) can be detected in peripheral nerve tissue. Molecules required for initiation of a cognate or innate immune response can be induced. Hence a local immune circuitry is in place to defend the integrity of the PNS or, when regulatory mechanisms fail to interact with the extraneural immune system, to cause its destruction. Various cell types armed with immunocompetent properties and found in normal PNS play distinct roles in immune reactions in the PNS, and their interplay is discussed in this chapter (Fig. 25–1).

B-Cell Tolerance T-Cell Tolerance Breakdown of Tolerance Termination of the Immune Response

IMMUNOCOMPETENT CELLULAR COMPONENTS OF THE PNS Macrophages A considerable number of local macrophages reside within the endoneurium of peripheral nerve. First recognized by Arvidson because of their phagocytosing capacity, resident macrophages are known to comprise up to 9% of the cellular components of normal peripheral nerve.7,8,34 Resident endoneurial macrophages of normal nerve are elongated cells stretching along the longitudinal axis of peripheral nerves and show small ramifications with two or three terminal branches at their ends. They are found close to endoneurial blood vessels but also scattered throughout the endoneurium. Moreover, these cells usually lie outside the basal lamina of the blood vessels and have a dendritic appearance.54 Their perivascular distribution at the bloodnerve interface makes them uniquely suited to act as APCs in the PNS.37 Many of these resident macrophages constitutively express major histocompatibility complex (MHC) class II molecules, complement receptor type 3 (CR3), and CD4, but the level of expression of these and additional co-stimulatory molecules can be enhanced greatly in inflammatory conditions, including Guillain-Barré syndrome (GBS) and chronic inflammatory demyelinating 559

560

Neuroimmunology of the Peripheral Nervous System

Tissue resident/endoneurial macrophage -MHC I/II -B7-1/2 Perivascular macrophage -MHC I/II -B7-1/2 T

Cytokines NO proteases

Blood-nerve-barrier Schwann cell

Damaged neurons -MHC I

Antigen presentation (MHC I/II) ICAM, cytokines, NO, proteases

polyradiculoneuropathy (CIDP).52,54 Studies suggest that endoneurial macrophages might act as sensors of pathology at early disease stages much like their central nervous system counterparts, the microglial cells.63 Macrophages are likely to be involved in virtually all steps of an immune reaction within the PNS. From early immune surveillance, antigen presentation, and activation of the cellular immune cascade throughout the evolving immune response, they serve a multitude of functions both protective and injurious: antigen-specific demyelination and axonal damage, nonspecific secondary tissue destruction, removal of debris, and regeneration.34,54 It has been well established that macrophage-mediated segmental demyelination is the pathologic hallmark of autoimmune demyelinating polyneuropathies, including classic GBS and CIDP (see Chapters 27, 98, and 99).42 Histopathologically, macrophages invade the Schwann cell basal lamina, penetrate myelin lamellae, strip myelin lamellae, and phagocytose both damaged and apparently intact myelin.7 Macrophages may be derived from circulating monocytes that invade or reside in the PNS. Hematogenous macrophages find their way into the peripheral nerve en passant with T cells following a concerted action of adhesion molecules,4 matrix metalloproteinases,55 and chemotactic signals.56 The immunologic processes leading to macrophage-mediated segmental demyelination have only partially been explored. Macrophages, in contrast to T cells, do not act in an antigen-specific manner and need to be targeted by additional mechanisms. There is good evidence that antibodies, through binding to their Fc receptors, may

FIGURE 25–1 The local immune circuitry of the peripheral nervous system (PNS). The blood-nerve barrier restricts access of immune cells and soluble mediators to a certain degree; however, activated T cells (T) can cross this barrier and scan the nervous system for non–self-antigens. Perivascular as well as tissue-resident macrophages, the latter abundantly found within the endoneurium, can act as antigen-presenting cells (APC), expressing MHC class I and II molecules and co-stimulatory molecules, such as B7-1 or B7-2. Schwann cells and damaged neurons are able to present antigens as well. Inflammatory mediators (e.g., cytokines, proteases, nitric oxide [NO]), can be released by macrophages and Schwann cells and perpetuate local inflammation within the PNS. See Color Plate

direct macrophages toward their myelin or axonal targets in autoimmune neuropathies.54 Activation of Fc receptors on macrophages results in the release of toxic mediators (oxygen radicals, nitric oxide [NO] metabolites, and proteases) that can damage the myelin sheath. Activated macrophages also strip myelin lamellae and phagocytose off both damaged and apparently intact myelin. Once within the nerve, macrophages promote inflammation by releasing pro-inflammatory cytokines including interleukin (IL)-1 and IL-6, as well as tumor necrosis factor-␣ (TNF-␣).61,75,108 They thus not only act as chief effector cells in demyelination and tissue destruction but are also intimately involved in the control of the pathogenetic process. Conversely, macrophages also contribute to the termination of the immuno-inflammatory attack through the induction of T-cell apoptosis28 and the release of antiinflammatory cytokines, including transforming growth factor (TGF)-␤1 and IL-10.53 It can be speculated that macrophages support the normal intact microenvironment of the PNS, but when disturbed by autoimmunity, they escape from normal control and develop their destructive potential as a result of the misdirected specific immune response following, for example, a preceding infection as in GBS.

T Lymphocytes T cells recognize via the T-cell receptor short linear peptide fragments in the context of MHC molecules on the cell surface. Each T cell carries its own unique receptor. Part of the

561

Introduction to Immune Reactions in the Peripheral Nervous System

T-cell receptor recognizes the foreign peptide, and part of it recognizes the self MHC molecule.44 Only a minority of T lymphocytes are able to perform this task, and a rigorous process of selection is necessary to determine those T cells that form the pool of lymphocytes to be exported from the thymus during T-cell development. This process is termed thymic education and involves both positive and negative selection.76,80,85,109 T-cell receptors are produced as transmembrane molecules; they are composed as heterodimers and consist of an ␣ and a ␤ or a ␥ and a ␦ chain, with each chain containing a variable and a constant domain. Eighty-five percent to 90% of the T cells carry the ␣/␤ T-cell receptor and 5% to 15% have ␥/␦ receptors. Whereas ␣/␤ T-cell receptors recognize a complex formed by a peptide seated within the groove of an MHC molecule, most ␥/␦ T cells do not recognize antigen in the form of peptide-MHC complexes. ␥/␦ T cells may be A

α– β–

triggered by infectious agents.68 T-cell responses can also be mounted against lipids and glycolipids, which are presented by CD1 molecules (CD1a, CD1b, CD1c, and CD1d). The CD1 proteins are distantly related to MHC class I and class II molecules and represent a small to moderate-sized family of ␣2-microglobulin–associated transmembrane proteins.13,24 Expression of CD1a and CD1b has been detected in endoneurial macrophages in inflammatory neuropathies (Fig. 25–2).51,95 T lymphocytes can be divided in two major subclasses: CD4⫹ and CD8⫹ T cells. CD4⫹ T cells usually act as helper T (Th) cells and recognize preferentially extracellular peptide antigens presented by MHC class II molecules, whereas CD8⫹ T cells are usually cytotoxic and recognize cytosolic antigens presented by MHC class I molecules. The latter are expressed on all nucleated cells. Thus any infected cell can signal to CD8⫹ T cells to get destroyed,

chain Variable region

Disulfide bonds Cell membrane

Bone marrow

Prothymocyte

Constant region

α0β0 γ 0δ0 r γ/δ r D-Jβ γ –δ–

CD3+ CD4+ CD8+

α+β+

CD3+ CD4– CD8–

γ +δ+

CD3+ CD4– CD8–

γ +δ+

Thymus

CD3– CD4– CD8–

CD3+ CD4– CD8+

α+β+

CD3+ CD4+ CD8–

α+β+

Periphery

Selection

B FIGURE 25–2 Structure of the T-cell receptor and its distribution during T-cell development. A, The T-cell receptor, expressed on the cellular surface only, is a heterodimer composed of two transmembrane glycoprotein chains, the ␣ and the ␤ chain. Both consist of a variable and a constant domain, and carry carbohydrate side chains. An alternative type of T-cell receptor is made up of different polypeptides designated ␥ and ␦. B, During T-cell development various types of T cells are selected. See Color Plate

562

Neuroimmunology of the Peripheral Nervous System

Immature effector T cell – CD4+

removing sites of pathogen replication. MHC class II molecules are expressed on the surface of professional APCs only. Subtypes of both CD4⫹ and CD8⫹ T cells engage in immunoregulatory actions and are instrumental in checking T-cell activities and in maintaining peripheral tolerance. The CD4⫹CD25⫹ regulatory T cell (Treg) population appears particularly responsive to self-antigens and functionally prevents under normal circumstances emergence or expansion of autoreactive T-cell clones.49,81,99 Once a naive CD4⫹ T lymphocyte encounters a specific antigen on the surface of a professional APC in the context of co-stimulatory molecules, such as CD80 or CD86, it gets activated, proliferates, and differentiates into an appropriately armed effector Th lymphocyte.1 A new serial encounter model of antigen recognition invokes a dynamic physical contact with T cells literally crawling over and scanning the surface of multiple APCs.45,104 Three types of such effector Th cells can be distinguished. Th1 cells manufacture effector molecules that activate macrophages, and Th2 cells generate B-cell activating effector molecules. The third group, so-called Th0 cells, from which both these functional classes of Th cells derive, also secrete molecules of both Th1 and Th2 and may therefore have a distinct effector function (Fig. 25–3).

Mature effector T cell – CD4+

+ IFN-γ IL-12 _ IL-4 IL-10

In normal nerves, T lymphocytes can only rarely be identified, although they normally traffic into and through the endoneurium of the PNS. Local activation of T cells requires antigen presentation in the context of MHC class II antigens, which can be found on a restricted number of cells in the peripheral nerves. In immune-mediated disorders the number of T cells within the PNS increases dramatically by invasion and clonal expansion in situ, underlining the important role of this cell type in the local immune response. A central role of T cells in disease pathogenesis was established by demonstrating that transfer of autoreactive neuritogenic T-cell lines into healthy recipient animals can induce the clinical, electrophysiologic, and morphologic features of classic experimental autoimmune neuritis (EAN) (see Chapter 27), a model of the human GBS.42 The pivotal role of T cells is also underscored by the preventive and suppressive effects of manipulations that eliminate or silence T lymphocytes in EAN (see Chapter 27). T-cell–derived immune responses are initiated and activated in the systemic immune compartment by the trimolecular interaction between APCs, MHC molecules, and a specific T-cell receptor, and fine tuning is done by the simultaneous perception of co-stimulatory signals,

TH0

IL-4 + IL-10 IFN-γ _ IL-12

TH1

TH2

Macrophageactivating molecules

Others

B cellactivating molecules

Others

IFN-γ TNF-α GM-CSF

IL-2 IL-3 TNF-β

IL-4 IL-5

IL-3 IL-10 TGF-β

FIGURE 25–3 T-cell activation. Once a naive CD4⫹ T cell encounters its specific antigenic epitope, displayed in the context of MHC class II gene products and sufficient levels of co-stimulatory molecules on an appropriate antigen-presenting cell, it starts to proliferate and differentiates into an immature effector cell, termed Th0. This cell type carries the potential to become either a Th1 or a Th2 cell, which differ in their spectrum of cytokine production. See Color Plate

Introduction to Immune Reactions in the Peripheral Nervous System

such as B7-1 (CD80) and B7-2 (CD86), on the cell surface of APCs.36 The latter ligands bind to the CD28 or CTLA-4 receptors on T cells (Fig. 25–4).82,87 Genesis of inflammatory lesions in the peripheral nerve requires that activated T cells cross the BNB and enter the PNS. The complex process of their homing, adhesion, and transmigration64 has been studied in EAN in great detail (see Chapter 27).28 Although breakdown of the BNB is one of the earliest morphologically demonstrable events in lesion development, a major disruption of the BNB may not be necessary. Thus it has been postulated that activated T cells, irrespective of their antigenic specificity, can traverse a structurally intact barrier to execute immune surveillance. As in the systemic immune compartment, also in the PNS T lymphocytes also encounter their target antigen, recognize appropriate MHC molecules, and perceive additional co-stimulatory signals.28 This complex interaction prompts antigen-specific T cells to divide and to undergo clonal expansion. These locally expanded T lymphocytes then may exert various effects in peripheral nerve. First, CD4⫹ T cells of the Th1 inflammatory phenotype can damage myelin by secreting pro-inflammatory and myelinotoxic cytokines and operate by recruiting and instructing macrophages to produce and release an array of toxic molecules or to engage in increased phagocytotic activity. Second, CD4⫹ T cells of the Th2 phenotype may cause B-cell proliferation and transformation into plasma cells that manufacture antibodies against peripheral

APC CD40L

B7

MHC

Antigen TCR

CD28

CD40 T cell

FIGURE 25–4 Trimolecular complex. T-cell activation requires cognate dual recognition of peptide-MHC complex and co-stimulatory molecules displayed on antigen-presenting cells (APCs). T cell–APC interaction is further strengthened by reciprocal recognition of accessory molecules, such as B7-CD28 and CD40L-CD40. See Color Plate

563

myelin components. Finally, activated CD8⫹ or perhaps a subset of CD4⫹ T lymphocytes may be directly cytotoxic to Schwann cells. Conversely, specialized subpopulations of T cells may terminate the acute immuno-inflammatory process by the secretion of anti-inflammatory cytokines, such as IL-10 and TGF-␤ or other molecules (Fig. 25–5 and Table 25–1).17,33,37,70,72,98

B Lymphocytes The unique feature of B lymphocytes is their ability to express and secrete antibodies. Antibodies specifically bind antigens both in the recognition and the effector phase of a humoral immune response. Antibodies or immunoglobulins consist of two identical heavy chains and two identical light chains that are held together by disulfide bonds. The amino terminus of each chain possesses a variable domain that binds antigen through three hypervariable complementarity-determining regions. The C-terminal domains of the heavy and the light chains form the constant regions, which define the class and subclass of the antibody and govern whether the light chain is of the ␣ or ␤ type. Moreover, the Fc part governs the binding and activation of complement, and is responsible for binding of antibodies to macrophages (discussed subsequently). In the central part of the antibody molecule is the hinge region with variable glycoconjugate side chains that may be important for some of the non–antigen-specific physicochemical properties of antibodies, including hydrophobic bonds to membrane constituents.35,78 Five different classes of immunoglobulin (Ig) antibodies (IgD, IgM, IgG, IgA, and IgE), as well as four subclasses of IgG and two subclasses of IgA, are known, each of which exhibits different functional properties. Each type of antibody can be produced as a soluble circulating immunoglobulin molecule or as a cell surface molecule anchored through a transmembrane domain in the B-cell membrane, where it acts as the B-cell receptor. Activation of B cells to produce antibodies requires the help of CD4⫹ T cells if the antigens are proteins.62 T-cell help is not required if the antigens are polysaccharides and lipids. On the interaction with CD4⫹ Th cells, and only then, B cells undergo heavy chain isotype switching, which results in the production of antibodies with heavy chains of different classes. Besides the increase in number of antibodies secreted, the affinity of individual antibodies against the stimulating antigen will increase over time, as B cells expressing higher affinity receptors on their surface are selectively activated (Fig. 25–6).40 Autoantibodies targeting structures within the PNS have been implicated in the pathogenesis of immune-mediated diseases of the peripheral nerve (see Chapter 26). An example of the role for antibodies is provided by the production of an axonal neuropathy in rabbit EAN through immunization with GM1 gangliosides.106 Antibodies can conceivably

564

Neuroimmunology of the Peripheral Nervous System

Recognition

Activation

Effector phase

APC

Macrophage activation

APC Costimulation

B cell activation, antibody production

MHC TCR CD4+

CD4+

Memory T cell

IL-2R Cytokines e.g. IL-2

Inflammation

CD4+ T cell mediated cytolysis

CD8+

Infected target cell

FIGURE 25–5 Different stages of T-cell responses. First, T-cell activation requires contact with processed antigenic epitopes associated with MHC class II molecules displayed on antigen-presenting cells (APC) and perception of additional co-stimulatory molecules. Upregulation of receptors, such as that for interleukin-2 (IL-2), occurs, and IL-2 drives T cells into clonal proliferation during the activation stage. The effector stage is diverse. Different pathways lead to activation of macrophages and B cells. CD8⫹ cells cause MHC class II–restricted target cell lysis. CD4-driven responses can culminate in inflammatory tissue injury. See Color Plate

induce myelin damage by three mechanisms: (1) on binding to the Fc receptor of macrophages, they can direct these to the putative (auto)antigenic structures and induce so-called antibody-dependent cellular cytotoxicity; (2) by opsonizing target structures, they can promote their internalization by macrophages; and (3) on binding to the antigenic epitopes, they can activate the classic complement pathway with subsequent assembly of the terminal complement complex (C5b-9). This results in pore formation allowing calcium influx, which triggers myelin-integrated proteases that degrade the myelin sheath. The complement system is the major effector arm of humoral immunity. An important role of complement in the pathogenesis of inflammatory demyelination was underlined by the observation that decomplementation of animals with cobra venom factor or

inactivation by soluble complement receptor type 1 (CR1) partly suppressed EAN. Current knowledge suggests that complement may be important in recruiting macrophages into the endoneurium, in opsonizing myelin for phagocytosis, in amplifying ongoing inflammatory reactions, and in disintegrating the myelin sheath.41,75 Besides mediating structural damage, antibodies may impair nerve impulse propagation and neuromuscular transmission when binding at or close to the node of Ranvier or at the motor terminals.15,23,28,47,59

Mast Cells There is an extensive mast cell population within the PNS; however, their physiologic role remains largely unexplained.

Introduction to Immune Reactions in the Peripheral Nervous System

565

Table 25–1. Major Cytokines and Their Function Cytokine

Producer Cell

Actions

IFN-␥

T cells, NK cells

TNF-␣ TNF-␤ TGF-␤ IL-1 IL-2 IL-4 IL-5 IL-10 IL-12

Macrophages, NK cells, T cells T cells, B cells Chondrocytes, monocytes, T cells Macrophages, epithelial cells T cells T cells, mast cells T cells, mast cells T cells, macrophages B cells, macrophages

IL-13

T cells

IL-18

Activated macrophages

IL-21

T-helper cells

IL-23

Macrophages, dendritic cells

Macrophage activation, increased expression of MHC molecules and antigen processing components, Ig class switching Local inflammation, endothelial activation Killing, endothelial activation Inhibits cell growth, anti-inflammatory Fever, T-cell activation, macrophage activation T-cell proliferation B-cell activation, IgE switch suppresses Th1 cells Eosinophil growth, differentiation Macrophage suppression Activates NK cells, induces CD4⫹ T-cell differentiation to Th1 cells B-cell growth and differentiation, inhibits macrophage inflammatory cytokine production and Th1 cells Induces IFN-␥ production by T cells and NK cells, favors Th1 induction Enhances CD8⫹ T-cell response, regulates B-cell–mediated humoral immunity Induces IFN-␥ production and CD4⫹ T-cell differentiation to Th1 cells

IFN ⫽ interferon; Ig ⫽ immunoglobulin; IL ⫽ interleukin; MHC ⫽ major histocompatibility complex; NK cell ⫽ natural killer cell; Th1 ⫽ helper T lymphocyte type 1; TGF ⫽ transforming growth factor; TNF ⫽ tumor necrosis factor.

Immunoglobulin-free light chains are known to sensitize mast cells, such that a second encounter with the appropriate antigen results in mast cell activation,79 one of the putative scenarios that might be critical in the inflamed peripheral nerve. Through degranulation, endoneurial mast cells contribute to the genesis of immune-mediated demyelination by releasing vasoactive amines and arachidonic acid–derived metabolites that augment vascular permeability and disturb BNB integrity and nerve conduction.26,28 Pharmacologic experiments in which mast cell–stabilizing drugs such as reserpine and nedocromil prevented or attenuated the animal model support the concept of a pathogenic role of mast cell degranulation in PNS inflammation.14,86

Schwann Cells The contribution of Schwann cells to the initiation and termination of an immune response in the PNS is still a matter of debate.5 In principle it has been shown that they possess all immune molecules of the trimolecular complex as a basic prerequisite to interact with invading T cells. Schwann cells in vitro constitutively express at low levels MHC class I but no significant numbers of MHC class II molecules. Upregulation of MHC class I or expression of MHC class II can be induced by stimulation with interferon-␥ (IFN-␥) or upon co-culture with

activated T cells.6 Interestingly, TNF-␣ synergizes with IFN-␥ and further increases MHC expression, which also has functional importance.30 Moreover, adhesion molecules such as intracellular adhesion molecule-1 (ICAM-1/ CD54), which are constitutively expressed on cultured Schwann cells, are upregulated by these cytokines and may exert co-stimulatory functions. Cytokine-activated Schwann cells have been shown to process exogenous P2 protein or its neuritogenic peptide for antigen presentation.30 Under certain conditions they can even present endogenous myelin proteins such as myelin basic protein to autoaggressive lymphocytes in an MHC class II–restricted manner.101 Although the relevance of MHC class II expression on Schwann cells has been questioned in EAN (discussed subsequently),84 there is suggestive evidence that Schwann cells are capable of directly engaging in this immune function in vivo.73 A possible explanation for this discrepancy between the in vitro and in vivo situation may be provided by the finding that viable, electrically active neurons exert a regulatory function on glial cells by suppressing MHC expression. Schwann cells also express molecules that can terminate T-cell inflammation by induction of apoptosis (discussed subsequently) and downregulate immune functions. Fas and its ligand are central molecules of a family of death factors that govern T-cell survival in the immune system. Schwann cells apparently do not express Fas (CD95) or Fas

566

Neuroimmunology of the Peripheral Nervous System

A

N-terminus

N-terminus VH

Variable region

VL

CH1

Fab

CH2

Constant region

Bone marrow

CL

B0

Fc CH3

Ag B′

Ab P

M

Periphery

Lymphatic organs

C-terminus

B FIGURE 25–6 B-cell development. A, Basic structure of an immunoglobulin molecule. Each immunoglobulin molecule consists of two heavy and two light chains, linked by disulfide bonds. The amino-terminal domain of each chain is variable in sequence (Fab fragment), whereas the remaining domains are constant (Fc fragment). Antibodies recognize conformational epitopes via the Fab fragment, whereas the Fc fragment mediates binding to Fc receptors on immuno-inflammatory cells. B, B cells develop in the bone marrow (B0), where potentially self-reactive cells are eliminated soon after their antigen receptor is first expressed. In the lymph node the heterogeneous repertoire of mature B cells (B⬘) encounters antigens. Upon interaction with an antigen and specific helper T cells, B cells are activated to divide. Selected B lymphocytes differentiate into plasma cells (P), which secrete large amounts of antibodies, or into long-lived memory cells (M), which contribute to lasting protective immunity. See Color Plate

ligand (FasL) constitutively but can be induced to do so by pro-inflammatory Th1 cytokines within 48 hours.12,103 Cross-linking of Fas molecules on invading T cells by membrane-bound or secreted FasL could eliminate the autoaggressive immune effectors. This could explain T-cell apoptosis observed during the natural disease course of EAN (discussed subsequently). Second, expression of Fas on Schwann cell membranes could render them susceptible to T-cell attack. Apoptotic elimination of Schwann cells is observed during EAN and may further augment demyelination in the PNS.100 Probably local secretion of TNF-␣ is involved in regulation of apoptotic Schwann cell death.

Schwann cells produce IL-1, a potent cytokine that promotes T-cell activation and proliferation,89 and also IL-6, which may bias the local cytokine milieu to a Th2 type of reaction.11 Furthermore, Schwann cells can be induced in vitro to secrete prostaglandin E2 and thromboxane A2.20 These immunomodulators may inhibit or stimulate T cells, depending on their level of production. Schwann cells are also endowed with a cytokine-inducible nitric oxide synthase that is rapidly upregulated after simultaneous treatment with IFN-␥ and TNF-␣.31 Messenger RNA was detected within 12 hours, and nitrite secretion as a measure of NO production was detectable after 24 hours. Secretion of NO by

Introduction to Immune Reactions in the Peripheral Nervous System

Schwann cells exerts a strong suppressive effect on T-cell activation in a co-culture model. Schwann cell–derived NO intermediates have the potential to limit inflammatory demyelination, unless T cells are rescued by exogenous IL-2 (discussed subsequently). Another mediator of steroid hormones, lipocortin-1 (annexin-1), may also aid in the downregulation of inflammatory reactions.29 Schwann cells carry on their surface a number of regulatory complement proteins such as CR1 (CD35), decay accelerating factor CD55, and membrane cofactor proteins CD46 and CD59.58,83,96,97 This ensemble of proteins serves to attenuate the pro-inflammatory and demyelinating properties of activated complement.40,90,94 Conversely, although macrophages constitute the major source of complement at inflammatory foci, Schwann cells can also be induced by Th1 cytokines to generate the central complement component C3.22

TOLERANCE AND AUTOIMMUNITY The random generation of a highly diverse repertoire of B and T lymphocytes carrying specific receptors allows the immune system to recognize virtually any antigen, including autoantigens. These are recognized by selfreactive T cells or autoantibodies. Tolerance is the process that eliminates or downregulates such autoreactive cells. Consequently, a breakdown in this system can cause an autoimmune response or even autoimmune disease.

B-Cell Tolerance Autoantibodies are characteristic for several autoimmune diseases of the nervous system. They bind to surface molecules or receptors, leading to functional impairment of the affected cell. They can also bind to intracellular antigens and may thereby cause disease.60 The immune system has several mechanisms available to eliminate autoreactive B lymphocytes out of the B-cell pool: (1) by clonal deletion of immature B cells in the bone marrow67; (2) by deletion of autoreactive B lymphocytes in the T-cell zones of secondary lymphoid organs, such as lymph nodes or the spleen77; (3) by induction of cellular anergy (“functional inactivation”)32; and (4) by a process called receptor editing, which changes the receptor specificity once an autoantigen has been encountered.66 At present the relative contribution of these mechanisms in preventing autoimmune disorders in the PNS remains unclear.

T-Cell Tolerance The basic principle of T-cell tolerance is the deletion of self-reactive T lymphocytes in the thymus. During the developmental process of thymic education, self-reactive

567

T lymphocytes are usually eliminated by complex mechanisms consisting of positive and negative selection. To delete all self-reactive cells, the presence of all autoantigens is required in the thymus. However, this is not always the case.50 Further, some autoreactive T cells escape thymic education and enter the systemic immune compartment. Several mechanisms are required to keep T cells in check and maintain peripheral tolerance. Immunologic Ignorance Several mechanisms can cause immunologic ignorance. Antigens may be physically isolated from self-reactive T cells as a result of their separation in different biologic compartments (e.g., by the blood-brain barrier, marking the central nervous system as a potential immunologically privileged site).10 Also, the level of antigen expressed might be below a certain threshold required for T-cell activation,2,25 or the physical encounters with APCs are not sufficiently frequent. If tissues lack professional antigen presenters, resident antigens fail to activate T cells and are therefore ignored. Importantly, autoreactive T cells in such a setting remain functionally intact. Peripheral Clonal Deletion If abundant self-antigens in the periphery continually stimulate autoreactive T cells, these may succumb to activation-induced cell death through apoptosis. The lack of growth factors may also lead to elimination of autoreactive T cells. Inhibition T cells require the presence of co-stimulatory molecules when detecting antigens presented by an MHC molecule. CD152 (or CTLA-4) on T cells binds the co-stimulatory molecules CD80 (or B7-1) and CD86 (or B7-2) with a higher affinity than the co-stimulatory receptor CD28. Thus CD152 inhibits T-cell activation16 and results in anergy (i.e., functional unresponsiveness). Anergy If T cells recognize antigens presented without adequate levels of co-stimulatory molecules, they are rendered unresponsive and cannot even launch a response when restimulated by antigen presenters expressing sufficient amounts of co-stimulators. This state is termed anergy. T-cell anergy can also be induced by altered peptide ligands that contain modified T-cell receptor contact residues. Suppression Treg cells, by the production of various cytokines, inhibit or suppress the activation of T lymphocytes. Collective evidence suggests that these CD4⫹CD25⫹ lymphocytes (Treg cells) contribute notably to endogenous mechanisms that actively regulate induced autoimmune-mediated diseases of the nervous system (Fig. 25–7).18,21,74,88,99

568

Neuroimmunology of the Peripheral Nervous System

Immune reaction

Immunologic ignorance

Inhibition

Suppression

Deletion

Antigen

APC

APC

B7-1

MHC

CD28

TCR

APC

APC

APC FAS ligand FAS

CD152 T

T

IL-10 TGF-β

T

T

T

BNB

T

T

T

T

T cell activation

No cell activation

No cell activation

No cell activation

No cell activation

FIGURE 25–7 Peripheral T-cell tolerance. Because self-reactive T lymphocytes can escape thymic education, various mechanisms control the activity of these potentially autoreactive cells in the periphery. See text for details. BNB ⫽ blood-nerve barrier. See Color Plate

BREAKDOWN OF TOLERANCE If one of the regulatory mechanisms outlined earlier fails, the specific immune response is mounted against self-antigens, which leads to expansion of autoreactive effector T cells, generation of autoantibodies through T-cell help, or both, and may give rise to severe tissue damage9—a scenario that Paul Ehrlich termed horror autotoxicus over a century ago. Autoaggressive responses eventuate autoimmune disease. The autoimmune response persists because the immune system is not able to remove the offending autoantigen from the body; even worse, new hitherto hidden autoantigens could be released to amplify the response and broaden its epitope specificity, a process termed epitope spreading. The mechanisms of tissue damage in autoimmune diseases are essentially the same as those encountered in other inflammatory disorders. One hypothesis suggests that lymphopenia, often observed in autoimmune disorders and

caused by viral infections, may provoke a compensatory homeostatic expansion of T cells, forming self-reactive immune cell populations.57 Various mechanisms are operative to promote or prevent tissue damage once an autoimmune reaction has started. Recruitment of large numbers of host monocytes/macrophages and T cells and the release of cytotoxic cytokines and chemokines would serve to augment the injurious tissue reaction. Autoantibodies, either directly with the mediation of complement or via antibody-dependent cellular cytotoxicity, are also active partners. Conversely, a suppressive/inhibitory local environment with only few or not fully competent APCs, anatomic distance of autoreactive cells to target structures, anti-inflammatory cytokines, or inappropriate antibody concentrations may help to defend against autoimmune tissue destruction. Finally, induction of apoptosis in activated T cells may eliminate the autoreactive T cells before the full inflammatory cascade is set in motion.

Introduction to Immune Reactions in the Peripheral Nervous System

In none of the disorders of the PNS thought to be autoimmune in nature has the ultimate cause or the precise sequence of pathogenic events been unequivocally established. Structural similarities between microbial and self-antigens could activate autoreactive T or B cells, a mechanism termed molecular mimicry.3,19,69,105 T lymphocytes can recognize microbial as well as self-peptides with similar amino acid sequence.27,48 In contrast, a single T-cell receptor can recognize several peptides with various degrees in sequence homology.39 The principle of molecular mimicry has been invoked in the pathogenesis of GBS following infection with Campylobacter jejuni, Mycoplasma pneumoniae, or certain viruses such as cytomegalovirus (see Chapter 26).38,43,91,92,102 Another possible role in the genesis of autoimmunity has been attributed to a group of peptides derived from viral and bacterial pathogens, so-called superantigens. They have a distinct mode of binding to MHC molecules outside the groove that enables them to activate large numbers of T or B cells.71 Bystander activation, implicating the “accidental” activation of effector immune cells, may also play an important role.46,93

TERMINATION OF THE IMMUNE RESPONSE Termination of ongoing PNS inflammation can be mediated by continuous downregulation of the vicious cycle of the amplification-effector pathways. One such mode is silencing of macrophages, which appears to be mediated via the secretion of anti-inflammatory Th2 cytokines. These may also prevail over Th1 cytokines and their cellular sources. Another mechanism is apoptosis of autoimmune T cells, which occurs during the natural disease course of EAN.107 Although the vital mediators of T-cell apoptosis during neuritis have not been identified, various effector molecules, including NO, TNF-␣, and lipocortins, may act in concert to mediate T-cell death in the inflamed PNS. Accumulating evidence suggests that molecules of the innate immune system, such as complement components and pentraxins, have a role in the removal of apoptotic cells.65

REFERENCES 1. Abbas, A. K., and Sharpe, A. H.: T-cell stimulation: an abundance of B7s. Nat. Med. 5:1345, 1999. 2. Akkaraju, S., Ho, W. Y., Leong, D., et al.: A range of CD4 T cell tolerance: partial inactivation to organ-specific antigen allows nondestructive thyroiditis or insulitis. Immunity 7:255, 1997. 3. Albert L. J., and Inman, R. D.: Molecular mimicry and autoimmunity. N. Engl. J. Med. 341:2068, 1999.

569

4. Archelos, J. J., Previtali, S. C., and Hartung, H.-P.: The role of integrins in immune-mediated diseases of the nervous system. Trends Neurosci. 22:30, 1999. 5. Armati, P. J., and Pollard, J. D.: Immunology of the Schwann cell. Baillieres Clin. Neurol. 5:47, 1996. 6. Armati, P. J., Pollard, J. D., and Gatenby, P.: Rat and human Schwann cells in vitro can synthesize and express MHC molecules. Muscle Nerve 13:106, 1990. 7. Arnason, B. G. W., and Soliven, B.: Acute inflammatory demyelinating polyradiculoneuropathy. In Dyck, P. J., Thomas, P. K., Griffin, J. W., et al. (eds.): Peripheral Neuropathy, Vol. 3. Philadelphia, W. B. Saunders, p. 1437, 1993. 8. Arvidson, B.: Cellular uptake of exogenous horseradish peroxidase in mouse peripheral nerve. Acta Neuropathol. (Berl.) 37:35, 1977. 9. Bach, J. F.: Autoimmune diseases as the loss of active “self-control.” Ann. N. Y. Acad. Sci. 998:161, 2003. 10. Barker, C. F., and Billingham, R. E.: Immunologically privileged sites. Adv. Immunol. 25:1, 1977. 11. Bolin, L. M., Verity, A. N., Silver, J. E., et al.: Interleukin-6 production by Schwann cells and induction in sciatic nerve injury. J. Neurochem. 64:850, 1995. 12. Bonetti, B., Valdo, P., Ossi, G., et al.: T-cell cytotoxicity of human Schwann cells: TNFalpha promotes fasL-mediated apoptosis and IFN gamma perforin-mediated lysis. Glia 43:141, 2003. 13. Brigl, M., and Brenner, M. B.: CD1: Antigen presentation and T cell function. Annu. Rev. Immunol. 22:817, 2004. 14. Brosnan, C. F., and Tansey, F. A.: Delayed onset of experimental allergic neuritis in rats treated with reserpine. J. Neuropathol. Exp. Neurol. 43:84, 1984. 15. Buchwald, B., Weishaupt, A., Toyka, K. V., and Dudel, J.: Pre- and postsynaptic blockade of neuromuscular transmission by Miller-Fisher syndrome IgG at mouse motor nerve terminals. Eur. J. Neurosci. 10:281, 1998. 16. Chambers, C. A., and Allison, J. P.: Costimulatory regulation of T cell function. Curr. Opin. Cell Biol. 11:203, 1999. 17. Chen, W., and Wahl, S. M.: TGF-beta: receptors, signaling pathways and autoimmunity. Curr. Dir. Autoimmun. 5:62, 2002. 18. Chen, Y., Kuchroo, V. K., Inobe, J.-L., et al.: Regulatory T cell clones induced by oral tolerance: suppression of autoimmune encephalomyelitis in anti-myelin basic protein T cell receptor transgenic mice. J. Exp. Med. 265:1237, 1994. 19. Christen, U., and von Herrath, M. G.: Induction, acceleration or prevention of autoimmunity by molecular mimicry. Mol. Immunol. 40:1113, 2004. 20. Constable, A. L., Armati, P. J., Toyka, K. V., and Hartung, H. P.: Production of prostanoids by Lewis rat Schwann cells in vitro. Brain Res. 635:75, 1994. 21. D’Ambrosio, D., Sinigaglia, F., and Adorini, L.: Special attractions for suppressor T cells. Trends Immunol. 24:122, 2003. 22. Dashiell, S. M., Vanguri, P., and Koski, C. L.: Dibutyryl cyclic AMP and inflammatory cytokines mediate C3 expression in Schwann cells. Glia 20:308, 1997. 23. Dilley, A., Gregson, N. A., Hadden, R. D., and Smith, K. J.: Effects on axonal conduction of anti-ganglioside sera and sera from patients with Guillain-Barre syndrome. J. Neuroimmunol. 139:133, 2003.

570

Neuroimmunology of the Peripheral Nervous System

24. Dutronc, Y., and Porcelli, S. A.: The CD1 family and T cell recognition of lipid antigens. Tissue Antigens 60:337, 2002. 25. Ferber, I., Schörich, G., Schenkel, J., et al.: Levels of peripheral T cell tolerance induced by different doses of tolerogen. Science 263:674, 1994. 26. Frossi, B., De Carli, M., and Pucillo, C.: The mast cell: an antenna of the microenvironment that directs the immune response. J. Leukoc. Biol. 75:579, 2004. 27. Fujinami, R. S., and Oldstone, M. B.: Amino acid homology between the encephalitogenic site of myelin basic protein and virus: mechanism for autoimmunity. Science 230:1043, 1985. 28. Gold, R., Archelos, J. J., and Hartung, H.-P.: Mechanisms of immune regulation in the peripheral nervous system. Brain Pathol. 9:343, 1999. 29. Gold, R., Oelschlager, M., Pepinsky, R. B., et al.: Increased lipocortin-1 (annexin-1) expression in the sciatic nerve of Lewis rats with experimental autoimmune neuritis. Acta Neuropathol. (Berl.) 98:583, 1999. 30. Gold, R., Toyka, K. V., and Hartung, H. P.: Synergistic effect of IFN-gamma and TNF-alpha on expression of immune molecules and antigen presentation by Schwann cells. Cell. Immunol. 165:65, 1995. 31. Gold, R., Zielasek, J., Kiefer, R., et al.: Secretion of nitrite by Schwann cells and its effect on T-cell activation in vitro. Cell. Immunol. 168:69, 1996. 32. Goodnow, C. C., Crosbie, J., Adelstein, S., et al.: Altered immunoglobulin expression and functional silencing of self-reactive B lymphocytes in transgenic mice. Nature 334:676, 1988. 33. Gorelik, L., and Flavell, R. A.: Transforming growth factorbeta in T-cell biology. Nat. Rev. Immunol. 2:46, 2002. 34. Griffin, J. W., George, R., and Ho, T.: Macrophage system in peripheral nerves: a review. J. Neuropathol. Exp. Neurol. 52:553, 1993. 35. Harris, L. J., Larson, S. B., and McPherson, A.: Comparison of intact antibody structures and the implications for effector function. Adv. Immunol. 72:191, 1999. 36. Hartung, H.-P.: Pathogenesis of inflammatory demyelination: implications for therapy. Curr. Opin. Neurol. 8:191, 1995. 37. Hartung, H.-P., Pollard, J. D., Harvey, G. K., and Toyka, K. V.: Invited review–immunopathogenesis and treatment of the Guillain-Barré syndrome. Parts I and II. Muscle Nerve 18:137, 1995. 38. Hartung, H.-P., van der Meché, F. G. A., and Pollard, J. D.: Guillain-Barré syndrome, CIDP and other chronic immunemediated neuropathies. Curr. Opin. Neurol. 11:497, 1998. 39. Hemmer, B., Vergelli, M., Pinilla, C., et al.: Probing degeneracy in T-cell recognition using peptide combinatorial libraries. Immunol. Today 19:163, 1998. 40. Heyman, B.: Regulation of antibody responses via antibodies, complement, and Fc receptors. Annu. Rev. Immunol. 18:709, 2000. 41. Hila, S., Soane, L., and Koski, C. L.: Sublytic C5b-9-stimulated Schwann cell survival through PI 3-kinase-mediated phosphorylation of BAD. Glia 36:58, 2001. 42. Ho, T. W., McKhann, G. M., and Griffin, J. W.: Human autoimmune neuropathies. Annu. Rev. Neurosci. 21:187, 1998.

43. Hughes, R. A. C., Hadden, R. D. M., Gregson, N. A., and Smith, K. J.: Pathogenesis of Guillain-Barré syndrome. J. Neuroimmunol. 100:74, 1999. 44. Huppa, J. B., and Davis, M. M.: T-cell-antigen recognition and the immunological synapse. Nat. Rev. Immunol. 3:973, 2003. 45. Iezzi, G., Scheidegger, D., and Lanzavecchia, A.: Migration and function of antigen-primed nonpolarized T lymphocytes in vivo. J. Exp. Med. 193:987, 2001. 46. Infante-Duarte, C., Horton, H. F., Byrne, M. C., and Kamradt, T.: Microbial lipopeptides induce the production of IL-17 in Th cells. J. Immunol. 165:6107, 2000. 47. Jacobs, B. C., O’Hanlon, G. M., Bullens, R. W., et al.: Immunoglobulins inhibit pathophysiological effects of antiGQ1b-positive sera at motor nerve terminals through inhibition of antibody binding. Brain 126:2220, 2003. 48. Jahnke, U., Fischer, E. H., and Alvord, E. C. J.: Sequence homology between certain viral proteins and proteins related to encephalomyelitis and neuritis. Science 229:282, 1985. 49. Jonuleit, H., and Schmitt, E.: The regulatory T cell family: distinct subsets and their interrelations. J. Immunol. 171:6323, 2003. 50. Kamradt, T., and Mitchison, N. A.: Tolerance and autoimmunity. N. Engl. J. Med. 334:655, 2001. 51. Khalili-Shirazi, A., Gregson, N. A., Londei, M., et al.: The distribution of CD1 molecules in inflammatory neuropathy. J. Neurol. Sci. 158:154, 1998. 52. Kiefer, R., Dangond, F., Mueller, M., et al.: Enhanced B7 costimulatory molecule expression in inflammatory human sural nerve biopsies. J. Neurol. Neurosurg. Psychiatry 69:362, 2000. 53. Kiefer, R., Funa, K., Schweitzer, T., et al.: Transforming growth factor-beta 1 in experimental autoimmune neuritis: cellular localization and time course. Am. J. Pathol. 148:211, 1996. 54. Kiefer, R., Kieseier, B. C., Stoll, G., and Hartung, H. P.: The role of macrophages in immune-mediated damage to the peripheral nervous system. Prog. Neurobiol. 64:109, 2001. 55. Kieseier, B. C., Clements, J. M., Pischel, H. B., et al.: Matrix metalloproteinases MMP-9 and MMP-7 are expressed in experimental autoimmune neuritis and the Guillain-Barre syndrome. Ann. Neurol. 43:427, 1998. 56. Kieseier, B. C., Tani, M., Mahad, D., et al.: Chemokines and chemokine receptors in inflammatory demyelinating neuropathies: a central role for IP-10. Brain 125:823, 2002. 57. King, C., Ilic, A., Koelsch, K., and Sarvetnick, N.: Homeostatic expansion of T cells during immune insufficiency generates autoimmunity. Cell 117:265, 2004. 58. Koski, C. L., Estep, A. E., Sawant-Mane, S., et al.: Complement regulatory molecules on human myelin and glial cells: differential expression affects the deposition of activated complement proteins. J. Neurochem. 66:303, 1996. 59. Krampfl, K., Mohammadi, B., Buchwald, B., et al.: IgG from patients with Guillain-Barre syndrome interacts with nicotinic acetylcholine receptor channels. Muscle Nerve 27:435, 2003. 60. Matsumoto, I., Staub, A., Benoist, C., and Mathis, D.: Arthritis provoked by linked T and B cell recognition of a glycolytic enzyme. Science 286:1732, 1999.

Introduction to Immune Reactions in the Peripheral Nervous System 61. Maurer, M., Toyka, K. V., and Gold, R.: Cellular immunity in inflammatory autoimmune neuropathies. Rev. Neurol. (Paris) 158:S7, 2002. 62. Mitchison, N. A.: T-cell–B-cell cooperation. Nat. Rev. Immunol. 4:308, 2004. 63. Mueller, M., Wacker, K., Ringelstein, E. B., et al.: Rapid response of identified resident endoneurial macrophages to nerve injury. Am. J. Pathol. 159:2187, 2001. 64. Muller, W. A.: Leukocyte–endothelial-cell interactions in leukocyte transmigration and the inflammatory response. Trends Immunol. 24:327, 2003. 65. Nauta, A. J., Daha, M. R., van Kooten, C., and Roos, A.: Recognition and clearance of apoptotic cells: a role for complement and pentraxins. Trends Immunol. 24:148, 2003. 66. Nemazee, D.: Receptor selection in B and T lymphocytes. Annu. Rev. Immunol. 18:19, 2000. 67. Nemazee, D. A., and Burki, K.: Clonal deletion of B lymphocytes in a transgenic mouse bearing anti-MHC class I antibody genes. Nature 337:562, 1989. 68. Nikolich-Zugich, J., Slifka, M. K., and Messaoudi, I.: The many important facets of T-cell repertoire diversity. Nat. Rev. Immunol. 4:123, 2004. 69. Olson, J. K., Ludovic Croxford, J., and Miller, S. D.: Innate and adaptive immune requirements for induction of autoimmune demyelinating disease by molecular mimicry. Mol. Immunol. 40:1103, 2004. 70. O’Shea, J. J., Ma, A., and Lipsky, P.: Cytokines and autoimmunity. Nat. Rev. Immunol. 2:37, 2002. 71. Perron, H., Garson, J. A., Bedin, F., et al.: Molecular identification of a novel retrovirus repeatedly isolated from patients with multiple sclerosis. Proc. Natl. Acad. Sci. U. S. A. 94:7583, 1997. 72. Pestka, S., Krause, C. D., Sarkar, D., et al.: Interleukin-10 and related cytokines and receptors. Annu. Rev. Immunol. 22:929, 2004. 73. Pollard, J. D., Baverstock, J., and McLeod, J. G.: Class II antigen expression and inflammatory cells in the GuillainBarre syndrome. Ann. Neurol. 21:337, 1987. 74. Powrie, F., Carlino, J., Leach, M. W., et al.: A critical role for transforming growth factor-beta but not interleukin 4 in suppression of T helper type-1 mediated colitis by CD45RB(low) CD4⫹ T cells. J. Exp. Med. 183:2669, 1996. 75. Putzu, G. A., Figarella-Branger, D., Bouvier-Labit, C., et al.: Immunohistochemical localization of cytokines, C5b-9 and ICAM-1 in peripheral nerve of Guillain-Barre syndrome. J. Neurol. Sci. 174:16, 2000. 76. Rathmell, J. C., Thompson, C. B.: The central effect of cell death in the immune system. Annu. Rev. Immunol. 17:781, 1999. 77. Rathmell, J. C., Townsend, S. E., Xu, J. C., et al.: Expansion or elimination of B cell in vivo: dual roles for CD40- and Fas (CD95)-ligands modulated by the B cell antigen receptor. Cell 87:319, 1996. 78. Ravetch, J. V., and Bolland, S.: IgG Fc receptors. Annu. Rev. Immunol. 19:275, 2001. 79. Redegeld, F. A., Van Der Heijden, M. W., Kool, M., et al.: Functional role for Ig free light chains in immediate and delayed hypersensitivity responses. Inflamm. Res. 53(Suppl. 1):S6, 2004.

571

80. Sakaguchi, S.: Policing the regulators. Nat. Immunol. 2:283, 2001. 81. Sakaguchi, S.: Naturally arising CD4⫹ regulatory T cells for immunologic self-tolerance and negative control of immune responses. Annu. Rev. Immunol. 22:531, 2004. 82. Salomon, B., and Bluestone, J. A.: Complexities of CD28/B7: CTLA-4 costimulatory pathways in autoimmunity and transplantation. Annu. Rev. Immunol. 19:225, 2001. 83. Sawant-Mane, S., Piddlesden, S. J., Morgan, B. P., et al.: CD59 homologue regulates complement-dependent cytolysis of rat Schwann cells. J. Neuroimmunol. 69:63, 1996. 84. Schmidt, B., Toyka, K. V., Kiefer, R., et al.: Inflammatory infiltrates in sural nerve biopsies in Guillain-Barré syndrome and chronic inflammatory demyelinating neuropathy. Muscle Nerve 19:474, 1996. 85. Sebzda, E., Mariathasan, S., Ohteki, T., et al.: Selection of the T cell repertoire. Annu. Rev. Immunol. 17:829, 1999. 86. Seeldrayers, P. A., Yasui, D., Weiner, H. L., and Johnson, D.: Treatment of experimental allergic neuritis with nedocromil sodium. J. Neuroimmunol. 25:221, 1989. 87. Sharpe, A. H., and Freeman, G. J.: The B7-CD28 superfamily. Nat. Rev. Immunol. 2:116, 2002. 88. Shevach, E. M.: Regulatory T cells in autoimmunity. Annu. Rev. Immunol. 18:423, 2000. 89. Skundric, D. S., Lisak, R. P., Rouhi, M., et al.: Schwann cell-specific regulation of IL-1 and IL-1Ra during EAN: possible relevance for immune regulation at paranodal regions. J. Neuroimmunol. 116:74, 2001. 90. Song, W.C.: Membrane complement regulatory proteins in autoimmune and inflammatory tissue injury. Curr. Dir. Autoimmun. 7:181, 2004. 91. Susuki, K., Nishimoto, Y., Yamada, M., et al.: Acute motor axonal neuropathy rabbit model: immune attack on nerve root axons. Ann. Neurol. 54:383, 2003. 92. Susuki, K., Odaka, M., Mori, M., et al.: Acute motor axonal neuropathy after Mycoplasma infection: evidence of molecular mimicry. Neurology 62:949, 2004. 93. Tough, D. F., Sun, S., and Sprent, J.: T cell stimulation in vivo by lipopolysaccharide (LPS). J. Exp. Med. 185:2089, 1997. 94. Tsokos, G. C., and Fleming, S. D.: Autoimmunity, complement activation, tissue injury and reciprocal effects. Curr. Dir. Autoimmun. 7:149, 2004. 95. Van Rhijn, I., Van den Berg, L. H., Bosboom, W. M., et al.: Expression of accessory molecules for T-cell activation in peripheral nerve of patients with CIDP and vasculitic neuropathy. Brain 123(Pt. 10):2020, 2000. 96. Vedeler, C. A., Conti, G., Fujioka, T., et al.: The expression of CD59 in experimental allergic neuritis. J. Neurol. Sci. 165:154, 1999. 97. Vedeler, C., Ulvestad, E., Bjorge, L., et al.: The expression of CD59 in normal human nervous tissue. Immunology 82:542, 1994. 98. Walker, L. S.: CD4⫹ CD25⫹ Treg: divide and rule? Immunology 111:129, 2004. 99. Walker, L. S., and Abbas, A. K.: The enemy within: keeping self-reactive T cells at bay in the periphery. Nat. Rev. Immunol. 2:11, 2002. 100. Weishaupt, A., Bruck, W., Hartung, T., et al.: Schwann cell apoptosis in experimental autoimmune neuritis of the Lewis

572

101.

102.

103.

104.

Neuroimmunology of the Peripheral Nervous System rat and the functional role of tumor necrosis factor-alpha. Neurosci. Lett. 306:77, 2001. Wekerle, H., Schwab, M., Linington, C., and Meyermann, R.: Antigen presentation in the peripheral nervous system: Schwann cells present endogenous myelin autoantigens to lymphocytes. Eur. J. Immunol. 16:1551, 1986. Wim Ang, C., Jacobs, B. C., and Laman, J. D.: The GuillainBarre syndrome: a true case of molecular mimicry. Trends Immunol. 25:61, 2004. Wohlleben, G., Ibrahim, S. M., Schmidt, J., et al.: Regulation of Fas and FasL expression on rat Schwann cells. Glia 30:373, 2000. Wolf, K., Muller, R., Borgmann, S., et al.: Amoeboid shape change and contact guidance: T-lymphocyte crawling through fibrillar collagen is independent of matrix remodeling by MMPs and other proteases. Blood 102:3262, 2003.

105. Wucherpfennig, K. W., and Strominger, J. L.: Molecular mimicry in T-cell mediated autoimmunity: viral peptides activate human T cell clones specific for myelin basic protein. Cell 80:695, 1995. 106. Yuki, N., Yamada, M., Koga, M., et al.: Animal model of axonal Guillain-Barré syndrome induced by sensitization with GM1 gangliosides. Ann. Neurol. 49:712, 2001. 107. Zettl, U. K., Gold, R., Toyka, K. V., and Hartung, H. P.: In situ demonstration of T cell activation and elimination in the peripheral nervous system during experimental autoimmune neuritis in the Lewis rat. Acta Neuropathol. (Berl.) 91:360, 1996. 108. Zhu, J., Mix, E., and Link, H.: Cytokine production and the pathogenesis of experimental autoimmune neuritis and Guillain-Barre syndrome. J. Neuroimmunol. 84:40, 1998. 109. Zuniga-Pflucker, J. C.: T-cell development made simple. Nat. Rev. Immunol. 4:67, 2004.

26 Peripheral Nerve Antigens HUGH J. WILLISON, NORMAN A. GREGSON, GRAHAM M. O’HANLON, AND RICHARD A. C. HUGHES

Introduction Immune Responses in Peripheral Nerve Molecular Composition of Peripheral Nerve Cell Types, Subcellular Topographic Compartments, and the Blood-Nerve Barrier Schwann Cell and Myelin Antigens Neuronal Antigens Neurofilament Antigens Gangliosides Ganglioside Metabolism and Function

Voltage-Gated Calcium and Potassium Channels Molecular Mimicry and Microbial Antigens Campylobacter Infection Other Infections Molecular Mimicry Pathophysiologic Relevance of Autoantigens in Animal Models Galactocerebroside-Induced EAN MAG-Induced Neuropathy Models Ganglioside-Induced Neuropathy Models

INTRODUCTION The peripheral nervous system (PNS) is composed of a diverse array of cell types within various specialized compartments. It is an immunologically complex structure, and many sites within the PNS are subject to a wide range of inflammatory and autoimmune responses. Chapter 25, on immunologic responses, describes in detail the underlying immunologic principles pertaining to the PNS, and Chapter 97 describes the clinical immunology of PNS disease. The present chapter focuses particularly on a description of the PNS antigens that are capable of acting as autoimmune targets, and readers are referred to other chapters for further immunologic and clinical information. Some immunologic issues of special relevance to peripheral nerve, particularly innate immune mechanisms and molecular mimicry, are also included herein. The PNS contains many molecules potentially capable of acting as antigens that are either unique to the PNS, such as some of the myelin proteins, or highly enriched in it, such as gangliosides. Additionally, the PNS contains a large number of molecules that are common to other sites, both in the central nervous system (CNS), such as the

Pathophysiologic Relevance of Autoantigens in Human Neuropathies Neuronopathies Axonopathies Inflammatory Demyelinating Polyradiculoneuropathies Ganglioside Antibody–Related Diseases Galactocerebroside Charcot-Marie-Tooth Disease Paraproteinemic Neuropathies Protein Antigens Channelopathies

neuronal antigens Hu and CV2, and in non-neural regions of the body, such as basement membrane antigens. However, out of the many thousands of candidate antigens, only a limited number of molecular structures have been convincingly demonstrated to behave as peripheral nerve autoantigens. This chapter includes a description of these antigens, along with the relevant morphologic, pathophysiologic, and immunologic background, and the features of the antigenic repertoire within the PNS that make it an important site for both B- and T-cell responses to its constituent components.

IMMUNE RESPONSES IN PERIPHERAL NERVE The role of the immune system is to recognize the presence of foreign, usually microbial, material within an organism and to provide protection by the destruction and removal of this material. In autoimmune neuropathy, one of the major principles underlying nerve injury is a breakdown of this system, such that elements of the nerve are inadvertently seen as foreign. The provision of resistance 573

574

Neuroimmunology of the Peripheral Nervous System

to infection by the immune system involves the correct functioning of a large number of genes whose dysregulation can result in autoimmunity.294 In higher vertebrates the immune system can be divided into the innate and the adaptive subsystems, both being required for effective protection against infection and for aberrant development of autoimmune reactions. Central to this function is selftolerance, an immunologic awareness of self and the avoidance of a response against healthy self. This process is more complex than an intrinsic inability to react to selfantigens, since clearly this occurs both naturally and normally for peripheral nerve antigens. Antigen receptors of T and B cells show a theoretically infinite variability. T cells with stable receptors are controlled by exposure to selfantigens presented by major histocompatibility complex (MHC) molecules on the surface of thymic epithelia, dendritic cells, or macrophages. After maturation, cells showing strong binding or no binding with self-peptides and relevant MHC become apoptotic or anergic. Cells showing intermediate reactivity are preserved to form the population of “naïve” cells.203 B-cell development differs in that immunoglobulin gene rearrangement occurs both before and after exposure to antigen, and cells within the lymph nodes stimulated by antigen and T helper cells undergo further gene rearrangement and V-gene hypermutation, leading to antibody isotype switching and affinity maturation. Cells reactive with self-antigens normally fail to receive sufficient co-stimulation and become apoptotic or anergic. However, some B cells, particularly CD5 cells, do produce antibodies with broad reactivity that can react with certain autoantigens, including peripheral nerve glycolipids. For a new antigen to selectively stimulate and clonally expand T and B cells, the antigen must be presented under conditions that will optimize the interaction. The rapid recognition of foreign antigens by the molecular and cellular components of the innate system promotes inflammation and antigen processing and so promotes the adaptive response. The generation of a mature immune response takes place in the lymphoid organs, but the restimulation of memory cells and their rapid expansion can also occur in the peripheral tissues. The PNS, like the vascular system, penetrates all parts of the body and tissues, including the blood vessels. It therefore might be expected to have a high risk of trauma and infection. Nonetheless, like the CNS, it is relatively isolated from the immune system. The endoneurial vascular bed is provided by penetrating arterioles that feed an extensive pericyte-covered capillary network leading to small collecting venules. There are no large endoneurial veins. Tight junctions seal the endoneurial blood vessels and the perineurium, so that serum proteins, including antibodies, and passively migrating cells cannot access the nerve, except at peripheral nerve terminals and the dorsal root ganglia (DRGs). As in the CNS, the basal level of adhesion molecule expression on the endoneurial

endothelia is sufficient to allow some activated T cells to enter the nerve irrespective of their antigen specificity.250 Also like the CNS, peripheral nerves do not contain lymphatic vessels, and so the passage of immunocompetent cells from the endoneurium to the lymphatic organs is indirect via the blood.152 These structural features, and the absence of dendritic cells, suggest that the development of a primary immune response arising from self or foreign endoneurial antigens would be unlikely to occur. However, autoimmune neuritis can be induced experimentally by the transfer of activated T cells specific for certain nerve antigens into healthy syngeneic animals, implying that there is always some autoantigen presentation within the endoneurial compartment. It is not known which cells may be responsible for endoneurial presentation of antigens. Unlike the CNS, peripheral nerves are only sparingly populated by myeloid cells, and these are normally downregulated. Schwann cells do not normally express MHC molecules, but they clearly do so in certain infections such as leprosy, and they can also express the related molecules of the CD1 family.141,326 It is possible that the perivascular pericytes are the antigen-presenting cells in normal resting nerve. Both macrophages and mast cells are found in healthy nerves, and endoneurial macrophages are downregulated, although not to the same extent as brain microglia. Macrophage numbers increase markedly on injury and during inflammation, and there is even an increase in both number and level of activation of macrophages in the normal contralateral nerve after crushing one sciatic nerve.244 Nonetheless, the induced response is monophasic and brief, suggesting efficient suppression of unwanted autoimmune responses. Active induction of autoimmune neuritis is possible in mice and rats with the peripheral myelin proteins P2, P0, and PMP22 (see below), but this is highly strain dependent, indicating a strong genetic element in the susceptibility. The functioning of the innate immune system depends on soluble and cell surface receptors and effector molecules. Some of these are homologues of proteins found in invertebrates, where they may have similar functions. In general, mammals show an increase in diversity of these receptors and molecules, often as a result of gene duplication. The receptors, assisted by soluble factors such as lectins, complement, and antibodies, detect a variety of products from bacteria and viruses and result in the activation of the effector cells of the innate system, including macrophages, polymorphonuclear cells, natural killer (NK) cells, dendritic cells, mast cells, some cytotoxic T cells, and some B cells. Most of these cells bear more than one type of receptor, and the pattern of ligand binding to these can provide a pathogen-specific response, particularly demonstrated by the Toll receptors.319 The full interaction of myelomonocytic cells with other cells and microorganisms, as with the interactions of T cells and NK

575

Peripheral Nerve Antigens

cells, requires a clustering of a variety of receptors to facilitate phosphorylation of their cytoplasmic domains and further signaling.20,316 The effective functioning of the immune system in infection is further diversified by the fact that so many pathogens have acquired and evolved defensive mechanisms that often rely on functional mimicry of receptors in both the innate and adaptive systems (see below). Although certain receptors/ligands may be defined as critical for responsiveness, full function requires additional molecules. Through the activation of complement, many of the soluble factors, such as lectins and C-reactive protein, increase opsonization and facilitate phagocytosis by myeloid and polymorphonuclear cells. The blood-nerve barrier would normally exclude these proteins from the endoneurium, along with antibodies and complement proteins. The innate immune system strongly promotes the inflammatory process and the adaptive response. In autoimmune disease, the target tissue is usually defined by antibody or T cells but the development of inflammation in the tissue, and hence often the symptoms, is highly dependent on the activity of components of the innate system.108 Thus the extent of inflammation seen in acute and chronic anti-GM1 antibody–associated syndromes varies considerably, being high in the former (Guillain-Barré syndrome [GBS] associated with highaffinity immunoglobulin G [IgG]) and low in the latter (multifocal motor neuropathy [MMN] associated with lowaffinity immunoglobulin M [IgM]). Similarly, in chronic demyelination associated with anti–myelin-associated glycoprotein (MAG) IgM paraproteinemia, both antibody and complement are bound to myelin sheaths,320 but there is little or no inflammation and the development of the disease is very slow.92 The basis of self-tolerance in cells of the innate immune system is not indifference to self; rather, self-tolerance results from the inhibitory influence of a number of receptors reacting with surface molecules expressed normally in the tissues. Molecules that are currently considered to provide a “self ” signature for macrophages, NK cells, and other members of the innate system are CD200 (Ox-2), CD47, MHC class I molecules, and a variety of sialic acid–bearing glycoproteins and glycolipids. These molecules are expressed widely throughout all tissues, while their cognate receptors, CD200R, signal-regulatory protein (CD172a), NK inhibitory receptors, and the family of siglecs, are expressed by the innate effector cells.107,261 CD47 binding affects phagocytosis as well as dendritic cell maturation and polymorphonuclear cell function.60,180,235 Both CD200 and CD47 are prominent in the nervous system and found mainly on neurons, and their presence in the CNS contributes to the characteristic downregulated state of the endogenous myeloid cells, the microglia. There is little information on the status of these molecules in the peripheral nerves and ganglia.

The siglec family, which binds sialic acid–containing glycoconjugates, has at least 10 members, including sialoadhesin, CD22, CD33, and MAG,52 and may play a role in adjusting the activation threshold of immune cells. In quiescent cells they are mainly cis-engaged, but they become released by activation signals and reengaged by sialyl conjugates on neighboring normal cells. If neighboring cells do not express normal sialic acid conjugates (e.g., after exposure to bacterial or viral sialidase), then reengagement is not possible and activation becomes more probable. As described below, some infective agents that have acquired surface sialic acid may have subverted this function. It is thus clear that many properties of the innate immune system have an important bearing on any consideration of the regulation of peripheral nerve antigen responses, along with the more traditional considerations of adaptive T- and B-cell responses.

MOLECULAR COMPOSITION OF PERIPHERAL NERVE Cell Types, Subcellular Topographic Compartments, and the Blood-Nerve Barrier The PNS can be subdivided into functional, structural, or regional compartments that are all relevant to a consideration of antigenic components (Fig. 26–1).346 In functional terms, the PNS comprises motor, sensory, and autonomic components. In the case of motor neurons, cell bodies within the spinal cord extend axonal projections via the nerve trunks, to reach motor nerve terminals up to 1 m away. The cell body, myelin, and axon compartments are each relatively protected by the blood-brain and bloodnerve barriers and by the basal laminae and membranes of myelinating Schwann cells. Only the motor nerve terminal lies entirely outside the blood-nerve barrier, although it is itself protected by a cap of perisynaptic Schwann cells. Immunologic attack may be primarily directed at any of these sites to create distinctive phenotypes, and the wide range of motor nerve syndromes offers appropriate examples of this segregation. For example, certain autoimmune neuropathies principally target motor nerve myelin, such as MMN, or motor nerve axons, such as acute motor axonal neuropathy (AMAN). Furthermore, motor nerve involvement may be topographically restricted to particular anatomic sites, such as the restriction of the motor deficit to cranially innervated muscles in Miller Fisher syndrome (MFS). In the sensory arm of the PNS, the diversity of sensory fiber types and sizes, the different types of sensory ending, and the presence of the DRG with peripheral axons and central projections into the dorsal horn of the spinal cord contribute to an equal, or indeed greater, level of complexity than that seen in the motor arm. Autonomic nerves

576

Neuroimmunology of the Peripheral Nervous System

PERIPHERAL NERVE ANTIGEN COMPARTMENTS Dorsal root

Dorsal root ganglion

Nerve bundles Satellite cell Neuronal cell bodies Schwann cell basal lamina

Ventral root

Axon Adaxonal membrane Compact myelin

Nerve terminal

Periaxonal space

Internode

Abaxonal membrane

Node

Para node

Perisynaptic Schwann cell

Internode

Pre-synaptic membrane Post-synaptic membrane

Paranodal Schmidt-Lanterman loops incisure

Muscle Active zone

are also regarded as PNS components. Thus a multitude of clinical syndromes may selectively involve particular sites, based on antigen distribution, the accessibility to circulating immunopathogenic factors, and the innate immune considerations described above. In addition to the peripheral nerve terminals, areas of relative vulnerability to circulating immune factors are the dorsal and ventral roots and the DRG, owing to the partial deficiency of the bloodnerve barrier at these sites. Thus there may be a predilection for immune injury in the nerve roots that is independent of antigen composition, in that there is no evidence that the antigenic profile of the roots differs very substantially from more distal portions of the equivalent peripheral nerve trunk. Within this broad architecture, a number of topographic compartments exist. The PNS contains abundant blood vessels with microvascular endothelial cells that form the tight junctions characteristic of the blood-nerve barrier described above. The integrity of the endothelial layer is supported by endoneurial pericytes that form the basement membrane. Clearly such vascular structures are vital to the immunopathology of peripheral nerve, and may also be relevant to peripheral nerve vasculitis, but antigens expressed in these sites have not been studied in any detail. Anti–endothelial cell antibodies have been found to

FIGURE 26–1 Schematic diagram of the main antigen compartments in the peripheral nervous system. Regions that are potentially vulnerable to autoimmune attack and locations at which differential motor or sensory symptoms could be generated are indicated. (Adapted from Willison H. J., and O’Hanlon, G. M.: Anti-glycosphingolipid antibodies and Guillain-Barré syndrome. In Nachamkin, I., and Blaser, M. J. [eds.]: Campylobacter. Washington, DC, ASM Press, p. 259, 2000.)

a greater extent in rheumatoid arthritis complicated by peripheral neuropathy than in uncomplicated rheumatoid arthritis.270 However, a direct role for antigen-specific cellular or humoral mechanisms in the formation of vasculitic lesions in endoneurial blood vessels may be minor. The endoneurial compartment, containing intrafascicular nerve fibers, comprises connective tissue rich in collagen and extracellular matrix material, and is surrounded by the perineurium. Within the endoneurial compartment lie the nerve axons enveloped by Schwann cells and their basal laminae. The basal lamina is highly enriched in laminin-2, which binds dystroglycans and 62 integrin.318 Although basement membrane antigens, including various laminins, integrins, glycosaminoglycans, and collagens, are recognized as important autoantigens in several organ-specific autoimmune diseases, they have not been widely identified as such in autoimmune neuropathy syndromes. A single study has demonstrated the presence of antibodies to heparan sulfate glycosaminoglycans in up to one third of patients with GBS and other inflammatory neuropathies.246 Among the different fiber types, it is clear that there are features of the molecular organization common to all, and also fiber type specializations, that may both be of relevance to autoimmune neuropathy. All myelinated fibers contain

Peripheral Nerve Antigens

very high concentrations of ion channels at the node of Ranvier, particularly voltage-gated sodium channels in the nodal axolemma and rectifying potassium channels in juxtaparanodes, and the extent to which these channels act as antigens is a subject of much current research. Gangliosides are also believed to be highly expressed in nodal and paranodal regions, as discussed below.

Schwann Cell and Myelin Antigens Peripheral nerve myelin membranes are synthesized and maintained by the Schwann cells that enwrap large-caliber axons. These specialized membranes are key sites for the location of neuropathy-specific antigens, as described in detail below. Both myelinating and nonmyelinating Schwann cells are highly polarized, containing an adaxonal membrane that is separated from the axonal membrane by the periaxonal space, and an abaxonal membrane surrounded by the basal lamina. The myelin sheath also develops specializations associated with noncompact myelin at paranodal regions, Schmidt-Lanterman incisures, and the inner and outer mesaxons. The apposed cytoplasmic and extracellular surfaces of myelin-forming regions of the membrane compress to extrude almost all cytoplasmic components and extracellular fluid, resulting in the multilamellar structure that is characteristic of the internodal myelin sheath. This highly compacted membrane forms the bulk of the peripheral nerve myelin and has been the subject of extensive biochemical, metabolic, and immunologic investigation. It is believed that the molecular composition of compact myelin is common to all fiber types, although it remains possible that there could be subtle differences in the content of particular antigens, including gangliosides, as described below. The compact internodal myelin membrane is not directly accessible to antibody present in the extracellular space because it is in direct continuity with, and completely enveloped by, the Schwann cell plasma membrane and its mesaxonal and paranodal specializations. It is these latter membranes that form the outermost surface of the myelin sheath and are directly accessible to antibody. Penetration of soluble molecules from the extracellular fluid into the apposed membranes of compact myelin is limited. However, the intraperiod line is quite mobile and the extracellular surfaces can be separated by changes in pH and ionic composition, and it is clear that under some circumstances, such as that seen in the IgM paraproteinemic neuropathy associated with anti-MAG antibodies, antibody can penetrate into this compartment in surprisingly large amounts.192 Thus the concept of cryptic sites and the compartmentalization of antigens as a majorly limiting factor, even in compact myelin, should be viewed with some caution. The biochemical composition of isolated peripheral nerve myelin is remarkably simple, but also unique, having

577

a high proportion of lipid and relatively few individual protein components. This high lipid content is responsible for the low buoyant density of myelin, a property that has been exploited to prepare highly purified multilamellar myelin by a combination of density gradient and differential centrifugation. Protein accounts for 25% to 38% of the dry weight of peripheral nerve myelin, with the remaining approximately 70% being lipid with a high proportion of cholesterol and glycolipid, indicating the unique nature of the membrane. The importance of PNS myelin lipid, and in particular glycolipids and gangliosides, as antigens has been a major focus of research in recent years. The major myelin lipids are cholesterol, ethanolamine glycerophosphatide, sphingomyelin, and galactocerebroside. Ethanolamine plasmalogens, serine and choline glycerophosphatides, and sulfatides are also present but in smaller amounts. Isolated peripheral nerve myelin also contains low levels of several different gangliosides and complex neutral glycolipids that have been implicated as antigens in autoimmune neuropathy. These are discussed extensively below. Polyacrylamide gel electrophoresis (PAGE) of myelin proteins reveals one major band, the glycoprotein P0 (molecular weight [MW] 28 kDa). There are two less prominent bands, myelin basic protein (MBP), also termed P1 (MW 17 kDa), and P2 (MW 15 kDa), so called because of their position of migration in PAGE (Fig. 26–2). Several minor bands of immunologic interest can also be revealed by detailed separative techniques, including MAG (MW ~ 110 kDa) and peripheral myelin protein 22 (PMP22), so named because of its MW (22 kDa). Each of these proteins is capable of inducing an immune response following injection into experimental animals, and antibody and T-cell responses to some of these proteins have been reported in human inflammatory neuropathy. The P0 and PMP22 proteins are confined to compact myelin, but they are transmembrane proteins with extracellular domains that may be accessible to antibodies and immune cells (Fig. 26–3).274 MBP and P2 protein are relatively inaccessible, being confined to the major dense line of compact myelin corresponding to the apposed intracellular surfaces of the Schwann cell membrane. MAG is a transmembrane protein with a large, accessible extracellular domain, and is located in areas of the myelin sheath where Schwann cell cytoplasm is present. P0 Glycoprotein The P0 glycoprotein accounts for over 50% of peripheral nerve myelin protein and is absent from CNS myelin. Its primary sequence of 219 amino acid residues is highly conserved among species.269 It is a primitive member of the immunoglobulin supergene family, having a single IgG-like extracellular domain with a single glycosylation site, a very hydrophobic transmembrane domain, and a basic cytoplasmic domain.178 The extracellular domain is

578

Neuroimmunology of the Peripheral Nervous System

Bovine myelin CNS proteins

CNS

PNS

Rat myelin CNS

PNS

PNS proteins

MAG*

MAG*

P0 PLP 21.5K MBP

21.5K MBP 18.5K P1

18.5K MBP 17.0K MBP

17.0K MBP 14.0K MBP, PR P2

14.0K MBP

glycosylated at Asn93 with an N-linked nonasaccharide that contains the HNK-1 epitope, also present on PMP22 and MAG.36 There are three differences in the amino acid sequence of the extracellular domain of the rat P0 protein

FIGURE 26–2 Polyacrylamide gel electrophoresis (PAGE) in the presence of sodium dodecyl sulfate of bovine and rat central and peripheral nervous system myelin proteins. The proteins are each identified by size or name or both. The PAGE was performed according to the procedure of Laemmli.169 The approximate location of myelin-associated glycoprotein is also shown.

from that of the human, which would be expected to alter the tertiary and quaternary structure.29 The crystal structure of purified extracellular domain of P0 is consistent with the formation of tetramers at the extracellular surface

Extracellular Po

Intracellular

PMP22

MAG S MBP Gal-C S Sulfatide

Non-compact myelin

Compact myelin

FIGURE 26–3 Schematic depiction of the localization of peripheral nervous system myelin antigens. The dispositions of protein P0 tetramers, PMP22 dimers, and MBP monomers, as well as the glycolipids galactocerebroside and sulfatide, are shown. (Adapted from Scherer, S. S., and Arroyo, E. J.: Recent progress on the molecular organization of myelinated axons. J. Peripher. Nerv. Syst. 7:1, 2002.)

Peripheral Nerve Antigens

that associate with tetramers on the opposing surface (see Fig. 26–3). Both homophilic interactions between the proteins and interaction of protruding tryptophan residues and negatively charged lipids govern the compaction of the extracellular surfaces of myelin and maintain its spacing.281 The recent successful expression of the human protein in Escherichia coli and improvements in purification will assist future investigations of the antigenic properties of P0.29 Mice carrying a null mutation and lacking P0 do not form compact myelin, and mice lacking one P0 gene develop a late-onset demyelinating neuropathy, suggesting that the appropriate amount of P0 is important for myelin maintenance.276 Mutations of the P0 gene in humans cause hereditary motor and sensory neuropathy whose severity depends on the position of the mutation. The resulting clinical picture may be congenital hypomyelinating neuropathy, childhood-onset Déjérine-Sottas disease, or later onset Charcot-Marie-Tooth (CMT) disease type 1b (see Chapter 71).178 There is evidence that immune responses are important in controlling the phenotype of P0-deficient mice. Martini and colleagues crossed mice heterozygously deficient for P0 with null mutants for the recombinant activating gene 1 or tumor necrosis factor-a receptor. Both these mutants had severely impaired T-cell responses and less severe neuropathy.200 PMP22 Protein PMP22 is a minor peripheral nerve myelin protein of great importance because its gene is duplicated in the most common form of CMT disease and deleted in hereditary neuropathy with liability to pressure palsies.293 The C- and N-terminal domains are thought to be intracellular, and it is usually considered as a transmembrane protein with four membrane-spanning regions and two extracellular domains, although alternative structures in which the second and third hydrophobic domains are also extracellular have been proposed.309 The first extracellular domain has 38 amino acid residues and is variably glycosylated, with the carbohydrate portion including the HNK-1 epitope shared by P0 and MAG, suggesting a cell adhesion function.292 The second extracellular domain has 14 amino acid residues and is not glycosylated.69,178 In the rat, expression of PMP22 begins soon after birth and reaches adult quantities after 3 weeks.293 It is glycosylated at Asn41 of the first extracellular domain, and is then translocated to the cell membrane.69 Nonglycosylated protein does not reach the cell membrane and is broken down. Its function is not entirely clear; mice carrying null mutations for PMP22 do myelinate, but myelination is delayed.2 The protein is therefore likely to be important in stabilizing myelin. It has been shown to complex with P0 at the cell surface in HeLa cells expressing both proteins, an interaction that was not mediated by carbohydrate.68 In addition, PMP22 probably plays a role in regulating

579

Schwann cell proliferation because synthesis coincides with cessation of Schwann cell division.214 The gene for PMP22 was originally described as the growth arrest– specific gene (gas-3), which is expressed by resting but not proliferating mouse fibroblasts.295 The messenger RNA (mRNA) for PMP22 is present in many tissues, although the amounts are lower than in myelin and the promoter driving its transcription in myelin differs from that in other tissues.31 The PMP22 protein is a component of tight junctions between epithelial cells, where it is closely associated with the transmembrane junctional proteins zonula occludens 1 and occludin.226 It might therefore be important in the function of Schwann cell tight junctions. P2 Protein The P2 protein, forming 2% to 15% of myelin protein, is a basic protein confined to peripheral nerve myelin, and of its 131 amino acids, there are large sequence differences at nine positions among different mammalian species. It is situated in the period line of myelin corresponding to the apposed intracellular surfaces of the Schwann cell. Xray crystallography and modeling predict a barrel-shaped molecule composed of beta pleated sheets surrounding a space that might contain fatty acid, and it may therefore be involved in fatty acid transport.132 There are no reports of mutants lacking P2 or of mutation causing animal or human neuropathy. Its intracellular location renders it relatively inaccessible to the immune system, but if the myelin sheath is disrupted and phagocytosed, P2 might be processed by antigen-presenting cells and become accessible to the T-cell receptor. P2 has been studied extensively as an antigen in relation to experimental allergic neuritis (EAN), as described in Chapter 27. Myelin Basic Protein PNS myelin contains another basic protein, previously called P1, that was found to have the same amino acid sequence as MBP in the CNS181 and accounts for 5% to 15% of PNS myelin protein, about half of that in the CNS. The major form has a molecular weight of 18 kDa, but there are several isoforms with molecular weights ranging from 14 to 21 kDa, arising from differences in splicing of the primary transcript and variations in glycosylation and phosphorylation. Partial loss of the MBP gene in the shiverer mutant mouse causes failure to synthesize compact CNS myelin but does not prevent the formation of normal peripheral nerve myelin.39 Like P2, MBP is located in the period line, where it may play a role in myelin compaction (see Fig. 26–3), but the normality of peripheral nerve myelin in shiverer mice suggests that this role is redundant or readily replaced by another molecule, perhaps P2. It would not be expected to be accessible to the immune system unless the integrity of the Schwann cell basal lamina and Schwann cell membranes was disturbed.

580

Neuroimmunology of the Peripheral Nervous System

peripheral neuropathy in adult life. This finding indicates that, in as yet unidentified ways, periaxin is essential for the maintenance of a normal myelin sheath.285 Mutations in the periaxin gene have been identified as causing autosomal recessive demyelinating CMT disease (CMT4F) and Déjérine-Sottas disease.27,94,305 There is a single report of polyclonal IgM antibodies to periaxin being discovered in the serum of a small proportion of patients with IgG paraprotein-associated demyelinating neuropathy and with diabetes mellitus type 2 and neuropathy.174

Myelin-Associated Glycoprotein MAG exists in large and small forms (L- and S-MAG) as a result of alternative splicing of the C-terminal sequences, with unglycosylated weights of 67 and 72 kDa and, when glycosylated, weights of about 100 and 110 kDa, respectively. L-MAG and S-MAG predominate in myelinating and mature Schwann cells, respectively. MAG is situated in the periaxonal space of noncompacted myelin, at paranodes, mesaxons, and the Schmidt-Lanterman incisures (see Fig. 26–3).314 MAG is a transmembrane protein and a member of the immunoglobulin superfamily, with five extracellular immunoglobulin-like domains and an intracellular C-terminal tail (Fig. 26–4).209 The extracellular domains have N-linked glycosylation sites to which the HNK-1 epitope may be attached, and the large extracellular domains may interact homophilically with identical molecules on the opposing surface to maintain the periaxonal space. The intracellular domain of L-MAG contains a phosphorylation site that may act as a target for tyrosine kinases. It belongs to the siglec family, having a sialic acid binding site between the first and second extracellular domains, and plays a role in cell adhesion and Schwann cell axon signaling.357 The sialic acid binding site binds to ganglioside GD1a and other gangliosides with an 2–3 sialic acid linkage.330 In culture, MAG functions as a cell adhesion molecule, promotes or inhibits neurite outgrowth depending on conditions, and binds to collagens I and G.192,272

Myelin Glycolipids Peripheral nerve myelin is enriched with a wide range of glycolipids and some gangliosides that are important antigens for humoral responses. Gangliosides are principally enriched in neuronal membranes and are described in detail in the next section; the glycoconjugate structures that are predominantly localized to human PNS myelin are described here. Structural diagrams of the major glycolipids are shown in Table 26–1. Galactocerebroside (GalC), one of the simplest glycosphingolipids, comprising galactose linked to ceramide (monogalactosylceramide), was the first glycolipid identified as an important antigen in experimental autoimmune encephalitis and EAN, in which it is the principal target for complement-fixing antibodies. Galactosulfatide, also called sulfatide, is GalC sulfated on the third carbon of galactose. GalC and sulfatide are very typical and prominent lipids in myelin from both the CNS and PNS. The arrangement of GalC and sulfatide within the lipid bilayer is asymmetrical, with the glycolipids being localized to the outer leaflet, where they are available for antibody binding, as opposed to the phospholipids, which are principally localized to the inner leaflet. The importance of GalC in maintenance of myelin membranes has been demonstrated by targeted disruption

Periaxin Periaxin, a component of the dystroglycan-dystrophin– related protein-2 complex, has recently been described as a possible peripheral nerve antigen (see below). This complex acts to link the Schwann cell cytoskeleton to the extracellular matrix. Periaxin knockout mice myelinate normally, but then go on to develop a demyelinating

S S

S S

S S

S S

S S

L-MAG N

P C

C

P = Phosphorylation site

= Sialic acid binding site

S S

S S

S S

S S

S S

S-MAG N

= Glycosylation site

FIGURE 26–4 Diagram of the molecular structure of the long (L-MAG) and short (S-MAG) forms of myelinassociated glycoprotein. The glycosylation and sialic acid binding sites are shown. (After a diagram in Milner, R. J., Lai, C., Nave, K.-A., et al.: Organisation of myelin protein genes: myelin-associated glycoprotein. In Duncan, I., Skoff, R. P., and Colman, D. [eds.]: Myelination and Demyelination. New York, New York Academy of Sciences, p. 254, 1990.)

Gangliosides with an (2-8) linked disialosyl group

Terminal Gal(1-3)GalNAcconfigured gangliosides

Ganglioside and Glycolipid Antigenic Motifs

GQ1b

GD3

GD2

GD1b

GT1b

GM1

Structure

Table 26–1. Structural Diagrams of the Major Glycolipids*

asialo-GM1

Table continued on following page

Glial Structures Members of this group are present in compact myelin,49 especially small sensory fibers.230 GM1 is more abundant in VR than DR.45,229 Areas selectively bound by ligands include Schmidt-Lanterman incisures,211 paranodal end segments of the myelin sheath,49,82,83,161,165,211 nodal gap,273 and abaxonal Schwann cell cytoplasm.211 Neuronal Structures Ligand binding was observed with motor neurons,50,195 at the nodal axonal surface,82,243 at the neuromuscular junction,172,173,230,231,275,313 and to DRG neurons.165,195,230,231 Other Ligand binding was observed to muscle spindles231 and a series of Gal-GalNAc–bearing glycoproteins.9,49,191,132 Sensitization with GM1 ganglioside led to high anti-GM1 antibody titers and the development of acute-onset flaccid paralysis in rabbits.365 Endothelial cells express ganglioside antigens, including GM1 and GD1b. Circulating antiganglioside antibodies may damage cell-to-cell attachments, thus weakening the blood-nerve barrier and contributing to development of autoimmune demyelinating neuropathy.136 Glial Structures GD1b and GQ1b are enriched in the paranodal myelin.44,156,161,165,273 Anti-GD2 antibodies label the myelin sheath.366 GQ1b is enriched in the cranial nerves serving the extraocular muscle.45 GQ1b/GD3-reactive antibodies bind to cytoplasmic channels on the surface of myelinating Schwann cells, and to perisynaptic Schwann cells at the NMJ.87 Neuronal Structures Abundant in cultured DRG neurons.37 Antibodies bind DRG neurons,165,195,230,231,329 and are able to lyse them.234 Anti-GD1a antibodies induce ataxic neuropathy in rabbits.104,162 GT1b is confined to the axolemma.282 Thin-layer chromatography with immunostaining showed that GT1a is present in human oculomotor and lower cranial nerves.146 Complex gangliosides are present at the NMJ but are redundant for normal synaptic function.35 Antibodies bind the NMJ and can disrupt normal structure and function.34,87,232,249,347 Others Muscle spindles are stained strongly by an antibody to polysialylated gangliosides.347

Anatomic Localization and Immunopathologic Features

GalC

Fucosyl gangliosides

Chol-1 antigens

GM2

Terminal GalNAc-Gal-NeuNAcconfigured gangliosides

Fuc-GM1

GM1a

GM1b

Fuc-GD1b

GQ1ba

GalNAc-GM1b

GalNAc-GD1a

GD1a

Structure

Terminal NeuNAc(2-3)Galconfigured gangliosides

Ganglioside and Glycolipid Antigenic Motifs

Table 26–1. Structural Diagrams of the Major Glycolipids*—Continued

GalC is a major glycolipid component of myelin298 and may also be expressed by nonmyelinating Schwann cells.129 Intraneurally injected antibodies accumulated on the Schwann cell, particularly at the paranodal areas and Schmidt-Lanterman clefts.301 Sera from rabbits with EAN induced by sensitization with GalC binds to cultured Schwann cells.186

Anti–Fuc-GM1 antibodies bind to some small DRG neurons and their surrounding satellite cells.163,164

A series of at least six very minor gangliosides that are specific to cholinergic neurons.5,102,120,259 At least some of these are expressed on the cell body of motorneurons227,339 and at the NMJ.61 Proteins with cross-reactive carbohydrate sequences may also be present.62

Suggested as antigenic targets in autoimmune motor neuropathy.117,118,160,363 In biochemical studies, GM2 and GalNAc-GD1a were present in human peripheral nerves,117 but were undetected by immunolocalization studies.233

GD1a is a significant component of nerve in biochemical studies.45,229 High-titer IgG anti-GD1a antibodies selectively bind to motor, but not sensory, nerve nodes of Ranvier.58 GM1b has been detected in the brain but not extensively studied in the perpheral nervous system.77,151

Anatomic Localization and Immunopathologic Features

Gal

NeuNAc GalNAc

SO3

Glu

Fuc

SGLPG

GlcNAc

Glc UA

Ceramide

SGPG and SGLPG are found by biochemcial analysis in the myelin and axons of peripheral nerve.13,46,148 MAG is localized to regions of the myelin sheath in which the membranes are uncompacted, including the periaxonal membrane, the paranode, and Schmidt-Lanterman incisures.199,315 Non–MAG-reactive antibodies bind at the outer surface of myelin sheaths. Faint staining is also visible at the axolemmal-myelin interface, but compact myelin was not stained.355

Neuropathy-associated antibodies cross-react with MAG252 and the myelin proteins P0 and HNK-1.28,36 Sural biopsies from patient with these antibodies show deposits within the myelin sheath and signs of demyelination.317

Glial and Neuronal Structures In peripheral nerve, the most common sites in antibody binding studies include axons, resident macrophages and Schwann cytoplasm, especially the Schmidt-Lanterman incisures, nodes of Ranvier, and perineuronal sheath of satellite cells in DRG.190,219,254 Sural biopsies from patients with these antibodies show IgM and complement product C3d bound to the myelin sheaths of almost all fibers.76 A significant component of both sensory and motor nerves.45,229 Absent from several cranial nerves displaying CNS characteristics (e.g., optic nerve).45

CNS  central nervous system; DR  dorsal root; DRG  dorsal root ganglion; GalC  galactocerebroside; IgG  immunoglobulin G; LM1  sialosylneolactotetraosylceramide; MAG  myelin-associated glycoprotein; NMJ  neuromuscular junction; SGLPG  sulfated glucuronyllactosaminyl paragloboside; SGPG  sulfated glucuronyl paragloboside; VR  sulfated glucuronyl paragloboside; VR  ventral root. * The groupings described in this table are illustrative of cross-reactive epitopes reported in the literature, but are not intended as a definitive list. Gangliosides share common structural motifs outside of those presented here, and even within the presented groupings, it should be noted that an individual ganglioside can fall within two or more groups (e.g., GD1b). Equally, the fact that two ganglioside species appear in the same group does not necessarily imply that an antibody to one will cross-react with the other.231 Glycolipids that appear dissimilar when presented as shown here may adopt tertiary structures that display common motifs, and through the same process similar-looking structures may be antigenically different. † Gal  galactose; NeuNAc  neuraminic acid; GalNAc  N-acetylgalactosamine; Glu  glucose; Fuc  fucose; GluNAc  N-acetylglucosamine; Glc  glucosamine.

Structural components†

SGPG

SGPG/SGLPG

SO3

LM1

SO3

LM1

Sulfatide

584

Neuroimmunology of the Peripheral Nervous System

of UDP-galactose ceramide galactosyltransferase (CGT), the key enzyme in galactolipid biosynthesis. Such mice have a phenotype marked by slow conduction and delayed myelin breakdown that has been characterized in detail.47 Sulfatide has also been identified as an important antigen in autoimmune neuropathy, and it has been possible to delete the galactosylceramide sulfotransferase gene that specifically inhibits the synthesis of sulfatide, while maintaining GalC. These modifications may have a much less severe phenotype than the CGT knockout,109 but they exhibit ultrastructural abnormalities in the paranodal regions, and it appears that sulfatide may play a critical role in the localization and maintenance of ion channels in the underlying axonal membrane.122 Peripheral nerve has an abundance of neolacto-series gangliosides that are localized mainly in myelin.228 The term LM1 is used for the sialosylneolactotetraosylceramide, which is also known as sialosyl paragloboside, and Hex-LM1 is used for sialosyllactosaminyl paragloboside. LM1 and Hex-LM1 are major monosialosyl glycolipids in human peripheral nerve myelin and have been identified as autoantigens in both acute and chronic autoimmune neuropathies. The function of LM1 and Hex-LM1 has not been elucidated, although they are very widely distributed in human peripheral nerve myelin, as well as other tissues and cells such as the red blood cell. Another important class of glycolipid antigens is sulfated glucuronyl paragloboside (SGPG) and its higher lactosaminyl homologue, sulfated glucuronyllactosaminyl paragloboside (SGLPG).42,171,253 Both were discovered directly as a result of studying IgM paraproteins reactive to MAG, with which they share immunologically crossreactive structures, from patients with chronic polyneuropathy.13 SGPG and SGLPG also have structures similar to that of LM1, except for a 3-sulfated glucuronic acid instead of sialic acid on the terminal saccharide chain. The subcellular distribution of these sulfated glucuronyl glycolipids has been studied in myelin- and axon-enriched fractions of motor and sensory nerves from human subjects. SGPG is slightly more abundant in sensory nerve axolemmal fractions than motor nerve fractions, which may account for the relative predominance of sensory features in patients with anti-SGPG antibodies. However, it should also be recognized that these antibodies crossreact with MAG and other myelin proteins that may not have such a distribution.

Neuronal Antigens A number of molecules that are principally localized to the neuronal and axonal compartments of peripheral nerve have also been identified as relevant to the pathogenesis of human autoimmune neuropathy. These antigens include the paraneoplastic antigens Hu, CV2, and amphiphysin, which are each associated with the development of para-

neoplastic sensory neuronopathies. In addition, axonal neurofilaments, several ion channels, and numerous gangliosides are all reported as important antigens in diverse neuropathy syndromes, principally manifested by neuronal or axonal dysfunction. Neurologic paraneoplastic syndromes provide some of the clearest examples of autoimmune disease in the nervous system, affecting diverse central and peripheral structures, and are recognized by the presence of high-affinity antibodies against specific and, in many cases well-characterized, neural antigens. The origin of the immune response is presumed to be against the tumor cells in which the neural antigens are inappropriately expressed. In syndromes in which the antigens are expressed on the plasma membrane, the immunopathology is primarily dependent on antibody and antibody-mediated processes such as complement fixation. However, in those paraneoplastic syndromes involving an intracellular antigen, the mechanism of pathogenesis is not clear, but there is evidence for the involvement of cytotoxic T cells.3,306,307 In most of these syndromes the CNS is the predominant site of disease, and these are not discussed here. Paraneoplastic Neuronal Antigens: Hu, CV2, and Amphiphysin The paraneoplastic subacute sensory neuronopathy (PSSN) associated with anti-Hu antibodies,56 and less commonly anti-amphiphysin antibodies, and the more recently recognized subacute polyneuropathy associated with anti-CV2 antibodies7 are the three paraneoplastic syndromes in which peripheral sensory neurons are most clearly involved. The autoimmune nature of sensory neuropathy associated with small cell lung carcinoma (SCLC) was originally detected by complement fixation,53,342 and the antibodies were shown to react preferentially with neuronal nuclei and a small group of proteins of 32 to 42 kDa, including the Hu antigen.4,90 The Hu gene is located on the human chromosome 1 at 1p34,215,279 and the antigen has several forms generated by alternative splicing of four exons, the most prominent product being HuD (Fig. 26–5).303 Additional closely related gene products HuC and Hel-N1a have also been recognized.187,268 The proteins characteristically contain three RNA recognition regions, the first two in tandem separated from the third domain by a basic segment. Sequence homology with the Drosophila proteins ELAV and Sex-Lethal suggests that Hu proteins bind specifically to AU-rich elements in the untranslated 3 regions of some classes of mRNA,154,187 and this has been confirmed by direct analysis.187,334 Those mRNAs with AU-rich regions are frequently short-lived species, but the binding of proteins such as HuD increases their stability and lifespan. The human anti-HuD IgG binds to those regions of the molecule containing the first two RNA binding domains. At least two different epitopes exist, one of which is in the

Peripheral Nerve Antigens

FIGURE 26–5 Diagrammatic view of HuD. The molecule contains three globular regions, a, b, and c, with homologies to RNA-binding proteins. These contain -helical and random coil structures (gray) and -structures (black) that contain the major RNA-binding residues. Serum from patients with paraneoplastic disease contains antibodies directed against discrete isotopes in the globular regions a and b that do not interfere with their RNA-binding property. (Data from Wang X., and Tanaka Hall, T. M.: Structural basis for recognition of AU-rich element RNA by the HuD protein. Nat. Struct. Biol. 8:141, 2001.)

vicinity of each of the RNA binding domains.197 However, the binding of the antibodies does not appear to interfere with the binding to RNA.154 Although the Hu proteins are highly characteristic and specific to all central, peripheral, and enteric neurons, they are also expressed by most SCLC cell lines,197,279 as well as cells of other tumors associated with PSSN. The presence of the antigen in the associated tumor satisfies the hypothesis that the autoimmunity arises during an immune response against the tumor. One explanation for the presence of an immune response against the protein could be that the tumor expresses a mutated form of HuD. However, this was not found to be the case in a study of 26 lung cancer cell lines, three from patients with anti-Hu antibodies, of whom one had a paraneoplastic disorder.279 Low levels of anti-Hu antibodies have been found in cases of SCLC in the absence of any neurologic disease.204 Histopathologic examination of SCLC tumors has indicated that paraneoplastic disease is associated with strong expression of MHC antigens by the tumor cells.55

585

However, these observations were clearly made after the initiation of the immune response, and the expression of MHC antigens could reflect the strong inflammatory nature of the response in PSSN. Although high-titer IgG antibodies against Hu and related antigens occur in the circulation and cerebrospinal fluid (CSF) in PSSN, it has not been possible to demonstrate any cytopathic effect of this antibody on neurons or SCLC cells in culture, even though the antibody can be demonstrated in the nuclei of the exposed cells.113 Similarly, it has not been possible to demonstrate the development of any neurologic syndrome in animals following passive transfer of patient IgG,63 or by active immunization against the Hu antigens.289 High titers of anti-Hu antibodies have been found in association with inflammation of the gut and loss of enteric neurons in the absence of a tumor and without any other peripheral or central signs.291 T cells are prominent in the cellular infiltrates in the nervous system, and these are predominantly CD8 T cells.91 Some circulating CD4 T cells have been shown to respond to recombinant HuD,24 and more recently CD8 cytotoxic T cells have been found to kill syngeneic fibroblasts containing HuD and expressing MHC class I molecules.306 In tissue culture, neurons can be induced to express MHC class I molecules on their surface in the presence of interferon-, particularly when the neurons are electrically silent.221 Greater MHC class I expression has been seen as a result of neural activity.51 The expression of MHC class I molecules could well provide the means of targeting the specific cytotoxic CD8 T cells and explain the neuronal killing in this disease. However, neuronal expression of human leukocyte antigen has not been demonstrated in pathologic material,91 whereas in the DRG, satellite cells may normally express both MHC class I and II molecules.89 Even if neurons can be induced to express MHC class I antigens, the problem remains of how such an induction is initiated if it requires the presence of inflammation. The CV2 antigen was first identified as a paraneoplastic antigen when serum and CSF of some patients with neuropathy associated with cerebellar ataxia and cancer were found to react with a subpopulation of oligodendrocytes in the rat brainstem, cerebellum, and spinal cord. The IgG antibodies bound to a 66-kDa protein found in the newborn rat brain.111 Cloning and Northern blot analysis demonstrated a 3.8-kb mRNA that was brain specific and increased from rat E17 to reach a maximum at P7, after which it declined to negligible levels except in the spinal cord and cerebellum.251 The sequence indicated that the antigen was an Unc-like phosphoprotein (Ulip), and the CV2 antigen was assigned to the Ulip-3 group, which in the human and rat corresponds to collapsin response mediator protein-1 (CRMP-1). Collapsins are receptors for the extracellular semaphorins that induce growth cone collapse and are therefore also important in axonal growth and plasticity.

586

Neuroimmunology of the Peripheral Nervous System

Sera from patients bind variously to CRMP-1, -2, and -3 prepared as recombinant protein in HeLa cells, but all bind to CRMP-5.110 CRMP-5 shows only 50% sequence identity to the other four CRMPs, which share 68% to 74% identity with each other. There is therefore a degree of uncertainty as to the identity of the “antigen” complementary to the anti-CV2 antibodies, but on the basis of a reported screen, CRMP-5 appears to be the likely candidate.110 However, the possibility must remain that more than one antibody specificity exists. The identity of the anti-CV2 antigen as CRMP-5 is compatible with the fact that it is also present in oligodendrocytes, and that CV2positive sera do bind to retinal neurons, neurons in the dentate gyrus, and the rostral band of olfactory neurons as well as peripheral axons.8,110 The putative antigen for CV2 antibodies is known to be concerned with axonal growth, and the antibodies do react with peripheral axons,7 but again CRMP proteins are intracellular and the question of access remains unresolved. There is no information about whether such antibodies are pathogenic, or about whether patients have T-cell responses to CRMP. CRMP-5 has been demonstrated to occur in SCLC cells, satisfying the requirement for tumor expression of the neural antigen.359 It is not uncommon for antibodies against more than one neural antigen to occur in paraneoplastic syndromes. Thus, as indicated above, anti–CRMP-3 antibodies have been found in association with anti-Hu but also antibodies against voltage-gated Ca2 channels176 and, less commonly, amphiphysin.67 Amphiphysin-1 was originally noted as a neural target in patients with stiff-man syndrome and breast cancer.57 The 128-kDa amphiphysin-1 and -2 play a role in clathrin-mediated endocytosis, forming a heterodimer interacting with the GTPase dynamin to act in vesicle budding from the membrane.341 Amphiphysin is localized predominantly at synaptic terminals and is passed anterogradely along axons by fast axonal transport, but only about a third returns by retrograde flow, suggesting local breakdown.179 Antibodies against amphiphysin are mostly directed against the C-terminal half of the molecule and have been found in PSSN,67 opsoclonus-myoclonus,21 and a small proportion of patients with SCLC, irrespective of the presence of paraneoplastic disease.267 They have also been found in a patient with a sensorimotor axonopathy.245 The occurrence of the antibodies in the absence of neurologic symptoms raises questions as to their pathogenic significance.267 There is no record of a model neuropathy inducible by hyperimmunization of a laboratory animal.

Neurofilament Antigens Neurofilaments (NFs) belong to the class of intermediate filament proteins and are prominent structures of neurons and axons. The filaments contain three proteins of 200, 160, and 68 kDa (NF-H, NF-M, and NF-L, respectively); their proportions and phosphorylation vary in different

parts of the neuron. In axons NFs are highly phosphorylated, and NF-L forms the central core of the structure. Antibodies against NF have been found in a variety of neurologic conditions and in healthy control subjects.19,288,297,310 It is possible that in healthy controls such antibodies represent part of the “natural antibody” repertoire, but it is also likely that in neurodegeneration these are upregulated and diversified as a response to the release of antigen.59 In relation to natural antibodies, 4 of 75 patients with paraprotein-associated neuropathy had an IgM antibody against NF-H,224 and a monoclonal immunoglobulin A (IgA) from a patient with amyotrophic lateral sclerosis (ALS) was reactive with the 200-kDa NF.263 There is little indication that such antibodies are pathogenic, although the monoclonal IgA from the patient with ALS was found to cross-react with an undefined neuronal surface protein.263 Once again, it is difficult to see how antibodies against intracellular antigens could be strongly immunopathogenic. However, both IgG and IgM can be endocytosed by neurons with peripheral processes and retrogradely transported to the perikaryon.72–74,194 Anti-NF antibodies have been detected in patients with paraneoplastic optic neuropathy, and NF was found in the associated tumors.150 Other studies have found that NF expression in SCLC tumors is not common, unlike that of HuD.325 However, neurofilaments of 160 and 68 kDa have been demonstrated in the cells of thymic cortical tumors in association with myasthenia gravis.201,202,277 It has been suggested that the antiaxonal antibodies in these patients represent antibodies against the NF-M expressed in these tumors.202,277 Epitope mapping showed that the antibodies reacted with the sequence KSPVEEK, which was repeated five times in the NF molecule, and also with a cytoplasmic epitope (KSAIGIK) of the a subunit of acetylcholine receptor.277 The immunity against the NF-M expressed by the tumor might account for the anti–acetylcholine receptor antibodies in myasthenia patients with cortical thymic tumors. However, it has not been established by experiment that such an immune response can induce myasthenia gravis.

Gangliosides Structure and Nomenclature The term ganglioside refers to the large family of glycosphingolipids that contain sialic acid linked to the oligosaccharide core. In humans, most ganglioside sialic acid is in the N-acetyl form, as opposed to the N-glycosyl configuration that is common to many other species. The nomenclature of Svennerholm is usually used for gangliosides.43,302 The designations are based on the four major brain gangliosides with a ganglio-series tetraose chain of neutral sugars (i.e., asialo-GM1) that are sialylated in different positions. However, there are around 100 structurally distinct gangliosides, and it has become difficult to create a single, easily workable nomenclature.

587

Peripheral Nerve Antigens

In the initial classification, four gangliosides, GM1, GD1a, GD1b, and GT1b, were designated to belong to the G1 series, wherein “G” stands for ganglio-series ganglioside. The four major gangliosides differ with regard to the number and position of their sialic acids, with M, D, and T indicating mono-, di-, and trisialosyl groups, respectively. Thus there are two disialosylgangliosides, GD1a and GD1b. Although “b” is normally used to designate gangliosides with a disialosyl group attached to the internal galactose (so-called b-series gangliosides), the term GM1b is used for the monosialosyl gangliotetraosyl ceramide, in which the sialosyl group is attached to the terminal galactose, in contrast to GM1a (normally more simply referred to as GM1), in which the sialic acid is on the internal galactose. When three sialic acids link to the internal galactose of the ganglioside, they are designated to belong to the “c” series. Now that the biosynthetic pathway of gangliosides of the ganglio-series has been in large part elucidated, it is evident that this early description of the “a,” “b,” and “c” series predicted the crucial role of the sialyltransferases in ganglioside biosynthesis, as shown in Figure 26–6. Gangliosides lacking the terminal galactose, preterminal galactosyl-N-acetylgalactosamine, or internal galactose are assigned the number 2, 3, or 4, respectively.

Ganglioside Metabolism and Function Gangliosides are present throughout the body but are very highly concentrated in the nervous system. They are particularly enriched in neuronal membranes, but are also minor

constituents of myelin.175,311,358 Gangliosides are amphipathic molecules comprising a hydrophobic ceramide moiety that is embedded in the lipid membrane and a hydrophilic oligosaccharide moiety, which is exposed to the cytosolic compartment or to the extracellular space. Thus gangliosides are highly associated with membrane structures both in the plasma membrane, where their density is high, and in a wide variety of intracellular membrane compartments, including nuclear and mitochondrial membranes, endosomes, lysosomes, and the Golgi apparatus. Gangliosides cycle through these various intracellular compartments to and from the plasma membrane as small vesicles and through endosomal sorting. Gangliosides are synthesized in the Golgi apparatus by the stepwise addition of saccharides in reactions catalyzed by glycosyltransferases and sialyltransferases, and are degraded in a reverse fashion by the stepwise removal of saccharides in lysosomes. They can accumulate abnormally in lysosomal storage diseases as a result of mutations in the enzymes or transport proteins involved in these degradative pathways. Targeted disruption of genes involved in ganglioside biosynthesis has demonstrated the importance of gangliosides in a wide range of neural processes, including the maintenance of peripheral nerve integrity and function. Thus null-mutant mice lacking the gene coding for the glycosyltransferase 1,4-GalNac-transferase (1,4-GalNac-T; GM2 /GD2 synthase) lack complex gangliosides, and consequently neuronal membranes only bear the simple gangliosides GD3 and GM3, whose expression is upregulated.304 Such animals develop age-related axonal degeneration in

CER

Glucose

Galactose

GaINAc

NeuAc

CER

CER

CER LacCer

CER

CER GA2

FIGURE 26–6 Schematic representation of ganglioside biosynthesis. Ganglioside nomenclature is according to Svennerholm.302 Gangliosides are synthesized through addition of monosaccharides in a stepwise fashion by glycosyltransferases and sialyltransferases (arrows). CER  ceramide; GalNAc  N-acetylgalactosamine; LacCer  lactosyl ceramide; NeuAc  neuraminic acid or sialic acid. (Adapted from Bullens, R. W. M., O’Hanlon, G. M., Wagner, E. R., et al.: Complex gangliosides at the neuromuscular junction are essential receptors for autoantibodies and botulinum neurotoxin but redundant for normal synaptic function. J. Neurosci. 22:6876, 2002.)

CER GM3

CER

CER GM2

CER GA1

CER

CER

GD1a

CER

“a-series”

GT1c CER

GT1b CER

GT1a

GT2

GD1b CER

CER GD1c

CER

CER

CER

GT3

GD2

GM1

GM1b

CER GD3

GQ1c CER

GQ1b “b-series”

GP1c “c-series”

588

Neuroimmunology of the Peripheral Nervous System

the CNS and PNS, and also develop demyelination in peripheral nerves with pathologic features resembling those seen in MAG knockout mice, MAG being a known ligand for complex brain gangliosides.330 In double-knockout mice that lack both GM2/GD2 and GD3 synthase, expressing only GM3, peripheral nerve degeneration has also been observed, combined with mutilating skin lesions and nerve fiber proliferation in the skin, suggesting that reduced sensory function is present.119 These types of studies are continually revealing important facets of ganglioside biology in peripheral nerve. With respect to their behavior as antigen targets, the predominant relevant ganglioside localization is the plasma membrane, where the oligosaccharide portion of gangliosides is extracellularly sited. Gangliosides are not uniformly distributed in plasma membranes but are concentrated in association with cholesterol and phospholipids in clusters referred to as functional or lipid rafts.290,298 Such rafts are important sites for signal transduction, being highly enriched in glycosylphosphatidylinositol-anchored proteins on the extracellular face and Src kinases on the cytoplasmic face. Rafts are also associated with small plasma membrane invaginations called caveoli, which are sites of endocytosis. The high concentration and close proximity of gangliosides in lipid rafts may enhance their ability to bind antibody with high avidity or affinity. The pathogenic effects of antiganglioside antibodies are likely to depend not only on how avidly they bind gangliosides but the extent to which the target gangliosides are intimately involved in modulating neuronal function. Additionally, activated complement components may in turn affect the normal functioning of ganglioside raft–associated proteins. The relative contribution of these factors may vary from site to site, and among antiganglioside antibodies of differing reactivity. Relatively little is currently known about the distribution of gangliosides in different membrane and raft components throughout the nervous system, and the relevance that this distribution may have to their role as antigens in autoimmune neuropathy remains unclear. However, when considering the pathogenic relationship between the presence of an antibody and neuropathy, it is clearly important to have detailed knowledge of the glycosphingolipid composition and distribution within the PNS both in humans and in experimental animals, and furthermore in species from which gangliosides are purified for experimental and diagnostic use. These issues are not straightforward because the regulation of ganglioside expression and methods for the analysis and purification of gangliosides are complex. Gangliosides are developmentally regulated and spatially segregated, varying among different peripheral nerve fiber types and among different species. A complete map of the ganglioside composition of human nerves, and a comparison among species used for experimental modeling, would be a valuable resource but might still have limitations.

The biochemical and immunohistologic approaches to establishing the distribution of gangliosides each have their merits and limitations. Biochemical analysis has been useful to identify significant differences in the ganglioside composition of different nerves and can reveal subtle differences in the overall lipid composition. However, this approach is limited by the pleomorphic composition of tissue and the lack of information about microanatomic distribution. In such circumstances the second approach, that of immunohistology or other in situ ligand-binding studies (e.g., using ganglioside-binding bacterial toxins such as cholera toxin) can reveal fine structural detail about ganglioside distribution at the cellular and subcellular level (Fig. 26–7). For these studies, high-quality reagents such as affinity-purified antisera or monoclonal antibodies are essential. Many antiganglioside antibodies are not monospecific but may crossreact with structurally similar gangliosides and other glycoconjugate antigens, making extrapolation of results to ganglioside localization difficult. Biochemical and immunohistologic approaches characterized by homogenization and tissue sectioning, respectively, may expose gangliosides that normally occupy cryptic sites (e.g., within compact myelin). Gangliosides may also be complexed with intrinsic ligands such as siglecs, as discussed above. Thus diverse factors may misrepresent the ganglioside array that would be visible to circulating antibodies in physiologic environments. Furthermore, gangliosides can be heterogeneously distributed within a membrane, as in functional rafts as described above. The antigen density and the surrounding lipid environment can also markedly influence the ability of antiganglioside antibodies to bind; thus failure to detect a ganglioside by immunohistology does not necessarily indicate its biochemical absence.188 A high local concentration of ganglioside may allow for good immunohistologic detection, whereas a ganglioside that is evenly distributed throughout a membrane may have the same total tissue concentration in biochemical evaluation, yet not be detectable by immunohistology. Interpretation of both biochemical and immunohistologic studies thus requires caution. The accessibility of PNS gangliosides to circulating antibodies, being protected in their neural environment by the blood-nerve barrier and other factors, is important.188 One explanation for the lack of CNS involvement in antiglycolipid antibody–associated neuropathy, despite the wide distribution of gangliosides in the CNS, would thus be the protection from autoimmune attack afforded by the blood-brain barrier. Blood-nerve barrier injury could be mediated by antiganglioside antibodies reacting with glycolipid antigens expressed by intraneural microvascular endothelial cells.137 Ganglioside Localization to Specific Nerve Sites and Fiber Types The simplest explanation for particular antiganglioside antibody–associated neuropathies being confined to a motor or

Peripheral Nerve Antigens

589

FIGURE 26–7 Immunofluorescent labeling of mouse tissue using antiganglioside antibodies. The monoclonal IgM antibodies Ha1 (human347) and CGM3 (mouse87) are both reactive with gangliosides possessing a disialosyl moiety, including GQ1b, GD1b, and GD3. Disialosyl gangliosides are detected throughout the cell bodies of dorsal root ganglion neurons (A), as delineated by antineurofilament (NF) antibodies (B). Disialosyl gangliosides are largely confined to the nerve terminal and associated nonmyelinating Schwann cells (C) at the neuromuscular junction, as defined by NF and -bungarotoxin (BTx) labeling (D). At the nodes of Ranvier, they occupy a focal nodal position suggestive of an axonal distribution (E) compared to the more diffuse paranodal distribution of ganglioside GM1, as detected by cholera toxin B subunit (CT) binding (F). Bars: 20 m.

sensory clinical phenotype is that the two systems contain different gangliosides; this issue has been addressed in many experimental studies, including immunohistologic analyses.86 A comparison of total ganglioside composition of human spinal roots showed that GM1 (associated with antibodies in motor neuropathy) is relatively enriched in the ventral roots compared with the dorsal roots.229 Similarly, the cranial motor nerves supplying the extraocular muscles contain higher concentrations of GQ1b, the ganglioside antigen associated with ophthalmoplegia.45 However, from these studies it is also apparent that key gangliosides are also present at sites unaffected by the disease process. Thus in many cases the absolute tissue distribution of gangliosides is an insufficient explanation for the regional localization of the clinical pathology. The DRG is a particularly interesting site to consider in these respects and illustrates some of the anomalies described above. Functional subpopulations of DRG neurons can be distinguished by the expression of lactoseries and globo-series carbohydrates that correlate with

their peptide and enzymatic phenotype.64,65 It is very likely that ganglio-series antigens are also selectively distributed and that such differences may underlie the specific nature of sensory deficits. In the rodent DRG, the GM1 ligand cholera toxin B subunit and anti-GM1 antibodies selectively identify a subset of DRG neurons. In contrast, cholera toxin and anti-GM1 antibodies bind the majority of neurons in the human DRG.230,231 Thus GM1 is clearly present in the human DRG in abundance, yet the clinical phenotype of anti-GM1–associated antibody neuropathy is strikingly devoid of sensory features. However, anti-GD1b antibodies are highly associated with sensory neuropathy and bind to the vast majority of DRG neurons.161,165,195 Disialosylated gangliosides, including GD1b, GT1b, GQ1b, GD3, and GD2, are prominent gangliosides in cultured rodent DRG neurons, and antidisialosyl antibodies can lyse these.234 These observations also suggest that care should be exercised when attempting to establish an animal model of a human disease.

590

Neuroimmunology of the Peripheral Nervous System

The node of Ranvier is another key site of injury in autoimmune neuropathy. Immune attack directed at antigenic determinants located at the paranodal Schwann cell surface may lead to paranodal demyelination, whereas antigens targeted on the exposed axolemma may result in axonal degeneration, both of which would result in conduction failure. Ligand-binding studies have suggested that GM1, GD1b, and polysialosylated gangliosides are enriched in the paranodal myelin loops of peripheral nerve (see Figs. 26–1 and 26–7).165,273 In the ocular motor nerves affected in MFS, GQ1b is particularly enriched at nodes of Ranvier.44 With respect to the neuronal components of the node of Ranvier, toxin- and antibody-binding studies have identified gangliosides on paranodal and internodal axolemma49,82,83,282 and the adaxonal membrane.211 Similarly, an antibody reactive with disialosylated gangliosides has been shown to bind to internodal axolemma and/or adaxonal Schwann cell cytoplasm.347 Antibodies to GD1a are associated with pure motor axonal neuropathy, and preferentially stain ventral root axons in comparison with dorsal root axons, indicating a good correlation between ganglioside localization and phenotype in this example.86,283 With some clinical justification and for a variety of hypothetical reasons, the presynaptic neuromuscular junction (NMJ) has recently been considered a potential target vulnerable to autoimmune attack in GBS. First, it lacks a blood-nerve barrier, thereby readily allowing access to circulating autoantibodies. Second, it is the site for other paralytic antibody-mediated diseases, including myasthenia gravis and Lambert-Eaton myasthenic syndrome. Third, it is rich in gangliosides, including GQ1b, GM1, and GD1a.34,87,173,230,231,249,275,313,347 Fourth, it is the binding site for a wide range of bacterial toxins that also use gangliosides as ectoacceptors.344 In particular, cholera and tetanus toxins are readily taken up into nerve terminals, loaded into synaptic vesicles, and ultimately transported back to the motor neuron cell body.103,333 As a result of this property, enzymic conjugates of cholera toxin, or its binding B subunit, are frequently used as retrograde neuronal markers. As expected, histologic analysis has demonstrated cholera toxin and anti-GM1/GD1b antibody binding to the NMJ.172,173,230,275 Antibodies reactive to polysialosylated gangliosides also bind to the NMJ (see Fig. 26–7). Some of the -series gangliosides specific to cholinergic neurons (Chol-1 antigens) are also expressed at the mature NMJ, making them potential targets for autoantibodies.63

Voltage-Gated Calcium and Potassium Channels Voltage-gated ion channels are glycosylated multimeric proteins forming a potential pore that opens and closes according to the membrane potential, leading to changes

in ion flux across the membrane. Voltage-gated potassium (K) channels have four  subunits that form the pore and a variety of  accessory subunits. There are many different K channels, and mutations of them cause episodic ataxia or epileptic syndromes.153 Antibodies against one of them, the dendrotoxin-sensitive fast potassium channel, cause Isaacs’ syndrome or acquired neuromyotonia (see below). In voltage-gated calcium (Ca2) channels, the pore is formed by a single  subunit with four domains each having six transmembrane segments and thus resembling the K channel. Different mutations of one Ca2 channel cause familial hemiplegic migraine, episodic ataxia, and spinocerebellar atrophy. Autoantibodies to presynaptic calcium channels present at motor nerve terminals are found in Lambert-Eaton myasthenic syndrome.

MOLECULAR MIMICRY AND MICROBIAL ANTIGENS Infection has long been considered to initiate autoimmune disease and particularly autoimmune neuropathy. Infective agents provide a large number of antigenic peptides and other structures that may stimulate naïve and memory T and B cells. The high levels of cross-reactivity shown by T cells means that some cells are likely to be stimulated that will also react with self-peptides.203 Infection can lead to an intense inflammatory activation, which increases the possibility of reactivating anergic autoreactive lymphocytes. Nonetheless, the situations in which infection stimulates autoimmunity that have been demonstrated are few and specific. In autoimmune neuropathies, GBS is the clearest example, but infection might also be a factor in chronic inflammatory neuropathies. An association between acute cytomegalovirus (CMV) infection and antiMAG paraproteinemic polyneuropathy was suggested362 but not confirmed.121,193 The symptoms of GBS usually become apparent some weeks after an infective episode, and thus the identification of the pathogen by its isolation is difficult. As a result, most of the information relating to pathogen identity is derived from testing for serum antibodies, but there are some problems with this approach. The serologic tests used are not always absolutely specific, and different centers use different assays, so that detailed comparisons are difficult. Even so, certain specific organisms are particularly strongly associated with the onset of the neuropathy. Of these, Campylobacter jejuni has been found to be the most prevalent agent, followed by CMV and Mycoplasma pneumoniae.95,127,145,213,322,324,352 An association with Haemophilus influenzae is currently controversial but seems likely. It is possible that the level of association of individual organisms with GBS differs according to geographic region and their general incidence or genetic subtype.

591

Peripheral Nerve Antigens

Campylobacter Infection As already indicated, antiganglioside antibodies are autoantibodies frequently found to be associated with GBS, and are strongly associated with prior C. jejuni enteritis.10,257,332,367 Several studies have also demonstrated that the AMAN/motor form of GBS is more frequent following C. jejuni infection.105,205,256,257,328,352 Campylobacter jejuni is a spiral, flagellated bacillus with a large genetic and phenotypic diversity. It is widespread in the environment, including fresh water. In humans, particularly in the northern hemisphere, it is now probably the most common agent of sporadic enteritis and appears to be a normal component of avian gut flora. The early finding that the lipopolysaccharide (LPS) fraction from C. jejuni isolates from GBS patients bears carbohydrate sequences identical to gangliosides suggests that this is the antigenic stimulus of the antiganglioside antibodies frequently found in patients.15–17,93,126,271,364 Immunization of experimental animals with LPS derived from isolates from GBS and MFS leads to the production of antiganglioside antibodies.6,87 However, neither the type of GBS nor the occurrence of antiganglioside antibodies is absolutely associated with C. jejuni infection. Recent genotyping studies on C. jejuni, using multienzyme electrophoresis, ribotype and flagellin gene (flaA) polymorphism, amplified fragment length polymorphism, and microarray, have all confirmed the high degree of genetic diversity.66,70,217 These studies have so far failed to find any clonal clustering of GBS isolates. In Japan and China serotype O:19,155,222 and in South Africa serotype O:41,170 are common isolates from GBS/MFS but not from uncomplicated enteritis patients. It is likely that some O:19 isolates are clonal and share an ancestor with Penner O:41 strains.217 However, the full extent of the ancestral relationships between GBS-associated strains of C. jejuni remains unknown. Endogenous synthesis of N-acetylneuraminic acid (sialic acid) from N-acetylmannosamine in C. jejuni is encoded by three neuB genes, only one of which, neuB1, affects LPS synthesis.185 Both (2-3)- and (2-8)-sialyltransferases are present,84,321 enabling the synthesis of mono-, disialo-, and trisialo-oligosaccharides. The genes encoding these transferases are widely distributed in isolates from both GBS/MFS and enteritis patients. The gene (wlaN) encoding the 1,3-galactosyltransferase is essential for the terminal galactose to produce GM1- and higher di- and trisialo ganglioside-like structures, and contains a homopolymeric sequence that introduces phase-variable behavior in the expression of this enzyme.184 This means that the expression of the GM1 epitopes on LPS will vary according to a single base insertion/deletion, and most isolates are likely to contain organisms with and without expression of this enzyme. Which members of the population become dominant

depends on growth conditions. Without wlaN expression, the LPS expresses a GM2 ganglioside epitope, which has been demonstrated to be immunogenic.260 Thus, although the production of antiganglioside antibodies is dependent on the presence and expression of these genes, the genes are sufficiently common in isolates of Campylobacter that they do not typify neurogenic strains.

Other Infections Antiganglioside antibodies also occur in GBS cases associated with other infections. Haemophilus influenzae has been associated with anti-GQ1b/GT1a antibodies in MFS147 and anti-GM1 antibodies in GBS.213 The LPSs of H. influenzae and H. ducreyi are known to be sialylated,98,196,198 and the organisms have multiple sialyltransferase genes.112,131 However, unlike Campylobacter, Haemophilus does not synthesize sialic acid but uses host sialic acid to sialylate its LPS.327 Antiganglioside antibodies are also found in some patients with serologic evidence of recent CMV and M. pneumoniae infection.95,97,101,128,143,158,362 These two organisms do not have the genes for glycolipid synthesis but incorporate molecules from the host cell into their own membranes. Interestingly, CMV is particularly associated with antibodies reactive to GM2 ganglioside,140 which is a very common ganglioside of tissues outside the nervous system. The occurrence of ganglioside structures is very prominent in the most common organisms associated with GBS/MFS. Also, the presence of surface sialic acid is considered to enable the adsorption of factor H and the inhibition of complement activation.

Molecular Mimicry The production of antiganglioside antibodies in response to the expression of ganglioside oligosaccharide sequences by a pathogen is most frequently referred to as “molecular mimicry.” This is a convenient shorthand phrase in autoimmunity to describe the similarity in structural sequences between a pathogen and a relevant self-antigen that might generate self-reactive T or B cells. However, in biologic systems, mimicry traditionally implies that the mimic obtains a selective advantage by expressing some phenotypic feature of another unrelated species. In fact, molecular mimicry is a prominent and familiar feature of pathogens and is related to their success. In most of the studied instances, the molecular mimics are “functional” mimics, lacking both sequential and structural homology to the host proteins, the function of which is subverted.296 Often the pathogen mimicry appears to have arisen by a process of convergent evolution, but some mimicry involves pathogen expression of homologues of host

592

Neuroimmunology of the Peripheral Nervous System

proteins, the genes for which are likely to have been acquired by horizontal transfer. In many cases the mimicry involves proteins that disrupt the immune response and can target both the innate and adaptive immune systems. The presence of sialic acid at the surface of microorganisms has been suggested to suppress serologic responses by an unknown mechanism. The discovery of the family of sialic acid–binding proteins on immune cells may offer a partial explanation. Thus the NK cell inhibitory receptor siglec-7 has been shown to bind to disialogangliosides.124,354 This receptor is also found on monocytes, dendritic cells, and a CD8 T-cell subset.52 Thus the expression of disialogangliosides by C. jejuni and other organisms could suppress the innate immune response and therefore possibly the adaptive response, with obvious selective advantage to the organism. How monosialoganglioside might have a similar effect is still not known. It is not known for how long C. jejuni has been a human pathogen, but it is quite possible that any displayed mimicry is an adaptation to the avian immune system. Nonetheless, it appears that the mimicry is successful in humans because, in most cases of C. jejuni enteritis, high titers of antiganglioside antibodies are not found. Similarly, in the majority of cases, the repeated injection of gangliosides into human subjects does not lead to a marked humoral response.88,177,255 This suggests that the development of high-titer antibodies in GBS/MFS may be due to either the host failing to respond appropriately, or to some strains of C. jejuni carrying an additional factor that defeats the subversion of the immune response. There is little direct evidence that C. jejuni provokes a poor immune response, or that it is particularly resistant to the innate immune system. Avian peritoneal macrophages are able to phagocytose and effectively kill the organism,216 although chickens produce weak B-cell responses against thymus independent type-1 antigens.130 In humans as few as 500 organisms can be infective,331 yet most laboratory animals are resistant and no good laboratory model of the human enteritis exists. Mice clear the organism quickly, although there are some strain differences.242 Campylobacter jejuni is susceptible to human complement,75 as is C. coli, but C. fetus (which causes bacteremia more commonly) is more resistant.26 Human macrophages have been shown to be able to phagocytose and kill C. jejuni efficiently independent of strain, but, interestingly, macrophages from about 10% of donors, although able to phagocytose the organisms, did not kill them.336 Plasmid involvement in virulence has been largely discounted because plasmids have been found in less than 10% of isolates. However, at least one strain has been shown to carry a plasmid that contains possible virulence genes.18 Similarly the occurrence of strains producing an enterotoxin with properties similar to cholera toxin and binding to ganglioside has been suggested, but consistent evidence is lacking.335 Nonetheless, a putative enterotoxigenic strain

has been shown to suppress the murine humoral response to red blood cells.241 An increasing incidence of GBS with a high association with enteritis and C. jejuni in Curacao could be due to a recent introduction of a new strain or increased exposure.323 The immune response against ganglioside oligosaccharide may not be just humoral but may also involve T cells. The preponderance of IgG1/IgG3 antiganglioside antibodies93,350 suggests a close T-cell involvement with B cells, which may arise from the antiganglioside B cells acting to present a C. jejuni–derived protein antigen.93,344 In addition, the neuropathy associated with antecedent C. jejuni infection is not always AMAN. Acute inflammatory demyelinating polyradiculoneuropathy (AIDP) is relatively frequent105,258,322 and occurs in the presence of antiganglioside antibodies.284 T cells responsive to C. jejuni antigens derived from nerve biopsy from a GBS patient were found to be predominantly of the / receptor type,23 and a cell line that was developed was shown to have an unusual restricted  gene usage.48 The number of circulating / cells expressing V 1 is increased in GBS.30 A similar expansion has been noted following a number of infections, including human immunodeficiency virus (HIV).32 The / T cells are very important in bacterial and viral infections and in particular for downregulating the / T-cell response and the concomitant inflammation.99 The / T cells have more resemblance to cells of the innate system, carrying a number of NK cell–type receptors and having wide antigen specificities that are not MHC restricted. Although it has been suggested that they respond to lipid antigens, this has not been clearly demonstrated. However, lipid antigens, including glycosphingolipids, are presented by the invariant MHClike molecule CD1d.212,218 There are four CD1 molecules in humans, CD1a through CD1d. They typically present antigen to stimulate / T cells, including CD4 /CD8

cells as well as CD4 and CD8 cells. The CD1 molecules are expressed on myeloid cells, in particular dendritic or predendritic cells, and occur in the peripheral nerve in inflammatory neuropathy.141 CD1d is able to present -galactosylceramide but not -galactosylceramide in humans.138 A CD1b-dependent stimulation of CD8 / T cells by glycolipids has been shown in multiple sclerosis patients.280 However, CD1b is known to present bacterial glycolipids by binding to fatty acid components and is not specific for the carbohydrate component. There thus remains the possibility that oligosaccharide haptens may be directly involved in T-cell activation, but neither T-cell receptor specificity nor the ability of a CD1 molecule to present LPS has been clearly demonstrated. Both CD1 and the / T cell are clearly important factors in the innate response to infection and may be implicated in the variant response shown by GBS patients to certain infective agents.

Peripheral Nerve Antigens

PATHOPHYSIOLOGIC RELEVANCE OF AUTOANTIGENS IN ANIMAL MODELS Galactocerebroside-Induced EAN It is difficult to induce immune responses to GalC in rodents, but repeated immunization of rabbits produces high titers of complement-fixing antibody and a subacute neuropathy. The neuropathy is characterized by macrophage-associated demyelination in the absence of T-cell infiltration, a point of difference from the EAN induced by proteins (discussed below).265,299 The presumption is that the disease is produced by a direct action of the antibodies on the myelin with complement fixation, with the resulting opsonization leading to vesicular dissolution of myelin and macrophage invasion without the detected participation of T cells. Whether anti-GalC is expressed on the plasma membrane of Schwann cells and leads to their lysis is not clear. In support of this hypothesis, anti-GalC serum induces demyelination when applied to myelinated tissue cultures,266 and following intraneural injection into the rat sciatic nerve.116,264,300 When antibody to GalC is injected into rats with adoptive transfer EAN induced by P2, the severity of the disease and amount of demyelination are increased.96 Galactocerebroside is abundant in CNS as well as peripheral nerve myelin. Therefore, it is not surprising that rabbits immunized with GalC and carrying high titers of circulating antibody experienced an inflammatory response–induced demyelination of the optic nerve.85 P0 EAN Active immunization of Lewis rats with bovine P0 produces an EAN whose clinical and light microscopic features resemble those induced by whole myelin, although detailed morphologic comparisons have not been made.208 With immunization using either of two synthetic peptides of P0 protein, the resulting EAN was more severe and prolonged with the cytoplasmic domain peptide 180–199 than with the extracellular domain peptide 56–71.368 Adoptive transfer of CD4 T-cell lines directed against the same peptides also induced EAN.183 Peptide 180–199 was immunodominant and was strongly recognized by T cells following immunization with whole myelin, whereas peptide 56–71 was not recognized. Antibodies to P0 appear early in the course of EAN induced by myelin and then subside. By contrast, antibodies to P2 appear later and persist, which argues for P0 being more important than P2 in producing myelin-induced disease.11,12 The relative importance of P0 was also demonstrated by showing that tolerance induced by nasal administration of P0 peptides 56–71 and 180–199 significantly reduced the severity of EAN induced with whole bovine myelin.369 Antibodies to P0 induce demyelination upon intraneural injection into rat sciatic nerves.116,356 However, it may be speculated that

593

antibodies directed against any extracellularly expressed antigen would induce demyelination if they were brought into direct contact with the external surface of the myelin sheath. Similarity between viral and myelin proteins might cause a viral infection to initiate a cross-reactive immune response, which would explain the pathogenesis of some cases of GBS. Attempting to further this hypothesis, Adelmann and Linington searched a protein sequence database for homologies between P0 and microbial proteins.1 They identified identical pentapeptides in P0 and proteins from Epstein-Barr virus, CMV, varicella, and HIV, all viruses that are recognized as precipitating GBS. Synthetic peptides from some of these proteins did not stimulate neuritogenic T-cell lines directed against P0 and did not induce EAN after injection into Lewis rats. PMP22 EAN There is a single report of the induction of EAN with recombinant rat PMP22 in Lewis rats.81 The disease induced was mild, but the histologic changes included perivascular mononuclear cell infiltrates and demyelinated nerve fibers, similar to the findings in AIDP. There is also a report of the detection of antibodies to PMP22 during the course of EAN induced in rats by myelin.144 Difficulties in purifying or expressing sufficient quantities of PMP22 have inhibited further investigation of its role in EAN.278 P2 EAN P2 was the first purified myelin protein to be identified as a neuritogenic antigen, and active immunization with it readily induces EAN in Lewis rats.114,133,134 Residues 61 through 70 form the longest amphipathic -helical domain in the P2 molecule, and represent the minimum sequence of peptide residues necessary to induce disease.238,239 This is consistent with the observation that such domains are immunodominant T-cell epitopes. Actively induced disease can be replicated in Lewis rats by adoptive transfer of CD4 T cells from rats immunized with P2 protein in a dose-dependent manner.182 Small numbers of cells induce mild disease in which the predominant change is demyelination, whereas larger numbers induce axonal degeneration and cause more severe disease.125,340 As in actively induced disease and also in GBS, the spinal cord is notably spared, although there is prominent meningeal infiltration surrounding the lumbosacral cord in EAN. The active and passive transfer models of P2-induced EAN in the Lewis rat have been enormously helpful in the dissection of the mechanisms of EAN and in studying potential therapies for autoimmune disease. For instance, intravenous administration of large amounts of recombinant human P2 ameliorated EAN, which had been induced either actively or passively by a synthetic peptide resembling bovine P2 residues 53 through 78.338 The likely explanation of this therapeutic effect is that antigen therapy

594

Neuroimmunology of the Peripheral Nervous System

induced T-cell apoptosis in the inflammatory lesions. More details of the immune responses in EAN and approaches to treatment are given in Chapter 27.

MAG-Induced Neuropathy Models The direct induction of EAN with MAG has not been reported. However, adoptive transfer of four T-cell lines directed against different peptide sequences predicted to be T-cell epitopes from the rat MAG extracellular domain induced mild inflammation of the central nervous system in Lewis rats.337 Two of the lines, one directed against residues 20 through 40 and another directed against residues 354 through 377, also induced mild inflammation but not demyelination in the dorsal roots and sciatic nerves. The presence of inflammation in both the PNS and CNS is not surprising because MAG is present in both. The experiment further illustrates that immunization of appropriate animals with many CNS or PNS myelin antigens will induce inflammation in neural tissue in which that epitope is expressed. Injection of human serum containing anti-MAG antibodies has induced demyelination in animal nerve using species that contain MAG expressing the HNK-1 carbohydrate epitope.100,349 Repeated injection of purified human anti-MAG IgM induced demyelination and widely spaced myelin in chick nerves.308 These experimental observations, and the demonstration of complement components and IgM on myelin sheaths in a location similar to the widely spaced myelin, make it very likely that the antibody is causative.

Ganglioside-Induced Neuropathy Models GM1 Model Despite the failure of several previous attempts, Yuki et al.365 recently succeeded in inducing an anti-GM1– mediated experimental neuropathy in rabbits by repeated immunization with a ganglioside mixture. All of the injected animals developed anti-GM1 antibodies, which switched class from IgM to IgG, and all suffered an acute monophasic neuropathy. The histologic appearances were those of an axonal neuropathy with wallerian degeneration but no lymphocytic infiltration or primary demyelination. Immunoglobulin deposition was demonstrated on peripheral nerve axons. A similar disease was also induced by immunization with purified GM1.365 This model is similar to the acute motor axonal neuropathy form of GBS. GD1b Model In a seminal study, another interesting model comprising an ataxic neuropathy was induced in rabbits with purified ganglioside GD1b.167 The affected rabbits had a dorsal root ganglionopathy with loss of DRG cells, degeneration of sensory axons and the dorsal column, but no inflammation

or involvement of the motor axons. Immunized rabbits all developed anti-GD1b antibodies, but only half developed the neuropathy. However, the demonstration that a monoclonal antibody directed against GD1b labeled about 50% of rabbit DRG neurons added strength to the argument that the antibody response was responsible for the neuropathy.

PATHOPHYSIOLOGIC RELEVANCE OF AUTOANTIGENS IN HUMAN NEUROPATHIES This chapter has been restricted to a description of antigens that have been shown to be or suspected as relevant to the pathogenesis of human neuropathy. Many of the specific clinical and immunologic phenotypes associated with these antigens are principally covered in Part N, Chapters 99 through 108, and the present section only briefly summarizes these findings in relation to the preceding text.

Neuronopathies Paraneoplastic neuropathies are often associated with antineuronal antibodies, and the most characteristic syndrome is a subacute sensory neuronopathy associated with SCLC and antibodies against Hu, a neuronal RNA-binding protein that is also present in SCLC (see above). The clinical picture is usually that of an asymmetrical proximal painful sensory syndrome often associated with other features, especially limbic encephalopathy, cerebellar, brainstem, and autonomic dysfunction; and weakness.38,56 Electrophysiologic testing also shows subclinical motor nerve conduction abnormalities.38 The finding of anti-Hu antibodies is highly specific for this condition and is usually associated with a SCLC, although a wide variety of other neoplasms have been reported.41,56 Antibodies against voltage-regulated Ca2 channels have been detected as co-occurring in patients with anti-Hu antibodies.176 A small number of patients with an acute sensory neuronopathy and antibodies to ganglioside GD1b and other gangliosides have been reported,240,343 consistent with the model described by Kusunoki et al. above.167 However, the association of this antibody with acute sensory neuropathy is not absolute. Another Japanese study identified 9 patients of 445 with GBS who had antibodies to ganglioside GD1b but no other glycolipids. Their clinical features were those of AIDP, but all had sensory changes.210

Axonopathies Paraneoplastic neuropathies associated with the antineuronal antibodies described in the previous paragraph, and others detailed below, may also present as distal sensory or

Peripheral Nerve Antigens

sensory and motor neuropathies caused by axonopathy. Antibodies to CV2 or CRMP are associated with a subacute sensorimotor neuropathy and frequently also with cerebellar symptoms.7,359 Endoneurial inflammation and demyelination have been found in some patients.359 However, antibodies to this group of antigens are not always associated with peripheral neuropathy. Of 116 patients with anti–CRMP-5 antibodies, only 47% had signs of peripheral neuropathy.359 As might be expected, the success or otherwise of treatment is dependent on the stage and severity of the disease. Removal of the tumor and immunosuppressive treatment while the neurologic impairment was not severe led to good recovery,79 but in patients with severe neurologic impairment, immunosuppression was not successful.22,25,56 However, if T cells are the agents of neuronal loss, possibly specific T-cell directed therapy might be more successful.3 Sensory neuropathy or sensory and motor neuropathy is also associated with other neoplasms and other antineuronal antibodies. For instance, the cerebellar degeneration associated with anti-Yo antibodies and ovarian carcinoma and other tumors may be associated with peripheral neuropathy.248 Another example of an immunologically mediated axonopathy is the AMAN associated with antibodies to gangliosides GM1, GD1a, and other gangliosides to be discussed in the next section. There have been several reports of antibodies to sulfatide being associated with sensory neuropathy.220,247 Antibodies to sulfatide stained a population of DRG neurons.254 In one report, IgM antibodies to sulfatide in the absence of a paraprotein were associated with a sensory axonopathy, and the antibodies stained axons.189 In the same report, antibodies to sulfatide in the presence of an IgM paraprotein were associated with a motor and sensory demyelinating neuropathy, and the sera stained myelin. However, in a conflicting report, of five patients with IgM paraproteins reactive to sulfatide alone, three had a sensory axonal neuropathy and two a predominantly motor demyelinating neuropathy.71 It is apparent that further research is needed to clarify the role and antigen specificity of antibodies to sulfatide in causing peripheral neuropathy.

Inflammatory Demyelinating Polyradiculoneuropathies P0 Antibodies Antibodies to P0 were found in 16% to 29% of patients with chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) in two studies139,356 but not in a third.168 Sera containing antibodies to P0 have been shown to produce demyelination following injection into rat sciatic nerve.116,356 Antibodies to a 35-kDa isoform of P0 have been found in CIDP but are not specific, being also found in motor neuron disease and other conditions.123,207,225 The

595

nature of this isoform of P0 is unclear, but it does not appear to represent a variation in glycosylation.206 Most effort has been directed toward detecting antibody responses to P0 in human neuropathies, despite the clear evidence that in EAN the T-cell responses are fundamentally important. However, Khalili-Shirazi et al. identified T-cell responsiveness to purified human P0 protein, judged by tritiated thymidine incorporation of blood mononuclear cells in the presence of antigen, in 6 of 19 patients, and to synthetic rat P0 peptides in 3 others.142 Responses in controls were rare. In the only other paper to address this issue,54 patients with GBS developed interleukin-4 (IL-4) responses to synthetic human peptides corresponding to those previously found to induce EAN.183 However, the responses were delayed, and IL-4 responses are characteristic of helper T cell type 2 responses, which are more associated with recovery from than initiation of autoimmune disease. PMP22 Antibodies There are no reports about T-cell responses to PMP22 in humans and conflicting reports concerning the presence of antibodies. Gabriel et al. identified IgG or IgM antibodies by enzyme-linked immunosorbent assay (ELISA) to synthetic peptides representing either the first or second extracellular domains in 58% of 19 GBS patients, 41% of 17 CIDP patients, 10% of 30 other neuropathy controls, and 4% of 51 normal subjects.80 The presence of antibodies was confirmed by identifying an appropriate 22-kDa band on a Western blot of human cauda equina in most of the GBS and CIDP patients but not usually in the others. However, two other groups have failed to identify such a clear relationship between antibodies to PMP22 and inflammatory demyelinating polyradiculoneuropathy. One study identified antibodies to recombinant PMP22 produced in E. coli in 23% of normal subjects and in 23% to 75% of patients with various sorts of neuropathy.262 However, these authors could not identify staining of PMP22-transfected cells with any of the positive sera. The other study did not identify antibody to PMP22 expressed in Chinese hamster ovary cells in any of 24 patients with CIDP and 25 patients with GBS.168 Therefore, the relevance of immunity to PMP22 in this disorder remains unclear, and resolution of these discrepant results will require further research. P2 Antibodies Because P2 was the first protein to be shown capable of inducing EAN, there have been several attempts to find antibodies to P2 in GBS and CIDP patients, and most have been negative,115 apart from one report in which a minority of patients had very high titers of IgM antibodies to bovine P2 protein.141 T-cell responses to P2 were sought in two studies. Khalili-Shirazi et al. used the mononuclear cell transformation technique with bovine P2 protein and

596

Neuroimmunology of the Peripheral Nervous System

identified responses in 6 of 19 GBS and 5 of 13 CIDP patients but also in occasional control subjects.142 These experiments would be worth repeating with human P2 protein. Dahle et al., using the ELISPOT technique, found that synthetic peptides of human P2 induced exaggerated production of interferon- by lymphocytes from one of seven patients with GBS.56

Ganglioside Antibody–Related Diseases Antibodies to gangliosides are found in both acute and chronic neuropathy syndromes. In the former, they tend to be of the IgG class, arise transiently after preceding infections, and disappear concomitant with clinical recovery. As described above, molecular mimicry is believed to be one major mechanism by which they arise. In chronic neuropathy syndromes, the antibodies tend to be of the IgM class and are persistently present over many years, often as IgM paraproteins, and in this situation the relation to molecular mimicry is less clear. Anti-GM1 IgM antibodies and their related clinical syndromes have been thoroughly studied since first being identified in MMN with demyelinating conduction block.149 IgM antibodies to GM1 are found in about 50% of MMN cases, but the figures vary depending upon the assay methodology used. Patients with MMN have a stereotyped clinical picture comprising a chronic asymmetrical motor syndrome, often with distal onset in an upper limb. Atypical cases may occur without focal conduction block in which the phenotype resembles chronic spinal muscular atrophy. Despite evidence of minor sensory involvement in some reported cases, anti-GM1 antibodies are clearly a marker for predominantly motor nerve syndromes. Antibodies to a wide range of glycolipids, including GM 1 , GM 1 (NeuGc), GM 1b , GalNAc-GM 1b , GD 1a , GalNAc-GD1a, GD1b, 9-O-acetyl-GD1b, GD3, GT1a, GT1b, GQ1b, GQ1ba, LM1, GalC, and SGPG, have now been reported either as case reports or larger clinical series in patients with GBS.237,253,351,360 The strongest associations are found in patients with AMAN and MFS. The prevalence of anti-GM1 antibodies in AMAN and AIDP varies from study to study, ranging from 0 to 80%, reflecting diverse variables including the prevalence of C. jejuni infection in the study population and the methodology used for antibody measurement.40,360 In the United States and Europe, approximately 20% of GBS cases have such antibodies. Anti-GD1a antibodies are also found in AMAN to a much greater extent than in AIDP. In a series of 138 AMAN and AIDP GBS patients from China, 60% of AMAN but only 4% of AIDP cases had anti-GD1a antibodies.106 Antibodies to GM1b and GalNAc-GD1a are also present in AMAN cases.135,157,361 In AIDP, antibodies to gangliosides and other glycolipids have been less consistently found, although they are certainly present in a proportion of cases. GBS occurring

in association with CMV infection has been linked with anti-GM2 antibodies.328 Mycoplasma pneumoniae infection preceding GBS has been reported in association with anti-GalC antibodies as described below. Anti-LM1 and anti-SGPG antibodies have also been reported in AIDP.253,353,360 MFS is very strongly associated with anti-GQ1b ganglioside antibodies.159 These antibodies are a very sensitive and specific marker for syndromes characterized by ophthalmoplegia.237,345 In MFS, anti-GQ1b IgG antibodies are elevated in over 90% of cases during the acute phase but may disappear rapidly, often being absent during convalescence. Diagnostic testing should therefore be conducted on serum samples drawn early in the course of the disease. Anti-GQ1b antibodies invariably cross-react with GT1a and in 50% of cases with GD3 and/or GD1b.

Galactocerebroside Antibodies to GalC have been found in GBS associated with M. pneumoniae infections,158,166 and the antibody activity could be absorbed by M. pneumoniae. Additionally, antibodies to GalC have been found in M. pneumoniae– associated encephalitis,223 providing strong evidence of molecular mimicry between glycolipid epitopes in M. pneumoniae and GalC.

Charcot-Marie-Tooth Disease Ritz et al. reported antibodies detected by ELISA against the whole recombinant human PMP22 molecule in 70% of patients with CMT1 and 60% with CMT2 as well as 23% of normal subjects.262 Gabriel et al. also found antibodies to PMP22 extracellular domain peptides by ELISA in 28% of 53 patients with CMT1a compared with only 10% of 30 patients with other neuropathies and 4% of 51 normal subjects. Of the patients with CMT1A, 12 had a stepwise progressive course, and of these four had antibodies.78 It was suggested that immune responses to myelin proteins, especially one that is overexpressed or mutated, might contribute to producing nerve damage in a subset of patients with hereditary neuropathy.

Paraproteinemic Neuropathies Myelin-Associated Glycoprotein About half of patients with paraproteinemia and demyelinating neuropathy have an IgM paraprotein that has antibody activity directed against MAG. The associated neuropathy is variable but is most commonly an insidiously progressive predominantly sensory neuropathy associated with a tremor in about a third of cases.224 The fine specificity of the antibody activity varies from patient to patient, with the paraproteins from many patients reacting with the HNK-1 epitope that is also present on P0, although the

597

Peripheral Nerve Antigens

fine structure of the oligosaccharides on the two proteins is different. The epitope is present only on tri- and quatroantennary oligosaccharides on P0, but on all types of oligosaccharide on MAG.36 As a consequence of the differences in fine specificity, the pattern of binding of sera with anti-MAG antibodies to peripheral nerve is variable. Most sera bind noncompact myelin in the periaxonal and abaxonal Schwann cell membranes and SchmidtLanterman incisures where MAG is located, but a few sera bind axons.189 In biopsied nerves from patients, bound IgM is found throughout the myelin sheath. Other Sulfated Antigens Most sera that react with MAG also react with SGPG, a myelin glycolipid that shares an epitope with MAG.71 Of patients with antibodies reactive with SGPG and not MAG or sulfatide, one had a slowly progressive sensory and motor demyelinating neuropathy and one a predominantly motor demyelinating neuropathy.71 There are a small number of reports of antibody activity against chondroitin sulfate in the serum of patients with IgM paraproteinemia and sensory axonal neuropathy.220,286 Chondroitin sulfate is a glycosaminoglycan present in axonal membranes and not myelin. In about half these cases the serum also shows antibody activity against sulfatide, and so may be directed against a shared galactosesulfatide epitope. GD1b Ganglioside An uncommon sensory ataxic neuropathy syndrome first described in 1985 is a variant of IgM paraproteinemic neuropathy in which the paraprotein reacts with NeuNAc(2-8)NeuNAc(2-3)-configured disialylated gangliosides, including GD1b, GD3, GD2, GT1a, GT1b, and GQ1b. In some publications this syndrome has been referred to as CANOMAD (chronic ataxic neuropathy with ophthalmoplegia, M protein, agglutination, and disialosyl antibodies).236,348 The clinical pattern comprises profound loss of limb kinesthesia (proprioception and vibration sense) with relative preservation of limb muscle strength. The patients therefore present with gait and upper limb ataxia. In addition, they often have craniobulbar motor involvement, either as a fixed set of symptoms or as relapsing-remitting symptoms; these comprise diplopia, ptosis, dysphagia, and dysarthria. The clinical course is chronic, often extending over more than 10 years, and is interspersed with episodes of relapse particularly affecting the craniobulbar and respiratory motor system. Patients are generally within the typical age group for paraproteinemic neuropathy (60 years) but may be younger. Electrophysiologically, motor studies show normal or moderately reduced conduction velocities and normal or prolonged distal motor latencies and F-wave latencies. Denervation changes are variable. Sensory nerve action potentials are absent or markedly reduced, and H reflexes are absent. Where performed, sural nerve biopsies

have shown varying degrees of demyelination and large axonal loss with relative preservation of small and unmyelinated fiber bundles. A partial response to intravenous immunoglobulin and other treatments has been reported in some cases.

Protein Antigens There are also a number of isolated reports of antibody activity directed against a variety of different proteins in the sera of patients with paraproteins. For instance, one patient with an axonal neuropathy had an IgM paraprotein and antibody activity against the 200-kDa neurofilament protein.33 However, this antibody also reacted with ribosomal proteins, and low-titer antibodies to this neurofilament protein are not uncommon in normal sera.297 Antibodies against the 170-kDa Schwann cell antigen L-periaxin have been detected in 3 of 17 patients with neuropathy and an IgG paraproteinemia.174 These antibodies were, however, IgM and polyclonal and could reflect a secondary response. Similar antibodies (but possibly monoclonal) were also found in some patients with type 2 diabetes and an associated neuropathy. No similar antibodies were found in nine patients with an IgM paraprotein and neuropathy. Intraneural injection of periaxinpositive sera with added complement produced sensory nerve conduction changes and structural indications of demyelination and axonal changes.

Channelopathies Isaacs’ Syndrome Isaacs’ syndrome consists of muscle cramps, slow relaxation of muscles following contraction, and hyperhidrosis. The electrophysiologic characteristics are spontaneous discharges of motor units as doublets, triplets, or multiplets at 5 to 150 Hz (myokymia) or faster (neuromyotonia). The serum of many patients contains antibodies directed against dendrotoxin-sensitive fast voltage-gated potassium channels, which are present at paranodes and internodes. The antibodies probably gain access to the motor nerve fibers in the anterior roots or at the motor nerve terminals, where the blood-nerve barrier is relatively deficient, and thus trigger the spontaneous discharges.14,287 Lambert-Eaton Syndrome Lambert-Eaton syndrome is a well-recognized syndrome often associated with a SCLC. The main clinical features are weakness, fatigue, and autonomic symptoms. It is due to antibodies directed against the P/Q type of voltagegated calcium channels that lead to the release of acetylcholine vesicles from the motor end plate following arrival of an action potential.176 It responds well to treatments

598

Neuroimmunology of the Peripheral Nervous System

that lower the antibody titer, including plasma exchange, intravenous immunoglobulin, and immunosuppression.

16.

REFERENCES 1. Adelmann, M., and Linington, C.: Molecular mimicry and the autoimmune response to the peripheral nerve myelin P0 glycoprotein. Neurochem. Res. 17:887, 1992. 2. Adlkofer, K., Martini, R., Aguzzi, A., et al.: Hypermyelination and demyelinating peripheral neuropathy in Pmp22deficient mice. Nat. Genet. 11:274, 1995. 3. Albert, M. L., Austin, L. M., and Darnell, R. B.: Detection and treatment of activated T cells in the cerebrospinal fluid of patients with paraneoplastic cerebellar degeneration. Ann. Neurol. 47:9, 2000. 4. Alderuccio, F., Rolland, J. M., Toner, G. C., et al.: Autoantibodies to neurons and to the cytoskeleton in small cell carcinoma with paraneoplastic sensory neuropathy. Autoimmunity 5:115, 1989. 5. Ando, S., Hirabayashi, Y., Kon, K., et al.: A trisialoganglioside containing a sialyl 2-6 N-acetylgalactosamine residue is a cholinergic-specific antigen, Chol-1a. J. Biochem. 111:287, 1992. 6. Ang, C. W., de Klerk, M. A., Endtz, H. P., et al.: GuillainBarré syndrome- and Miller Fisher syndrome-associated Campylobacter jejuni lipopolysaccharides induce anti-GM1 and anti-GQ1b antibodies in rabbits. Infect. Immun. 69:2462, 2001. 7. Antoine, J. C., Honnorat, J., Camdessanche, J. P., et al.: Paraneoplastic anti-CV2 antibodies react with peripheral nerve and are associated with a mixed axonal and demyelinating peripheral neuropathy. Ann. Neurol. 49:214, 2001. 8. Antoine, J. C., Honnorat, J., Vocanson, C., et al.: Posterior uveitis, paraneoplastic encephalomyelitis and auto-antibodies reacting with developmental protein of brain and retina. J. Neurol. Sci. 117:215, 1993. 9. Apostolski, S., Sadiq, S. A., Hays, A., et al.: Identification of Gal(1-3)GalNAc bearing glycoproteins at the nodes of Ranvier in peripheral nerve. J. Neurosci. Res. 38:134, 1994. 10. Arasaki, K., Kusonoki, S., Kudo, N., and Kanazawa, I.: Acute conduction block in vitro following exposure to antiganglioside sera. Muscle Nerve 16:587, 1993. 11. Archelos, J., Toyka, K., and Hartung, H.: B cell responses to the PNS protein P0 in experimental autoimmune neuritis. J. Neurol. Sci. 128:111, 1995. 12. Archelos, J. J., Roggenbuck, K., Schneider-Schaulies, J., et al.: Detection and quantification of antibodies to the extracellular domain of P0 during experimental allergic neuritis. J. Neurol. Sci. 117:197, 1993. 13. Ariga, T., Kohriyama, T., and Freddo, L.: Characterization of sulfated glucuronic acid containing glycolipids reacting with IgM M-proteins in patients with neuropathy. J. Biol. Chem. 262:848, 1987. 14. Arimura, K., Sonoda, Y., Watanabe, O., et al.: Isaacs’ syndrome as a potassium channelopathy of the nerve. Muscle Nerve 25(Suppl. 11):S55, 2002. 15. Aspinall, G. O., Fujimoto, S., McDonald, A. G., et al.: Lipopolysaccharides from Campylobacter jejuni associated

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

with Guillain Barre syndrome patients mimic human gangliosides in structure. Infect. Immun. 62:2122, 1994. Aspinall, G. O., McDonald, A. G., Pang, H., et al.: Lipopolysaccharides of Campylobacter jejuni serotype O:19: structure of core oligosaccharide regions from the serostrain and two bacterial isolates from patients with the Guillain-Barré syndrome. Biochemistry 33:241, 1994. Aspinall, G. O., McDonald, A. G., Raju, T. S., et al.: Chemical structures of the core regions of Campylobacter jejuni serotypes O:1, O:4, O:23, and O:36 lipopolysaccharides. Eur. J. Biochem. 213:1017, 1993. Bacon, D. J., Alm, R. A., Burr, D. H., et al.: Involvement of a plasmid in virulence of Campylobacter jejuni 81-176. Infect. Immun. 68:4384, 2000. Bahmanyar, S., Moreau-Dubois, M.-C., Brown, P., et al.: Serum antibodies to neurofilament antigens in patients with neurological and other diseases and in healthy controls. J. Neuroimmunol. 5:191, 1983. Barclay, A. N., Wright, G. J., Brooke, G., and Brown, M. H.: CD200 and membrane protein interactions in the control of myeloid cells. Trends Immunol. 23:285, 2002. Bataller, L., Graus, F., Saiz, A., et al.: Clinical outcome in adult onset idiopathic or paraneoplastic opsoclonus-myoclonus. Brain 124:437, 2001. Batchelor, T. T., Platten, M., and Hochberg, F. H.: Immunoadsorption therapy for paraneoplastic syndromes. J. Neurooncol. 40:131, 1998. Ben-Smith, A., Gaston, J. S., Barber, P. C., and Winer, J. B.: Isolation and characterisation of T lymphocytes from sural nerve biopsies in patients with Guillain-Barré syndrome and chronic inflammatory demyelinating polyneuropathy. J. Neurol. Neurosurg. Psychiatry 61:362, 1996. Benyahia, B., Liblau, R., Merle-Beral, H., et al.: Cell-mediated autoimmunity in paraneoplastic neurological syndromes with anti-Hu antibodies. Ann. Neurol. 45:162, 1999. Blaes, F. Immunotherapeutic approaches to paraneoplastic neurological disorders. Expert Opin. Invest. Drugs 9:727, 2000. Blaser, M. J., Smith, P. F., and Kohler, P. F.: Susceptibility of Campylobacter isolates to the bactericidal activity of human serum. J. Infect. Dis. 151:227, 1985. Boerkoel, C. F., Takashima, H., Stankiewicz, P., et al.: Periaxin mutations cause recessive Dejerine-Sottas neuropathy. Am. J. Hum. Genet. 68:325, 2001. Bollensen, E., Steck, A. J., and Schachner, M.: Reactivity with the peripheral myelin glycoprotein P(0) in serum from patients with monoclonal IgM gammopathy and polyneuropathy. Neurology 38:1266, 1988. Bond, J. P., Saavedra, R. A., and Kirschner, D. A.: Expression and purification of the extracellular domain of human myelin protein zero. Protein Expr. Purif. 23:398, 2001. Borsellino, G., Koul, O., Placido, R., et al.: Evidence for a role of gammadelta T cells in demyelinating diseases as determined by activation states and responses to lipid antigens. J. Neuroimmunol. 107:124, 2000. Bosse, F., Zoidl, G., Wilms, S., et al.: Differential expression of two mRNA species indicates a dual function of peripheral myelin protein, PMP-22, in cell growth and myelination. J. Neurosci. Res. 37:529, 1994.

Peripheral Nerve Antigens 32. Boullier, S., Cochet, M., Poccia, F., and Gougeon, M. L.: CDR3-independent gamma delta V delta 1 T cell expansion in the peripheral blood of HIV-infected persons. J. Immunol. 154:1418, 1995. 33. Brindel, I., Preud’homme, J. L., Diaz, J. J., et al.: A human monoclonal IgM lambda specific for an epitope shared by the 200 kDa neurofilament protein, histones and ribosomal proteins. J. Autoimmun. 8:915, 1995. 34. Buchwald, B., Weishaupt, A., Toyka, K. V., and Dudel, J.: Pre- and postsynaptic blockade of neuromuscular transmission by Miller-Fisher syndrome IgG at mouse motor nerve terminals. Eur. J. Neurosci. 10:281, 1998. 35. Bullens, R. W. M., O’Hanlon, G. M., Wagner, E. R., et al.: Complex gangliosides at the neuromuscular junction are essential receptors for autoantibodies and botulinum neurotoxin but redundant for normal synaptic function. J. Neurosci. 22:6876, 2002. 36. Burger, D., Simon, M., Peruisseau, G., and Steck, A. J.: The epitope(s) recognised by HNK-1 antibody and IgM paraprotein in neuropathy is present on several N-linked oligosaccharide structures on human P0 and myelin-associated glycoprotein. J. Neurochem. 54:1569, 1990. 37. Calderon, R. O., Attema, B., and DeVries, G. H.: Lipid composition of neuronal cell bodies and neurites from cultured dorsal root ganglia. J. Neurochem. 64:424, 1995. 38. Camdessanche, J. P., Antoine, J. C., Honnorat, J., et al.: Paraneoplastic peripheral neuropathy associated with antiHu antibodies: a clinical and electrophysiological study of 20 patients. Brain 125:166, 2002. 39. Campagnoni, A. T., and Macklin, W. B.: Cellular and molecular aspects of myelin protein gene expression. Mol. Neurobiol. 2:41, 1988. 40. Carpo, M., Pedotti, R., Allaria, S., et al.: Clinical presentation and outcome of Guillain-Barre and related syndromes in relation to anti-ganglioside antibodies. J. Neurol. Sci. 168:78, 1999. 41. Chalk, C. H., Lennon, V. A., Stevens, J. C., and Windebank, A. J.: Seronegativity for type 1 antineuronal nuclear antibodies (“anti-Hu”) in subacute sensory neuronopathy patients without cancer. Neurology 43:2209, 1993. 42. Chassande, B., Leger, J. M., Younes-Chennoufi, A. B., et al.: Peripheral neuropathy associated with IgM monoclonal gammopathy: correlations between M-protein antibody activity and clinical/electrophysiological features in 40 cases. Muscle Nerve 21:55, 1998. 43. Chester, M. A., for the IUPAC-IUB Joint Commission on Biochemical Nomenclature (JCBN): Nomenclature of glycolipids—recommendations 1997. Eur. J. Biochem. 257:293, 1998. 44. Chiba, A., Kusunoki, S., Obata, H., et al.: Serum anti-GQ1b IgG antibody is associated with ophthalmoplegia in Miller Fisher syndrome and Guillain-Barre syndrome: clinical and immunohistochemical studies. Neurology 43:1911, 1993. 45. Chiba, A., Kusunoki, S., Obata, H., et al.: Ganglioside composition of the human cranial nerves, with special reference to pathophysiology of Miller Fisher syndrome. Brain Res. 745:32, 1997. 46. Chou, D. K. H., Ilyas, A. A., and Evans, J. E.: Structure of sulfated glucuronyl glycolipids in the nervous system react-

47.

48.

49.

50.

51.

52. 53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

599

ing with HNK-1 antibody and some IgM paraproteins in neuropathy. J. Biol. Chem. 261:11717, 1986. Coetzee, T., Fujita, N., Dupree, J., et al.: Myelination in the absence of galactocerebroside and sulfatide: normal structure with abnormal function and regional instability. Cell 86:209, 1996. Cooper, J. C., Ben-Smith, A., Savage, C. O. S., and Winer, J. B.: Unusual T cell receptor phenotype V gene usage of / T cells in a line derived from the peripheral nerve of a patient with Guillain-Barré syndrome. J. Neurol. Neurosurg. Psychiatry 69:522, 2000. Corbo, M., Quattrini, A., Latov, N., and Hays, A. P.: Localization of GM1 and Gal(1-3)GalNAc antigenic determinants in peripheral nerve. Neurology 43:809, 1993. Corbo, M., Quattrini, A., Lugaresi, A., et al.: Patterns of reactivity of human anti-GM1 antibodies with spinal-cord and motor neurons. Ann. Neurol. 32:487, 1992. Corriveau, R. A., Huh, G. S., and Shatz, C. J.: Regulation of class I MHC gene expression in the developing and mature CNS by neural activity. Neuron 21:505, 1998. Crocker, P. R., and Varki, A.: Siglecs, sialic acids and innate immunity. Trends Immunol. 22:337, 2001. Croft, P. H., Henson, R. A., Urich, H., and Wilkinson, P. C.: Sensory neuropathy with bronchial carcinoma: a study of four cases showing serological abnormalities. Brain 87:501, 1964. Dahle, C., Ekerfelt, C., Vrethem, M., et al.: T helper type 2 like cytokine responses to peptides from P0 and P2 myelin proteins during the recovery phase of Guillain-Barré syndrome [abstract]. J. Neurol. Sci. 153:54, 1997. Dalmau, J., Graus, F., Cheung, N.-K. V., et al.: Major histocompatibility proteins, anti-Hu antibodies, and paraneoplastic encephalomyelitis in neuroblastoma and small cell lung cancer. Cancer 75:99, 1995. Dalmau, J., Graus, F., Rosenblum, M. K., and Posner, J. B.: Anti-Hu–associated paraneoplastic encephalomyelitis/sensory neuronopathy: a clinical study of 71 patients. Medicine (Baltimore) 71:59, 1992. David, C., Solimena, M., and De Camilli, P.: Autoimmunity in stiff-man syndrome with breast cancer is targeted to the C-terminal region of human amphiphysin, a protein similar to the yeast proteins, Rvs167 and Rvs161. FEBS Lett. 351:73, 1994. De Angelis, M. V., Di Muzio, A., Lupo, S., et al.: Anti-GD1a antibodies from an acute motor axonal neuropathy patient selectively bind to motor nerve fiber nodes of Ranvier. J. Neuroimmunol. 121:79, 2001. Dehaut, F., Haddad, K., Alhayek, G., and PouplardBarthelaix, A.: Autoantibodies against H- and M-subunits of neurofilaments are induced by PC12 cell grafts or lesions into different sites of rat brain. Neurosci. Lett. 165:59, 1994. Demeure, C. E., Tanaka, H., Mateo, V., et al.: CD47 engagement inhibits cytokine production and maturation of human dendritic cells. J. Immunol. 164:2193, 2000. Derrington, E. A., and Borroni, E.: The developmental expression of the cholinergic-specific antigen Chol-1 in the central and peripheral nervous system of the rat. Dev. Brain Res. 52:131, 1990. Derrington, E. A., Kelic, S., and Whittaker, V. P.: A novel cholinergic-specific antigen (Chol-2) in mammalian brain. Brain Res. 620:16, 1993.

600

Neuroimmunology of the Peripheral Nervous System

63. Dick, D. J., Harris, J. B., Falkous, G., et al.: Neuronal antinuclear antibody in paraneoplastic sensory neuronopathy. J. Neurol. Sci. 85:1, 1988. 64. Dodd, J., and Jessell, T. M.: Lactoseries carbohydrates specify subsets of dorsal root ganglion neurons projecting to the superficial dorsal horn of rat spinal cord. J. Neurosci. 5:3278, 1985. 65. Dodd, J., Solter, D., and Jessell, T. M.: Monoclonal antibodies against carbohydrate differentiation antigens identify subsets of primary sensory neurones. Nature 311:469, 1984. 66. Dorrell, N., Mangan, J. A., Laing, K. G., et al.: Whole genome comparison of Campylobacter jejuni human isolates using a low-cost microarray reveals extensive genetic diversity. Genome Res. 11:1706, 2001. 67. Dropcho, E. J.: Antiamphiphysin antibodies with small-cell lung carcinoma and paraneoplastic encephalomyelitis. Ann. Neurol. 39:659, 1996. 68. D’Urso, D., Ehrhardt, P., and Muller, H. W.: Peripheral myelin protein 22 and protein zero: a novel association in peripheral nervous system myelin. J. Neurosci. 19:3396, 1999. 69. D’Urso, D., and Muller, H. W.: Ins and outs of peripheral myelin protein-22: mapping transmembrane topology and intracellular sorting. J. Neurosci. Res. 49:551, 1997. 70. Engberg, J., Nachamkin, I., Fussing, V., et al.: Absence of clonality of Campylobacter jejuni in serotypes other than HS:19 associated with Guillain-Barre syndrome and gastroenteritis. J. Infect. Dis. 184:215, 2001. 71. Eurelings, M., Moons, K. G., Notermans, N. C., et al.: Neuropathy and IgM M-proteins: prognostic value of antibodies to MAG, SGPG, and sulfatide. Neurology 56:228, 2001. 72. Fabian, R. H.: Uptake of antineuronal IgM by CNS neurons: comparison with antineuronal IgG. Neurology 40:419, 1990. 73. Fabian, R. H.: Retrograde axonal transport and transcytosis of immunoglobulins: implications for the pathogenesis of autoimmune motor neuron disease [review]. Adv. Neurol. 56:433, 1991. 74. Fabian, R. H., and Petroff, G.: Intraneuronal IgG in the central nervous system: uptake by retrograde axonal transport. Neurology 37:1780, 1987. 75. Fernandez, H., Giusti, G., and Bertoglio, J. C.: Effect of the complement system on the sensitivity of Campylobacter jejuni and Campylobacter coli to human blood serum. Braz. J. Med. Biol. Res. 28:227, 1995. 76. Ferrari, S., Morbin, M., Nobile-Orazio, E., et al.: Antisulfatide polyneuropathy: antibody-mediated complement attack on peripheral myelin. Acta Neuropathol. 96:569, 1998. 77. Furuya, S., Hashikawa, T., Irie, F., et al.: Neuronal expression of a minor monosialosyl ganglioside GM1b in rat brain: immunochemical characterization using a specific monoclonal antibody. Neurosci. Res. 22:411, 1995. 78. Gabriel, C., Gregson, N., Wood, N. W., and Hughes, R. A. C.: An immunological study of hereditary motor and sensory neuropathy type 1a (HSMN1a). J. Neurol. Neurosurg. Psychiatry 72:230, 2002. 79. Gabriel, C. M., Gregson, N. A., and Hughes, R. A. C.: Sensory neuropathy and anti-Hu antibodies in a patient with seminoma. Eur. J. Neurol. 3:471, 1996.

80. Gabriel, C. M., Gregson, N. A., and Hughes, R. A. C.: Antibodies to PMP22 in patients with inflammatory neuropathies. J. Neurol. Neurosurg. Psychiatry 66:260, 1999. 81. Gabriel, C. M., Hughes, R. A. C., Moore, S. E., et al.: Induction of experimental neuritis with peripheral myelin protein 22. Brain 121:1895, 1998. 82. Ganser, A. L., and Kirschner, D. A.: Differential expression of gangliosides on the surfaces of myelinated nerve fibers. J. Neurosci. Res. 12:245, 1984. 83. Ganser, A. L., Kirschner, D. A., and Willinger, M.: Ganglioside localization on myelinated nerve fibres by cholera toxin binding. J. Neurocytol. 12:921, 1983. 84. Gilbert, M., Brisson, J. R., Karwaski, M. F., et al.: Biosynthesis of ganglioside mimics in Campylobacter jejuni OH4384: identification of the glycosyltransferase genes, enzymatic synthesis of model compounds, and characterization of nanomole amounts by 600-mHz (1)H and (13)C NMR analysis. J. Biol. Chem. 275:3896, 2000. 85. Goban, Y., Saida, T., Saida, K., et al.: Ultrastructural study of central nervous system demyelination in galactocerebroside sensitized rabbits. Lab. Invest. 55:86, 1986. 86. Gong, Y., Tagawa, Y., Lunn, M. P. T., et al.: Localization of major gangliosides in the PNS: implications for immune neuropathies. Brain 125:2491, 2002. 87. Goodyear, C. S., O’Hanlon, G. M., Plomp, J. J., et al.: Monoclonal antibodies raised against Guillain Barré syndrome-associated Campylobacter jejuni lipopolysaccharides react with neuronal gangliosides and paralyse nerve muscle preparations. J. Clin. Invest. 104:697, 1999. 88. Granieri, E., Casetta, I., Govoni, V., et al.: Ganglioside therapy and Guillain-Barre syndrome: a historical cohort study in Ferrara, Italy, fails to demonstrate an association. Neuroepidemiology 10:161, 1991. 89. Graus, F., Campo, E., Cruz Sanchez, F., et al.: Expression of lymphocyte, macrophage and class I and II major histocompatibility complex antigens in normal human dorsal root ganglia. J. Neurol. Sci. 98:203, 1990. 90. Graus, F., Cordon-Cardo, C., and Posner, J.: Neuronal antinuclear antibody in sensory neuronopathy from lung cancer. Neurology 35:538, 1985. 91. Graus, F., Ribalta, T., Campo, E., et al.: Immunohistochemical analysis of the immune reaction in the nervous system in paraneoplastic encephalomyelitis. Neurology 40:219, 1990. 92. Gregson, N. A., and Leibowitz, S.: IgM paraproteinaemia, polyneuropathy and myelin-associated glycoprotein (MAG). Neuropathol. Appl. Neurobiol. 11:329, 1985. 93. Gregson, N. A., Rees, J. H., and Hughes, R. A. C.: Reactivity of serum IgG anti-GM1 antibodies with the lipopolysaccharide fractions of Campylobacter jejuni isolates from patients with Guillain Barre syndrome (GBS). J. Neuroimmunol. 73:28, 1997. 94. Guilbot, A., Williams, A., Ravise, N., et al.: A mutation in periaxin is responsible for CMT4F, an autosomal recessive form of Charcot-Marie-Tooth disease. Hum. Mol. Genet. 10:415, 2001. 95. Hadden, R. D., Karch, H., Hartung, H. P., et al.: Preceding infections, immune factors, and outcome in Guillain-Barre syndrome. Neurology 56:758, 2001. 96. Hahn, A. F., Feasby, T. E., Wilkie, L., and Lovgren, D.: Antigalactocerebroside antibody increases demyelination in

Peripheral Nerve Antigens

97.

98.

99.

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

110.

111.

112.

adoptive transfer experimental allergic neuritis. Muscle Nerve 16:1174, 1993. Hao, Q., Saida, T., Kuroki, S., et al.: Antibodies to gangliosides and galactocerebroside in patients with Guillain-Barre syndrome with preceding Campylobacter jejuni and other identified infections. J. Neuroimmunol. 81:116, 1998. Harvey, H. A., Swords, W. E., and Apicella, M. A.: The mimicry of human glycolipids and glycosphingolipids by the lipooligosaccharides of pathogenic Neisseria and Haemophilus. J. Autoimmun. 16:257, 2001. Hayday, A. C.:  Cells: a right time and a right place for a conserved third way of protection. Annu. Rev. Immunol. 18:975, 2000. Hays, A. P., Latov, N., Takatsu, M., and Sherman, W. H.: Experimental demyelination of nerve induced by serum of patients with neuropathy and an anti MAG IgM M protein. Neurology 37:242, 1987. Heckmann, J. G., Sommer, J. B., Druschky, A., et al.: Acute motor axonal neuropathy associated with IgM anti-GM1 following Mycoplasma pneumoniae infection. Eur. Neurol. 41:175, 1999. Hirabayashi, Y., Nakao, T., Irie, F., et al.: Structural characterization of a novel cholinergic neuron-specific ganglioside in bovine brain. J. Biol. Chem. 267:12973, 1992. Hirakawa, M., McCabe, T., and Kawata, M.: Time-related changes in the labeling pattern of motor and sensory neurons innervating the gastrocnemius muscle, as revealed by the retrograde transport of the cholera toxin B subunit. Cell Tissue Res. 267:419, 1992. Hitoshi, S., Kusunoki, S., Murayama, S., et al.: Rabbit experimental sensory ataxic neuropathy: anti-GD1b antibody-mediated trkC downregulation of dorsal root ganglia neurons. Neurosci. Lett. 260:157, 1999. Ho, T. W., Mishu, B., Li, C. Y., et al.: Guillain-Barre syndrome in northern China: relationship to Campylobacter jejuni infection and anti-glycolipid antibodies. Brain 118:597, 1995. Ho, T. W., Willison, H. J., Nachamkin, I., et al.: Anti-GD1a antibody is associated with axonal but not demyelinating forms of Guillain-Barre syndrome. Ann. Neurol. 45:168, 1999. Hoek, R. M., Ruuls, S. R., Murphy, C. A., et al.: Downregulation of the macrophage lineage through interaction with OX2 (CD200). Science 290:1768, 2000. Hong, J., Ohmura, K., Mahmood, U., et al.: Arthritis critically dependent on innate immune system players. Immunity 16:157, 2002. Honke, K., Hirahara, Y., Dupree, J., et al.: Paranodal junction formation and spermatogenesis require sulfoglycolipids. Proc. Natl. Acad. Sci. U. S. A. 99:4227, 2002. Honnorat, J., Antoine, J. C., and Belin, M. F.: Are the “newly discovered” paraneoplastic anticollapsin responsemediator protein 5 antibodies simply anti-CV2 antibodies? Ann. Neurol. 50:688, 2001. Honnorat, J., Antoine, J. C., Derrington, E., et al.: Antibodies to a subpopulation of glial cells and a 66 kDa developmental protein in patients with paraneoplastic neurological syndromes. J. Neurol. Neurosurg. Psychiatry 61:270, 1996. Hood, D. W., Cox, A. D., Gilbert, M., et al.: Identification of a lipopolysaccharide alpha-2,3-sialyltransferase from Haemophilus influenzae. Mol. Microbiol. 39:341, 2001.

601

113. Hormigo, A., and Lieberman F.: Nuclear localization of anti-Hu antibody is not associated with in vitro cytotoxicity. J. Neuroimmunol. 55:205, 1994. 114. Hughes, R. A. C.: Guillain-Barré Syndrome. Heidelberg, Springer-Verlag, 1990. 115. Hughes, R. A. C., Gregson N. A., Hadden R. D. M., and Smith K. J.: Pathogenesis of Guillain-Barré syndrome. J. Neuroimmunol. 100:74, 1999. 116. Hughes, R. A. C., Powell H. C., Braheny S. L., and Brostoff S. W.: Endoneurial injection of antisera to myelin antigens. Muscle Nerve 8:516, 1985. 117. Ilyas, A. A., Li, S. C., Chou, D. K. H., et al.: Gangliosides GM2, IV4GalNAcGM1b, and IV4GalNAcGD1a as antigens for monoclonal immunoglobulin M in neuropathy associated with gammopathy. J. Biol. Chem. 263:4369, 1988. 118. Ilyas, A. A., Willison, H. J., Dalakas, M. C., et al.: Identification and characterization of gangliosides reacting with IgM paraproteins in three patients with neuropathy associated with biclonal gammopathy. J. Neurochem. 51:851, 1988. 119. Inoue, M., Fujii, Y., Furukawa, K., et al.: Refractory skin injury in complex knock-out mice expressing only the GM3 ganglioside. J. Biol. Chem. 277:29881, 2002. 120. Irie, F., Kurono, S., Li, Y.-T., et al.: Isolation of three novel cholinergic neuron-specific gangliosides from bovine brain and their in vitro syntheses. Glycoconj. J. 13:177, 1996. 121. Irie, S., Kanazawa, N., Ogino, M., et al.: No cytomegalovirus DNA in sera from patients with anti-MAG/SGPG antibodyassociated neuropathy. Ann. Neurol. 47:274, 2000. 122. Ishibashi, T., Dupree, J. L., Ikenaka, K., et al.: A myelin galactolipid, sulfatide, is essential for maintenance of ion channels on myelinated axon but not essential for initial cluster formation. J. Neurosci. 22:6507, 2002. 123. Ishida, K., Takeuchi, H., Takahashi, R., et al.: A possible novel isoform of peripheral myelin PO protein: a target antigen recognized by an autoantibody in a patient with malignant lymphoma and peripheral neuropathy. J. Neurol. Sci. 188:43, 2001. 124. Ito, A., Handa, K., Withers, D. A., et al.: Binding specificity of siglec7 to disialogangliosides of renal cell carcinoma: possible role of disialogangliosides in tumor progression. FEBS Lett. 504:82, 2001. 125. Izumo, S., Linington, C., Wekerle, H., and Meyermann, R.: Morphological study on EAN mediated by T-cell line specific for bovine P2 protein in Lewis rats. Lab. Invest. 53:209, 1985. 126. Jacobs, B. C., Hazenberg, M. P., van Doorn, P. A., et al.: Cross-reactive antibodies against gangliosides and Campylobacter jejuni lipopolysaccharides in patients with Guillain-Barre or Miller Fisher syndrome. J. Infect. Dis. 175:729, 1997. 127. Jacobs, B. C., Rothbarth, P. H., van der Meché, F. G. A., et al.: The spectrum of antecedent infections in Guillain-Barré syndrome: a case control study. Neurology 51:1110, 1998. 128. Jacobs, B. C., van Doorn, P. A., Groeneveld, J. H., et al.: Cytomegalovirus infections and anti-GM2 antibodies in Guillain-Barre syndrome. J. Neurol. Neurosurg. Psychiatry 62:641, 1997. 129. Jessen, K. R., Morgan, L., Brammer, M., and Mirsky, R.: Galactocerebroside is expressed by non-myelin-forming Schwann cells in situ. J. Cell Biol. 101:1135, 1985.

602

Neuroimmunology of the Peripheral Nervous System

130. Jeurissen, S. H., Janse, E. M., van Rooijen, N., and Claassen, E.: Inadequate anti-polysaccharide antibody responses in the chicken. Immunobiology 198:385, 1998. 131. Jones, P. A., Samuels, N. M., Phillips, N. J., et al.: Haemophilus influenzae type b strain A2 has multiple sialyltransferases involved in lipooligosaccharide sialylation. J. Biol. Chem. 277:14598, 2002. 132. Jones, T. A., Bergfors, T., Sedzik, J., and Unge, T.: The threedimensional structure of P2 myelin protein. EMBO J. 7:1597, 1988. 133. Kadlubowski, M., and Hughes, R. A. C.: Identification of the neuritogen responsible for experimental allergic neuritis. Nature 277:140, 1979. 134. Kadlubowski, M., Hughes, R. A. C., and Gregson, N. A.: Experimental allergic neuritis in the Lewis rat: characterisation of the activity of peripheral myelin and its major basic protein P2. Brain Res. 184:439, 1980. 135. Kaida, K., Kusunoki S., Kamakura K., et al.: Guillain-Barré syndrome with antibody to a ganglioside, N-acetylgalactosaminyl GD1a. Brain 123:116, 2000. 136. Kanda, T., Yamawaki, M., Iwasaki, T., and Mizusawa, H.: Glycosphingolipid antibodies and blood-nerve barrier in autoimmune demyelinative neuropathy. Neurology 54:1459, 2000. 137. Kanda, T., Yoshino, H., Ariga, T., et al.: Glycosphingolipid antigens in cultured microvascular bovine brain endothelial cells: sulfoglucuronosyl paragloboside as a target of monoclonal IgM in demyelinative neuropathy. J. Cell Biol. 126:235, 1994. 138. Kawano, T., Cui, J., Koezuka, Y., et al.: CD1d-restricted and TCR-mediated activation of valpha14 NKT cells by glycosylceramides. Science 278:1626, 1997. 139. Khalili-Shirazi, A., Atkinson, P., Gregson, N., and Hughes, R. A. C.: Antibody responses to P0 and P2 myelin proteins in Guillain-Barré syndrome and chronic idiopathic demyelinating polyradiculoneuropathy. J. Neuroimmunol. 46:245, 1993. 140. Khalili-Shiraz, A., Gregson, N., Gray, I., et al.: Antiganglioside antibodies in Guillain-Barré syndrome after a recent cytomegalovirus infection. J. Neurol. Neurosurg. Psychiatry 66:376, 1999. 141. Khalili-Shiraz, A., Gregson, N. A., Londei, M., et al.: The distribution of CD1 molecules in inflammatory neuropathy. J. Neurol. Sci. 158:154, 1998. 142. Khalili-Shirazi, A., Hughes, R. A. C., Brostoff, S., et al.: T cell response to myelin proteins in Guillain-Barré syndrome. J. Neurol. Sci. 111:200, 1992. 143. Kitazawa, K., Tagawa, Y., Honda, A., and Yuki, N.: GuillainBarré syndrome associated with IgG anti-GM1b antibody subsequent to Mycoplasma pneumoniae infection. J. Neurol. Sci. 156:99, 1998. 144. Koehler, N. K. U., Martin, R., and Wiethölter, H.: The antibody repertoire in experimental allergic neuritis: evidence for PMP-22 as a novel neuritogen. J. Neuroimmunol. 71:179, 1996. 145. Koga, M., Ang, C. W., Yuki, N., et al.: Comparative study of preceding Campylobacter jejuni infection in Guillain-Barré syndrome in Japan and the Netherlands. J. Neurol. Neurosurg. Psychiatry 70:693, 2001. 146. Koga, M., Yoshino, H., Morimatsu, M., and Yuki, N.: AntiGT1a IgG in Guillain-Barré syndrome. J. Neurol. Neurosurg. Psychiatry 72:767, 2002.

147. Koga, M., Yuki, N., Tai, T., and Hirata, K.: Miller Fisher syndrome and Haemophilus influenzae infection. Neurology 57:686, 2001. 148. Kohriyama, T., Kusunoki, S., and Ariga, T.: Subcellular localization of sulfated glucuronic acid-containing glycolipids reacting with anti-myelin-associated glycoprotein antibody. J. Neurochem. 48:1516, 1987. 149. Kornberg, A. J., and Pestronk, A.: Chronic motor neuropathies: diagnosis, therapy, and pathogenesis. Ann. Neurol. 37(Suppl. 1):S43, 1995. 150. Kornguth, S. E., Kalinke, T., Grunwald, G. B., et al.: Antineurofilament antibodies in the sera of patients with small cell carcinoma of the lung and with visual paraneoplastic syndrome. Cancer Res. 46:2588, 1986. 151. Kotani, M., Kawashima, I., Ozawa, H., et al.: Immunohistochemical localization of minor gangliosides in the rat central nervous system. Glycobiology 4:855, 1994. 152. Kuhlmann, T., Bitsch, A., Stadelmann, C., et al.: Macrophages are eliminated from the injured peripheral nerve via local apoptosis and circulation to regional lymph nodes and the spleen. J. Neurosci. 21:3401, 2001. 153. Kullmann, D. M.: The neuronal channelopathies. Brain 125:1177, 2002. 154. Kumagai, T., Kitagawa, Y., Hirose, G., and Sakai, K.: Antibody recognition and RNA binding of a neuronal nuclear autoantigen associated with paraneoplastic neurological syndromes and small cell lung carcinoma. J. Neuroimmunol. 93:37, 1999. 155. Kuroki, S., Saida, T., Nukina, M., et al.: Campylobacter jejuni strains from patients with Guillain-Barré syndrome belong mostly to Penner serogroup 19 and contain betaN-acetylglucosamine residues. Ann. Neurol. 33:243, 1993. 156. Kusunoki, S.: Antiglycolipid antibody in inflammatory neuropathy. Clin. Neurol. 35:1370, 1995. 157. Kusunoki, S.: Antiglycolipid antibodies in Guillain-Barré syndrome and autoimmune neuropathies. Am. J. Med. Sci. 319:234, 2000. 158. Kusunoki, S., Chiba, A., Hitoshi, S., et al.: Anti-Gal-C antibody in autoimmune neuropathies subsequent to mycoplasma infection. Muscle Nerve 18:409, 1995. 159. Kusunoki, S., Chiba, A., and Kanazawa, I.: Anti-GQ1b IgG antibody is associated with ataxia as well as ophthalmoplegia. Muscle Nerve 22:1071, 1999. 160. Kusunoki, S., Chiba, A., Kon, K., et al.: N-acetylgalactosaminyl GD1a is a target molecule for serum antibody in Guillain-Barre syndrome. Ann. Neurol. 35:570, 1994. 161. Kusunoki, S., Chiba, A., Tai, T., and Kanazawa, I.: Localization of GM1 and GD1b antigens in the human peripheral nervous system. Muscle Nerve 16:752, 1993. 162. Kusunoki, S., Hitoshi, S., Kaida, K., et al.: Monospecific anti-GD1b IgG is required to induce rabbit ataxic neuropathy. Ann. Neurol. 45:400, 1999. 163. Kusunoki, S., Inoue, K., Iwamori, M., et al.: Discrimination of human dorsal root ganglion cells by anti-fucosyl GM1 antibody. Brain Res. 494:391, 1989. 164. Kusunoki, S., Inoue, K., Iwamori, M., et al.: Fucosylated glycoconjugates in human dorsal root ganglion cells with unmyelinated axons. Neurosci. Lett. 126:159, 1991. 165. Kusunoki, S., Mashiko, H., Mochizuki, N., et al.: Binding of antibodies against GM1 and GD1b in human peripheral nerve. Muscle Nerve 20:840, 1997.

Peripheral Nerve Antigens 166. Kusunoki, S., Shiina, M., and Kanazawa, I.: Anti-Gal-C antibodies in GBS subsequent to mycoplasma infection: evidence of molecular mimicry. Neurology 57:736, 2001. 167. Kusunoki, S., Shimizu, J., Chiba, R., et al.: Experimental sensory neuropathy induced by sensitisation with ganglioside GD1b. Ann. Neurol. 39:324, 1996. 168. Kwa, M. S., van Schaik, I. N., Brand, A., et al.: Investigation of serum response to PMP22, connexin 32 and P(0) in inflammatory neuropathies. J. Neuroimmunol. 116:220, 2001. 169. Laemmli, U. K.: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680, 1970. 170. Lastovica, A. J., Goddard, E. A., and Argent, A. C.: Guillain-Barré syndrome in South Africa associated with Campylobacter jejuni O:41 strains. J. Infect. Dis. 176:s139, 1997. 171. Latov, N.: Antibodies to glycoconjugates in neuropathy and motor-neuron disease. Proc. Brain Res. 101:295, 1994. 172. Latov, N., Hays, A. P., Donofrio, P. D., et al.: Monoclonal IgM with unique specificity to gangliosides GM1 and GD1b and to lacto-N-tetraose associated with human motor neuron disease. Neurology 38:763, 1988. 173. Latov, N., Hays, A. P., Yu, R. K., et al.: Antibodies to glycoconjugates in human motor neuron disease. Neurochem. Pathol. 8:181, 1988. 174. Lawlor, M. W., Richards, M. P., De Vries, G. H., et al.: Antibodies to L-periaxin in sera of patients with peripheral neuropathy produce experimental sensory nerve conduction deficits. J. Neurochem. 83:592, 2002. 175. Ledeen, R. W.: Gangliosides of the neuron. Trends Neurosci. 10:169, 1985. 176. Lennon, V. A., Kryzer, T. J., Griesmann, G. E., et al.: Calcium-channel antibodies in the Lambert-Eaton syndrome and other paraneoplastic syndromes. N. Engl. J. Med. 332:1467, 1995. 177. Levitt, D., Griffin, N. B., and Egan, M. L.: Mitogeninduced plasma cell differentiation in patients with multiple sclerosis. J. Neuroimmunol. 124:2117, 1980. 178. Lewis, R. A., Sumner, A. J., and Shy, M. E.: Electrophysiological features of inherited demyelinating neuropathies: a reappraisal in the era of molecular diagnosis. Muscle Nerve 23:1472, 2000. 179. Li, J. Y., De Camilli, P., and Dahlstrom, A.: Intraneuronal trafficking and distribution of amphiphysin and synaptojanin in the rat peripheral nervous system and the spinal cord. Eur. J. Neurosci. 9:1864, 1997. 180. Lindberg, F. P., Bullard, D. C., Caver, T. E., et al.: Decreased resistance to bacterial infection and granulocyte defects in IAP-deficient mice. Science 274:795, 1996. 181. Linington, C., and Brostoff, S. W.: Peripheral nerve antigens. In Dyck, P. J., Thomas, P. K., Griffin, J. W., et al. (eds.): Peripheral Neuropathy, 3rd ed. Philadelphia, W. B. Saunders, p. 404, 1993. 182. Linington, C., Izumo, S., Suzuki, M., et al.: A permanent rat T cell line that mediates experimental allergic neuritis in the rat in vivo. J. Immunol. 133:1946, 1984. 183. Linington, C., Lassmann, H., Ozawa, K., et al.: Cell adhesion molecules of the immunoglobulin supergene family as tissue-specific autoantigens: induction of experimental allergic neuritis (EAN) by P0 protein-specific T cell lines. Eur. J. Immunol. 22:1813, 1992.

603

184. Linton, D., Gilbert, M., Hitchen, P. G., et al.: Phase variation of a -1,3 galactosyltransferase involved in generation of the ganglioside GM1-like lipo-oligosaccharide of Campylobacter jejuni. Mol. Microbiol. 37:501, 2000. 185. Linton, D., Karlyshev, A. V., Hitchen, P. G., et al.: Multiple N-acetyl neuraminic acid synthetase (neuB) genes in Campylobacter jejuni: identification and characterization of the gene involved in sialylation of lipo-oligosaccharide. Mol. Microbiol. 35:1120, 2000. 186. Lisak, R. P., Saida, T., and Kennedy, P. G. E.: EAE, EAN and galactocerebroside sera bind to oligodendrocytes and Schwann cells. J. Neurol. Sci. 48:287, 1980. 187. Liu, J., Dalmau, J., Szabo, A., et al.: Paraneoplastic encephalomyelitis antigens bind to the Au-rich elements of mRNA. Neurology 45:544, 1995. 188. Lloyd, K. O., Gordon, C. M., Thampoe, I. J., and DiBenedetto, C.: Cell surface accessibility of individual gangliosides in malignant melanoma cells to antibodies is influenced by the total ganglioside composition of the cells. Cancer Res. 52:4948, 1992. 189. Lopate, G., Kornberg, A. J., Yue, J., et al.: Anti-myelin associated glycoprotein antibodies: variability in patterns of IgM binding to peripheral nerve. J. Neurol. Sci. 188:67, 2001. 190. Lopate, G., Pestronk, A., Kornberg, A. J., et al.: IgM antisulfatide autoantibodies: patterns of binding to cerebellum, dorsal root ganglion and peripheral nerve. J. Neurol. Sci. 151:189, 1997. 191. Lugaresi, A., Corbo, M., Thomas, F. P., et al.: Identification of glycoconjugates which are targets for anti-Gal(-1-3)GalNAc autoantibodies in spinal motor neurons. J. Neuroimmunol. 34:69, 1991. 192. Lunn, M. P., Crawford, T. O., Hughes, R. A., et al.: Antimyelin-associated glycoprotein antibodies alter neurofilament spacing. Brain 125:904, 2002. 193. Lunn, M. P., Muir, P., Brown, L. J., et al.: Cytomegalovirus is not associated with IgM anti-myelin-associated glycoprotein/sulphate-3-glucuronyl paragloboside antibody-associated neuropathy. Ann. Neurol. 46:267, 1999. 194. Madden, K. S., Sanders, V. M., and Felten, D. L.: Catecholamine influences and sympathetic neural modulation of immune responsiveness [review]. Annu. Rev. Pharmacol. Toxicol. 35:417, 1995. 195. Maehara, T., Ono, K., Tsutsui, K., et al.: A monoclonal antibody that recognizes ganglioside GD1b in the rat central nervous system. Neurosci. Res. 29:9, 1997. 196. Mandrell, R. E., McLaughlin, R., Aba Kwaik, Y., et al.: Lipooligosaccharides (LOS) of some Haemophilus species mimic human glycosphingolipids, and some LOS are sialylated. Infect. Immun. 60:1322, 1992. 197. Manley, G. T., Smitt, P. S., Dalmau, J., and Posner, J. B.: Hu antigens: reactivity with Hu antibodies, tumour expression, and major immunogenic sites. Ann. Neurol. 38:102, 1995. 198. Mansson, M., Hood, D. W., Li, J., et al.: Structural analysis of the lipopolysaccharide from nontypeable Haemophilus influenzae strain 1003. Eur. J. Biochem. 269:808, 2002. 199. Martini, R., and Schachner, M.: Immunoelectron microscopic localization of neural cell adhesion molecules (L1, N-CAM, and MAG) and their shared carbohydrate epitope and myelin basic protein in developing sciatic nerve. J. Cell Biol. 103:2439, 1986.

604

Neuroimmunology of the Peripheral Nervous System

200. Martini, R., Zielasek, J., Toyka, K. V., et al.: Protein zero (P0)-deficient mice show myelin degeneration in peripheral nerves characteristic of inherited human neuropathies. Nat. Genet. 11:281, 1995. 201. Marx, A., Kirchner, T., Greiner, A., et al.: Neurofilament expression in thymic epithelial tumors and anti-axonal autoantibodies in myasthenia gravis: a model for autoimmunity by abnormal T cell selection [in German]. Verh. Dtsch. Ges. Pathol. 76:256, 1992. 202. Marx, A., Wilisch, A., Schultz, A., et al.: Expression of neurofilaments and of a titin epitope in thymic epithelial tumors: implications for the pathogenesis of myasthenia gravis. Am. J. Pathol. 148:1839, 1996. 203. Mason, D.: A very high level of crossreactivity is an essential feature of the T-cell receptor. Immunol. Today 19:395, 1998. 204. Mason, W. P., Verschuuren, J., Graus, F., et al.: Anti-Hu antibodies in patients with small-cell lung cancer but no paraneoplastic disorder. Ann. Neurol. 38:341, 1995. 205. McKhann, G. M., Cornblath, D. R., Griffin, J. W., et al.: Acute motor axonal neuropathy: a frequent cause of acute flaccid motor paralysis in China. Ann. Neurol. 33:333, 1993. 206. Meléndez-Vásquez, C., and Gregson, N. A.: Characterization and partial purification of a novel 36 kDa peripheral myelin protein recognized by the sera of patients with neurological disorders. J Neuroimmunol 91:10, 1998. 207. Meléndez-Vásquez, C., Redford, J., Choudhary, P. P., et al.: Immunological investigation of chronic inflammatory demyelinating polyradiculoneuropathy. J. Neuroimmunol. 73:124, 1997. 208. Milner, P., Lovelidge, C. A., Taylor, W. A., and Hughes, R. A. C.: P0 myelin protein produces experimental allergic neuritis in Lewis rats. J. Neurol. Sci. 79:275, 1987. 209. Milner, R. J., Lai, C., Nave, K.-A., et al.: Organisation of myelin protein genes: myelin-associated glycoprotein. In Duncan, I., Skoff, R. P., and Colman, D. (eds.): Myelination and Demyelination. New York, New York Academy of Sciences, p. 254, 1990. 210. Miyazaki, T., Kusunoki, S., Kaida, K., et al.: Guillain-Barre syndrome associated with IgG monospecific to ganglioside GD1b. Neurology 56:1227, 2001. 211. Molander, M., Berthold, C.-H., Persson, H., et al.: Monosialoganglioside (GM1) immunofluorescence in rat spinal roots studied with a monoclonal antibody. J. Neurocytol. 26:101, 1997. 212. Moody, D. B., Besra, G. S., Wilson, I. A., and Porcelli, S. A.: The molecular basis of CD1-mediated presentation of lipid antigens. Immunol. Rev. 172:285, 1999. 213. Mori, M., Kuwabara, S., Miyake, M., et al.: Haemophilus influenzae infection and Guillain-Barre syndrome. Brain 123:2171, 2000. 214. Muller, H. W.: Tetraspan myelin protein PMP22 and demyelinating peripheral neuropathies: new facts and hypotheses. Glia 29:182, 2000. 215. Muresu, R., Baldini, A., Gress, T., et al.: Mapping of the gene coding for a paraneoplastic encephalomyelitis antigen (HuD) to human chromosome site 1p34. Cytogenet. Cell Genet. 65:177, 1994. 216. Myszewski, M. A., and Stern, N. J.: Phagocytosis and intracellular killing of Campylobacter jejuni by elicited chicken peritoneal macrophages. Avian Dis. 35:750, 1991.

217. Nachamkin, I., Engberg, J., Gutacker, M., et al.: Molecular population genetic analysis of Campylobacter jejuni HS:19 associated with Guillain-Barre syndrome and gastroenteritis. J. Infect. Dis. 184:221, 2001. 218. Naidenko, O. V., Koezuka, Y., and Kronenberg, M.: CD1mediated antigen presentation of glycosphingolipids. Microbes Infect. 2:621, 2000. 219. Nardelli, E., Bassi, A., Mazzi, G., et al.: Systemic passive transfer studies using IgM monoclonal antibodies to sulfatide. J. Neuroimmunol. 63:29, 1995. 220. Nemni, R., Fazio, R., Quattrini, A., et al.: Antibodies to sulfatide and to chondroitin sulfate C in patients with chronic sensory neuropathy. J. Neuroimmunol. 43:79, 1993. 221. Neumann, H., Cavalie, A., Jenne, D. E., and Wekerle, H.: Induction of MHC class I genes in neurons. Science 269:549, 1995. 222. Nishimura, M., Nukina, M., Kuroki, S., et al.: Characterization of Campylobacter jejuni isolates from patients with GuillainBarré syndrome. J. Neurol. Sci. 153:91, 1997. 223. Nishimura, M., Saida, T., Kuroki, S., et al.: Post-infectious encephalitis with anti-galactocerebroside antibody subsequent to Mycoplasma pneumoniae infection. J. Neurol. Sci. 140:91, 1996. 224. Nobile-Orazio, E., Manfredini, E., Carpo, M., et al.: Frequency and clinical correlates of anti-neural IgM antibodies in neuropathy associated with IgM monoclonal gammopathy. Ann. Neurol. 36:416, 1994. 225. Nobile-Orazio, E., Manfredini, E., Sgarzi, M., et al.: Serum IgG antibodies to a 35-kDa P0-related glycoprotein in motor neuron disease. J. Neuroimmunol. 53:143, 1994. 226. Notterpek, L., Roux, K. J., Amici, S. A., et al.: Peripheral myelin protein 22 is a constituent of intercellular junctions in epithelia. Proc. Natl. Acad. Sci. U. S. A. 98:14404, 2001. 227. Obrocki, J., and Borroni, E.: Immunocytochemical evaluation of a cholinergic-specific ganglioside antigen (Chol-1) in the central nervous system of the rat. Exp. Brain Res. 72:71, 1988. 228. Ogawa-Goto, K., and Abe, T.: Gangliosides and glycosphingolipids of peripheral nervous system myelins—a minireview. Neurochem. Res. 23:305, 1998. 229. Ogawa-Goto, K., Funamoto, N., Ohta, Y., et al.: Myelin gangliosides of human peripheral nervous system: an enrichment of GM1 in the motor nerve myelin isolated from cauda equina. J. Neurochem. 59:1844, 1992. 230. O’Hanlon, G. M., Paterson, G. J., Veitch, J., et al.: Mapping immunoreactive epitopes in the human peripheral nervous system using human monoclonal anti-GM1 ganglioside antibodies. Acta Neuropathol. 95:605, 1998. 231. O’Hanlon, G. M., Paterson, G. J., Wilson, G., et al.: AntiGM1 ganglioside antibodies cloned from autoimmune neuropathy patients show diverse binding patterns in the rodent nervous system. J. Neuropathol. Exp. Neurol. 55:184, 1996. 232. O’Hanlon, G. M., Plomp, J. J., Chakrabarti, M., et al.: AntiGQ1b ganglioside antibodies mediate complementdependent destruction of the motor nerve terminal. Brain 124:893, 2001. 233. O’Hanlon, G. M., Veitch, J., Gallardo, E., et al.: Peripheral neuropathy associated with anti-GM2 ganglioside antibodies:

Peripheral Nerve Antigens

234.

235.

236.

237.

238.

239.

240.

241.

242.

243.

244.

245.

246.

247.

248.

249.

250.

clinical and immunopathological studies. Autoimmunity 32:133, 2000. Ohsawa, T., Miyatake, T., and Yuki, N.: Anti-B-series ganglioside-recognizing autoantibodies in an acute sensory neuropathy patient cause cell death of rat dorsal root ganglion neurons. Neurosci. Lett. 157:167, 1993. Oldenborg, P. A., Zheleznyak, A., Fang, Y. F., et al.: Role of CD47 as a marker of self on red blood cells. Science 288:2051, 2000. O’Leary, C. P., and Willison, H. J.: Autoimmune ataxic neuropathies (sensory ganglionopathies). Curr. Opin. Neurol. 10:366, 1997. O’Leary, C.P., and Willison, H. J.: The role of antiglycolipid antibodies in peripheral neuropathies. Curr. Opin. Neurol. 13:583, 2000. Olee, T., Powell H. C., and Brostoff S. W.: New minimum length requirement for a T cell epitope for experimental allergic neuritis. J. Neuroimmunol. 27:187, 1990. Olee, T., Weise, M., Powers, J., and Brostoff, S. W.: A T cell epitope for experimental allergic neuritis is an amphipathic alpha-helical structure. J. Neuroimmunol. 21:235, 1989. Pan, C. L., Yuki, N., Koga, M., et al.: Acute sensory ataxic neuropathy associated with monospecific anti-GD1b IgG antibody. Neurology 57:1316, 2001. Pancorbo, P.L., de Pablo, M.A., Ortega, E., et al.: Potential intervention of Campylobacter jejuni in the modulation of murine immune response. Curr. Microbiol 43:209, 2001. Pancorbo, P. L., Gallego, A. M., de Pablo, M., et al.: Inflammatory and phagocytic response to experimental Campylobacter jejuni infection in mice. Microbiol. Immunol. 38:89, 1994. Paparounas, K., O’Hanlon, G. M., O’Leary, C. P., et al.: Anti-ganglioside antibodies can bind peripheral nerve nodes of Ranvier and activate the complement cascade without inducing acute conduction block in vitro. Brain 122:807, 1999. Paul, J. A., and Gregson, N. A.: An immunohistochemical study of phospholipase A2 in peripheral nerve during Wallerian degeneration. J. Neuroimmunol. 39:31, 1992. Perego, L., Previtali, S. C., Nemni, R., et al.: Autoantibodies to amphiphysin I and amphiphysin II in a patient with sensory-motor neuropathy. Eur. Neurol. 47:196, 2002. Pestronk, A., Choksi, R., Yee, W. C., et al.: Serum antibodies to heparan sulfate glycosaminoglycans in Guillain-Barre syndrome and other demyelinating polyneuropathies. J. Neuroimmunol. 91:204, 1998. Pestronk, A., Li, F., Griffin, J., et al.: Polyneuropathy syndromes associated with serum antibodies to sulfatide and myelin-associated glycoprotein. Neurology 41:357, 1991. Peterson, K., Rosenblum, M. K., Kotanides, H., and Posner, J. B.: Paraneoplastic cerebellar degeneration. I. A clinical analysis of 55 anti-Yo antibody-positive patients. Neurology 42:1931, 1992. Plomp, J. J., Molenaar, P. C., O’Hanlon, G. M., et al.: Miller Fisher anti-GQ1b antibodies: -latrotoxin-like effects on motor end plates. Ann. Neurol. 45:189, 1999. Pollard, J. D., Westland, K. W., Harvey, G. K., et al.: Activated T cells of nonneural specificity open the bloodnerve barrier to circulating antibody. Ann. Neurol. 37:467, 1995.

605

251. Quach, T. T., Rong, Y., Belin, M. F., et al.: Molecular cloning of a new unc-33-like cDNA from rat brain and its relation to paraneoplastic neurological syndromes. Brain Res. Mol. Brain Res. 46:329, 1997. 252. Quarles, R. H., Ilyas, A. A., and Willison, H. J.: Antibodies to glycolipids in demyelinating diseases of the human peripheral nervous system. Chem. Phys. Lipids 42:235, 1986. 253. Quarles, R. H., and Weiss, M. D.: Autoantibodies associated with peripheral neuropathy. Muscle Nerve 22:800, 1999. 254. Quattrini, A., Corbo, M., Dhaliwal, S. K., et al.: Antisulfatide antibodies in neurological disease: binding to rat dorsal root ganglia neurons. J. Neurol. Sci. 112:152, 1992. 255. Raschetti, R., Maggini, M., Popoli, P., et al.: Gangliosides and Guillain-Barre syndrome. J. Clin. Epidemiol. 48:1399, 1995. 256. Rees, J. H., Gregson, N. A., Griffiths, P. L., and Hughes, R. A. C.: Campylobacter jejuni and Guillain-Barre syndrome. Q. J. Med. 86:623, 1993. 257. Rees, J. H., Gregson, N. A., and Hughes, R. A. C.: Antiganglioside GM1 antibodies in Guillain-Barré syndrome and their relationship to Campylobacter jejuni infection. Ann. Neurol. 38:809, 1995. 258. Rees, J. H., Soudain, S. A., Gregson, N. A., and Hughes, R. A. C.: Campylobacter jejuni infection and Guillain-Barré syndrome. N. Engl. J. Med. 333:1374, 1995. 259. Richardson, P. J., Walker, J. H., Jones, R. T., and Whittaker, V. P.: Identification of a cholinergic-specific antigen Chol-1 as a ganglioside. J. Neurochem. 38:1605, 1982. 260. Ritter, G., Fortunato, S. R., Cohen, L., et al.: Induction of antibodies reactive with GM2 ganglioside after immunization with lipopolysaccharides from Campylobacter jejuni. Int. J. Cancer 66:184, 1996. 261. Ritz, J.: The role of NK cells in immune surveillance. N. Engl. J. Med. 320:1748, 1989. 262. Ritz, M. F., Lechner-Scott, J., Scott, R. J., et al.: Characterisation of autoantibodies to peripheral myelin protein 22 in patients with hereditary and acquired neuropathies. J. Neuroimmunol. 104:155, 2000. 263. Sadiq, S. A., Van den Berg, L. H., Thomas, F. P., et al.: Human monoclonal antineurofilament antibody crossreacts with a neuronal surface protein. J. Neurosci. Res. 29:319, 1991. 264. Saida, K., Saida, T., and Brown, M. J.: In vitro demyelination induced by intraneural injection of antigalactocerebroside serum: a morphological study. Am. J. Pathol. 95:99, 1979. 265. Saida, T., Saida, K., and Dorfman, S. H.: Experimental allergic neuritis induced by sensitisation with galactocerebroside. Science 204:1103, 1979. 266. Saida, T., Saida, K., and Silberberg, D. H.: Demyelination produced by experimental allergic neuritis serum and antigalactocerebroside antiserum in central nervous system cultures: an ultrastructural study. Acta Neuropathol. 48:19, 1979. 267. Saiz, A., Dalmau, J., Butler, M. H., et al.: Anti-amphiphysin I antibodies in patients with paraneoplastic neurological disorders associated with small cell lung carcinoma. J. Neurol. Neurosurg. Psychiatry 66:214, 1999.

606

Neuroimmunology of the Peripheral Nervous System

268. Sakai, K., Gofuku, M., Kitagawa, Y., et al.: A hippocampal protein associated with paraneoplastic neurologic syndrome and small cell lung carcinoma. Biochem. Biophys. Res. Commun. 199:1200, 1994. 269. Sakamoto, Y., Kitamura, K., Yoshimura, K., et al.: Complete amino acid sequence of P0 protein in bovine peripheral nerve myelin. J. Biol. Chem. 262:4208, 1987. 270. Salih, A. M., Nixon, N. B., Dawes, P. T., and Mattey, D. L.: Soluble adhesion molecules and anti-endothelial cell antibodies in patients with rheumatoid arthritis complicated by peripheral neuropathy. J. Rheumatol. 26:551, 1999. 271. Salloway, S., Mermel, L. A., Seamans, M., et al.: MillerFisher syndrome associated with Campylobacter jejuni bearing lipopolysaccharide molecules that mimic human ganglioside GD3. Infect. Immun. 64:2945, 1996. 272. Schachner, M., and Bartsch, U.: Multiple functions of the myelin-associated glycoprotein MAG (siglec-4a) in formation and maintenance of myelin. Glia 29:154, 2000. 273. Scherer, S. S.: Molecular specializations at nodes and paranodes in peripheral nerve. Microsc. Res. Tech. 34:452, 1996. 274. Scherer, S. S., and Arroyo, E. J.: Recent progress on the molecular organization of myelinated axons. J. Peripher. Nerv. Syst. 7:1, 2002. 275. Schluep, M., and Steck, A. J.: Immunostaining of motor nerve terminals by IgM M protein with activity against gangliosides GM1 and GD1b from a patient with motor neuron disease. Neurology 38:1890, 1988. 276. Schmid, C. D., Stienekemeier, M., Oehen, S., et al.: Immune deficiency in mouse models for inherited peripheral neuropathies leads to improved myelin maintenance. J. Neurosci. 20:729, 2000. 277. Schultz, A., Hoffacker, V., Wilisch, A., et al.: Neurofilament is an autoantigenic determinant in myasthenia gravis. Ann. Neurol. 46:167, 1999. 278. Sedzik, J., Kotake, Y., and Uyemura, K.: Purification of PASII/PMP22—an extremely hydrophobic glycoprotein of PNS myelin membrane. Neuroreport 9:1595, 1998. 279. Sekido, Y., Bader, S. A., Carbone, D. P., et al.: Molecular analysis of the HuD gene encoding a paraneoplastic encephalomyelitis antigen in human lung cancer. Cancer Res. 54:4988, 1994. 280. Shamshiev, A., Donda, A., Carena, I., et al.: Self glycolipids as T-cell autoantigens. Eur. J. Immunol. 29:1667, 1999. 281. Shapiro, L., Doyle, J. P., Hensley, P., et al.: Crystal structure of the extracellular domain from P0, the major structural protein of peripheral nerve myelin. Neuron 17:435, 1996. 282. Sheikh, K. A., Deerinck, T. J., Ellisman, M. H., and Griffin, J. W.: The distribution of ganglioside-like moieties in peripheral nerves. Brain 122:449, 1999. 283. Sheikh, K. A., Gong, Y., Schnaar, R. L., and Griffin, J. W.: Anti-ganglioside antibody mediated axonal degeneration. J. Peripher. Nerv. Syst. 6:175, 2001. 284. Sheikh, K. A., Nachamkin, I., Ho, T. W., et al.: Campylobacter jejuni lipopolysaccharides in Guillain-Barre syndrome: molecular mimicry and host susceptibility. Neurology 51:371, 1998. 285. Sherman, D. L., Fabrizi, C., Gillespie, C. S., and Brophy, P. J.: Specific disruption of a Schwann cell dystrophin-related protein complex in a demyelinating neuropathy [see comment]. Neuron 30:677, 2001.

286. Sherman, W. H., Latov, N., Hays, A. P., et al.: Monoclonal IgMk antibody precipitating with chondroitin sulfate C from patients with axonal polyneuropathy and epidermolysis. Neurology 33:192, 1983. 287. Shillito, P., Molenaar, P. C., Vincent, A., et al.: Acquired neuromyotonia: evidence for autoantibodies directed against K channels of peripheral nerves. Ann. Neurol. 38:714, 1995. 288. Silber, E., Semra, Y. K., Gregson, N. A., and Sharief, M. K.: Patients with progressive multiple sclerosis have elevated antibodies to neurofilament subunit. Neurology 58:1372, 2002. 289. Sillevis Smitt, P. A. E., Manley, G. T., and Posner, J. B.: Immunization with the paraneoplastic encephalomyelitis antigen HuD does not cause neurologic disease in mice. Neurology 45:1873, 1995. 290. Simons, K., and Ikonen, E.: Functional rafts in cell membranes. Nature 387:569, 1997. 291. Smith, V. V., Gregson, N., Foggensteiner, L., et al.: Acquired intestinal aganglionosis and circulating autoantibodies without neoplasia or other neural involvement. Gastroenterology 112:1366, 1997. 292. Snipes, G. J., Suter, U., and Shooter, E. M.: Human peripheral myelin protein-22 carries the L2/HNK-1 carbohydrate adhesion epitope. J. Neurochem. 61:1961, 1993. 293. Snipes, G. J., Suter, U., Welcher, A. A., and Shooter, E. M.: Characterization of a novel peripheral nervous system myelin protein (PMP-22 /SR13). J. Cell Biol. 117:225, 1992. 294. Sorensen, T. I., Nielsen, G. G., Andersen, P. K., and Teasdale, T. W.: Genetic and environmental influences on premature death in adult adoptees. N. Engl. J. Med. 318:727, 1988. 295. Spreyer, P., Kuhn, G., Hanemann, C. O., et al.: Axonregulated expression of a Schwann cell transcript that is homologous to a “growth arrest-specific” gene. EMBO J. 10:3661, 1991. 296. Stebbins, E. C., and Galán, J. E.: Structural mimicry in bacterial virulence. Nature 412:701, 2001. 297. Stefansson, K., Marton, L. S., Dieperink, M. E., et al.: Circulating autoantibodies to the 200,000-dalton protein of neurofilaments in the serum of healthy individuals. Science 228:1117, 1985. 298. Stoffel, W., and Bosio, A.: Myelin glycolipids and their functions. Curr. Opin. Neurobiol. 7:654, 1997. 299. Stoll, G., Schwendemann, G., Heininger, K., et al.: Relation of clinical, serological, morphological and electrophysiological findings in galactocerebroside induced experimental allergic neuritis. J. Neurol. Neurosurg. Psychiatry 49:258, 1986. 300. Sumner, A.: Electrophysiological and morphological effects of the injection of Guillain-Barré sera in the sciatic nerve of the rat. Rev. Neurol. (Paris) 138:17, 1982. 301. Sumner, A. J., Saida, K., and Saida, T.: Acute conduction block associated with experimental antiserum-mediated demyelination of peripheral nerve. Ann. Neurol. 11:469, 1982. 302. Svennerholm, L.: Designation and schematic structure of gangliosides and allied glycosphingolipids. Proc. Brain Res. 101:xi, 1994. 303. Szabo, A., Dalmau, J., Manley, G., et al.: HuD, a paraneoplastic encephalomyelitis antigen, contains RNA-binding

Peripheral Nerve Antigens

304.

305.

306.

307.

308.

309.

310.

311. 312.

313.

314.

315.

316.

317.

318.

319.

320.

domains and is homologous to Elav and Sex-lethal. Cell 67:325, 1991. Takamiya, K., Yamamoto, A., Furukawa, K., et al.: Mice with disrupted GM2/GD2 synthase gene lack complex gangliosides but exhibit only subtle defects in their nervous system. Proc. Natl. Acad. Sci. U. S. A. 93:10662, 1996. Takashima, H., Boerkoel, C. F., De Jonghe P., et al.: Periaxin mutations cause a broad spectrum of demyelinating neuropathies. Ann. Neurol. 51:709, 2002. Tanaka, K., Tanaka M., Inuzuka, T., et al.: Cytotoxic T lymphocyte-mediated cell death in paraneoplastic sensory neuronopathy with anti-Hu antibody. J. Neurol. Sci. 163:159, 1999. Tanaka, M., Tanaka, K., Tsuji, S., et al.: Cytotoxic T cell activity against the peptide, AYRARALEL, from Yo protein of patients with the HLA A24 or B27 supertype and paraneoplastic cerebellar degeneration. J. Neurol. Sci. 188:61, 2001. Tatum, A. H.: Experimental paraprotein neuropathy: demyelination by passive transfer of human IgM antimyelin-associated glycoprotein. Ann. Neurol. 33:502, 1993. Taylor, V., Zgraggen, C., Naef, R., and Suter, U.: Membrane topology of peripheral myelin protein 22. J. Neurosci. Res. 62:15, 2000. Terryberry, J., Thor, G., and Peter, J. B.: Autoantibodies in neurodegenerative diseases: antigen-specific frequencies and intrathecal synthesis. Neurobiol. Aging 19:205, 1998. Tettamanti, G., and Riboni, L.: Gangliosides and modulation of the function of neural cells. Adv. Lipid Res. 25:235, 1993. Thomas, F. P., Adapon, P. H., Goldberg, G. P., et al.: Localization of neural epitopes that bind to IgM monoclonal autoantibodies (M-proteins) from two patients with motor neuron disease. J. Neuroimmunol. 21:31, 1989. Thomas, F. P., Lee, A. M., Romas, S. N., and Latov, N.: Monoclonal IgMs with anti-Gal(1–3)GalNAc activity in lower motor neuron disease: identification of glycoprotein antigens in neural tissue and cross-reactivity with serum immunoglobulins. J. Neuroimmunol. 23:167, 1989. Trapp, B. D.: Distribution of the myelin-associated glycoprotein and P0 protein during myelin compaction in quaking mouse peripheral nerve. J. Cell Biol. 107:675, 1988. Trapp, B. D., Andrews, S. B., Wong, A., et al.: Co-localization of the myelin-associated glycoprotein and the microfilament components, F-actin and spectrin, in Schwann cells of myelinated nerve fibers. J. Neurocytol. 18:47, 1989. Triantafilou, M., and Triantafilou, K.: Lipopolysaccharide recognition: CD14, TLRs and the LPS activation cluster. Trends Immunol. 23:301, 2002. Trojaborg, W., Hays, A. P., van den Berg, L., et al.: Motor conduction parameters in neuropathies associated with anti-MAG antibodies and other types of demyelinating and axonal neuropathies. Muscle Nerve 18:730, 1995. Tsiper, M. V., and Yurchenco, P. D.: Laminin assembles into separate basement membrane and fibrillar matrices in Schwann cells. J. Cell Sci. 115:1005, 2002. Underhill, D. M., and Ozinsky, A.: Toll-like receptors: key mediators of microbe detection. Curr. Opin. Immunol. 14:103, 2002. Vallat, J. M., Leboutet, M. J., Jauberteau, M. O., et al.: Widenings of the myelin lamellae in a typical Guillain-Barré syndrome. Muscle Nerve 17:378, 1994.

607

321. van Belkum, A., Van Den Braak, N., Godschalk, P., et al.: A Campylobacter jejuni gene associated with immunemediated neuropathy. Nat. Med. 7:752, 2001. 322. van der Meché, F. G. A., Visser, L. H., Jacobs, B. C., et al.: Guillain-Barre syndrome: multifactorial mechanisms versus defined subgroups. J. Infect. Dis. 176(Suppl. 2):S99, 1997. 323. van Koningsveld, R., Rico, R., Gerstenbluth, I., et al.: Gastroenteritis-associated Guillain-Barre syndrome on the Caribbean island Curacao. Neurology 56:1467, 2001. 324. van Koningsveld, R., van Doorn, P. A., Schmitz, P. I., et al.: Mild forms of Guillain-Barre syndrome in an epidemiologic survey in The Netherlands. Neurology 54:620, 2000. 325. van Muijen, G. N., Ruiter, D. J., van Leeuwen, C., et al.: Cytokeratin and neurofilament in lung carcinomas. Am. J. Pathol. 116:363, 1984. 326. Van Rhijn, I., Van den Berg, L. H., Bosboom, W. M., et al.: Expression of accessory molecules for T-cell activation in peripheral nerve of patients with CIDP and vasculitic neuropathy. Brain 123:2020, 2000. 327. Vimr, E., Lichtensteiger, C., and Steenbergen, S.: Sialic acid metabolism’s dual function in Haemophilus influenzae. Mol. Microbiol. 36:1113, 2000. 328. Visser, L. H., van der Meché, F. G. A., Meulstee, J., et al.: Cytomegalovirus infection and Guillain-Barre syndrome: the clinical, electrophysiologic, and prognostic features. Dutch Guillain-Barre Study Group [see comments]. Neurology 47:668, 1996. 329. Vriesendorp, F. J.: Insights into Campylobacter jejuniinduced Guillain-Barre syndrome from the Lewis rat model of experimental allergic neuritis. J. Infect. Dis. 176:S164, 1997. 330. Vyas, A. A., Patel, H. V., Fromholt, S. E., et al.: Gangliosides are functional nerve cell ligands for myelin-associated glycoprotein (MAG), an inhibitor of nerve regeneration [see comments.]. Proc. Natl. Acad. Sci. U. S. A. 99:8412, 2002. 331. Walker, R. I., Caldwell, M. B., Lee, E. C., et al.: Pathophysiology of Campylobacter enteritis. Microbiol. Rev. 50:81, 1986. 332. Walsh, F. S., Cronin, M., Koblar, S., et al.: Association between glycoconjugate antibodies and Campylobacter infection in patients with Guillain-Barre syndrome. J. Neuroimmunol. 34:43, 1991. 333. Wan, X. C. S., Trojanowski, J. Q., and Gonatas, J. O.: Cholera toxin and wheat germ agglutinin conjugates as neuroanatomical probes: their uptake and clearance, transganglionic and retrograde transport and sensitivity. Brain Res. 243:215, 1982. 334. Wang, X., and Tanaka Hall, T. M.: Structural basis for recognition of AU-rich element RNA by the HuD protein. Nat. Struct. Biol. 8:141, 2001. 335. Wassenaar, T. M.: Toxin production by Campylobacter spp. Clin. Microbiol. Rev. 10:466, 1997. 336. Wassenaar, T. M., Engelskirchen, M., Park, S., and Lastovica, A.: Differential uptake and killing potential of Campylobacter jejuni by human peripheral monocytes/macrophages. Med. Microbiol. Immunol. (Berl.) 186:139, 1997. 337. Weerth, S., Berger, T., Lassman, H., and Linington, C.: Encephalitogenic and neuritogenic T cell responses to the

608

338.

339. 340.

341.

342. 343.

344.

345.

346.

347.

348.

349.

350.

351. 352.

353.

354.

Neuroimmunology of the Peripheral Nervous System myelin-associated glycoprotein (MAG) in the Lewis rat. J. Neuroimmunol. 95:157, 1999. Weishaupt, A., Gold, R., Gaupp, S., et al.: Antigen therapy eliminates T cell inflammation by apoptosis: effective treatment of experimental autoimmune neuritis with recombinant myelin protein P2. Proc. Natl. Acad. Sci. U. S. A. 94:2164, 1997. Whittaker, V. P., and Kelic, S.: Cholinergic-specific glycoconjugates. Neurochem. Res. 20:1377, 1995. Wiethölter, H., Hulser, P.-J., Linington, C., et al.: Electrophysiological follow-up of experimental allergic neuritis mediated by a permanent T cell line in rats. J. Neurol. Sci. 83:1, 1988. Wigge, P., and McMahon, H. T.: The amphiphysin family of proteins and their role in endocytosis at the synapse. Trends Neurosci. 21:339, 1998. Wilkinson, P. C.: Serological findings in carcinomatous neuropathy. Lancet 1:1301, 1964. Willison, H. J., Almemar, A., Veitch, J., and Thrush, D.: Acute ataxic neuropathy with cross-reactive antibodies to GD1b and GD3 gangliosides. Neurology 44:2395, 1994. Willison, H. J., and Kennedy, P. G. E.: Gangliosides and bacterial toxins in Guillain-Barre-syndrome. J. Neuroimmunol. 46:105, 1993. Willison, H. J., and O’Hanlon, G. M.: The immunopathogenesis of Miller Fisher syndrome. J. Neuroimmunol. 100:3, 1999. Willison, H. J., and O’Hanlon, G. M.: Anti-glycosphingolipid antibodies and Guillain-Barre syndrome. In Nachamkin, I., and Blaser, M. J. (eds.): Campylobacter. Washington, DC, ASM Press, p. 259, 2000. Willison, H. J., O’Hanlon, G. M., Paterson, G. J., et al.: A somatically mutated human antiganglioside IgM antibody that induces experimental neuropathy in mice is encoded by the variable region heavy chain gene, V1–18. J. Clin. Invest. 97:1155, 1996. Willison, H. J., O’Leary, C. P., Veitch, J., et al.: The clinical and laboratory features of chronic sensory ataxic neuropathy with anti-disialosyl IgM antibodies. Brain 124:1968, 2001. Willison, H. J., Trapp, B. D., Bacher, J. D., et al.: Demyelination induced by intraneural injection of human anti MAG antibodies. Muscle Nerve 11:1169, 1988. Willison, H. J., and Veitch, J.: Immunoglobulin subclass distribution and binding characteristics of anti-GQ1b antibodies in Miller Fisher syndrome. J. Neuroimmunol. 50:159, 1994. Willison, H. J., and Yuki, N.: Peripheral neuropathies and anti-glycolipid antibodies. Brain 125:1, 2002. Winer, J. B., Hughes, R. A. C., Anderson, M. J., et al.: A prospective study of acute idiopathic neuropathy. II: Antecedent events. J. Neurol. Neurosurg. Psychiatry 51:613, 1988. Yako, K., Kusunoki, S., and Kanazawa, I.: Serum antibody against a peripheral nerve myelin ganglioside, LM1, in Guillain-Barre syndrome. J. Neurol. Sci. 168:85, 1999. Yamaji, T., Teranishi, T., Alphey, M. S., et al.: A small region of the natural killer cell receptor, Siglec-7, is responsible for

355.

356.

357.

358.

359.

360. 361.

362. 363.

364.

365.

366.

367.

368.

369.

its preferred binding to alpha 2,8-disialyl and branched alpha 2,6-sialyl residues: a comparison with Siglec-9. J. Biol. Chem. 277:6324, 2002. Yamawaki, M., Ariga, T., Bigbee, J. W., et al.: Generation and characterization of anti-sulfoglucuronosyl paragloboside monoclonal antibody NGR50 and its immunoreactivity with peripheral nerve. J. Neurosci. Res. 44:586, 1996. Yan, W. X., Archelos, J. J., Hartung, H. P., and Pollard, J. D.: P0 protein is a target antigen in chronic inflammatory demyelinating polyradiculoneuropathy. Ann. Neurol. 50:286, 2001. Yinchang, Y., Crawford, T. O., Griffin, J. W., et al.: Myelinassociated glycoprotein is a myelin signal that modulates the caliber of myelinated axons. J. Neurosci. 18:1953, 1998. Yu, R. K., and Ariga, T.: The role of glycosphingolipids in neurological disorders—mechanisms of immune action. Ann. N. Y. Acad. Sci. 845:285, 1998. Yu, Z., Kryzer, T. J., Griesmann, G. E., et al.: CRMP-5 neuronal autoantibody: marker of lung cancer and thymoma-related autoimmunity. Ann. Neurol. 49:146, 2001. Yuki, N.: Anti-ganglioside antibody and neuropathy: review of our research. J. Peripher. Nerv. Syst. 3:3, 1998. Yuki, N., Ho, T. W., Tagawa, Y., et al.: Autoantibodies to GM1b and GalNAc-GD1a: relationship to Campylobacter jejuni infection and acute motor axonal neuropathy in China. J. Neurol. Sci. 164:134, 1999. Yuki, N., and Tagawa, Y.: Acute cytomegalovirus infection and IgM anti-GM2 antibody. J. Neurol. Sci. 154:14, 1998. Yuki, N., Taki, T., and Handa, S.: Antibody to GalNAc-GD1a and GalNAc-GMlb in Guillain-Barre syndrome subsequent to Campylobacter jejuni enteritis. J. Neuroimmunol. 71:155, 1996. Yuki, N., Taki, T., Inagaki, F., et al.: A bacterium lipopolysaccharide that elicits Guillain-Barre syndrome has a GM1 ganglioside-like structure. J. Exp. Med. 178:1771, 1993. Yuki, N., Yamada, M., Koga, M., et al.: Animal model of axonal Guillain-Barre syndrome induced by sensitization with GM1 ganglioside [see comments]. Ann. Neurol. 49:712, 2001. Yuki, N., Yamada, M., Tagaawa, Y., and Takahashi, H.: Pathogenesis of the neurotoxicity caused by anti-GD2 antibody therapy. J. Neurol. Sci. 149:127, 1997. Yuki, N., Yoshino, H., Sato, S., and Miyatake, T.: Acute axonal polyneuropathy associated with anti-GM1 antibodies following Campylobacter enteritis [see comments]. Neurology 40:1900, 1990. Zhu, J., Pelidou, S. H., Deretzi, G., et al.: P0 glycoprotein peptides 56–71 and 180–199 dose-dependently induce acute and chronic experimental autoimmune neuritis in Lewis rats associated with epitope spreading. J. Neuroimmunol. 114:99, 2001. Zou, L. P., Ma, D.-H., Levi, M., et al.: Antigen-specific immunosuppression: nasal tolerance to P0 protein peptides for the prevention and treatment of experimental autoimmune neuritis in Lewis rats. J. Neuroimmunol. 94:109, 1999.

27 Experimental Autoimmune Neuritis RALF GOLD, GUIDO STOLL, BERND C. KIESEIER, HANS-PETER HARTUNG, AND KLAUS V. TOYKA

Immunologic Principles Categories of the Immune Response Tolerance and Autoimmunity The Local Immune Circuit The EAN Models Actively Induced EAN Adoptive Transfer EAN

Chronic (Relapsing) EAN: Mode of Immunization and Species-Specific Aspects Experimental Treatments in EAN Induction Phase: The Role of Antigen Presentation and Co-stimulation Transmigration and Early Effector Phase: Adhesion Molecules and MMPs

IMMUNOLOGIC PRINCIPLES Categories of the Immune Response The immune system is a multifaceted system of cells and molecules with specialized tasks in defending the organism from external agents. In addition, it plays a central role in maintaining antigenic homeostasis in the body. Two types of responses to invading organisms can occur: an acute response initiated within hours, mediated by the so-called innate immune system, and a delayed response occurring within days delivered by the adaptive or acquired immune system. The main distinctions between these two systems relate to the mechanisms and receptors used for immune recognition and to the maturational processing induced by somatic mutations of highly specific antigen receptors. Adaptive and innate immunity are functionally connected, allowing for intensive interactions (see Chapter 98).17,104 The Innate Immune System Innate immune responses consist of all the immune defense mechanisms that do not require antigen-specific immunologic memory. The innate immune system forms a fast first-line defense. The strategy of the innate immune response preferentially focuses on a few highly conserved structures present in a large variety of microorganisms. These structures are referred to as pathogen-associated molecular patterns, and the corresponding receptors of the

Effector Phase: T-Cell–Directed Treatments Effector Phase: Macrophage- and Antibody-Directed Treatments Termination of the Immune Response Survival Factors in the PNS Future Perspectives for Treatment

innate immune system are called pattern-recognition receptors.66 These are expressed on many effector cells of the innate immune system, most importantly on macrophages, dendritic cells, and B lymphocytes—collectively termed professional antigen-presenting cells (APCs). The total number of receptors involved in the innate immune response is thought to be on the order of hundreds, in contrast to the approximately 1015 somatically generated immunoglobulins and T-cell receptors (TCRs) of the adaptive (antigen-specific) immune response. Recent evidence shows that this recognition can mainly be attributed to the family of TOLL-like receptors (TLRs). Binding of pathogenassociated molecular patterns to TLRs induces the production of reactive oxygen and nitrogen intermediates and the proinflammatory cytokines, and upregulates expression of co-stimulatory molecules, subsequently initiating adaptive immunity.135 The Adaptive Immune System The adaptive immune response is based on two classes of highly specialized cells, T and B lymphocytes. Each of these cells usually expresses a single kind of a structurally unique receptor, resulting in a broad and extremely diverse repertoire of antigen recognition, usually with high affinity and avidity. Both B and T lymphocytes are derived from primordial stem cells in primary lymphoid tissues such as bone marrow 609

610

Neuroimmunology of the Peripheral Nervous System

and fetal liver. Their early phase of development is not dependent on the presence of any antigen, but once these cells express a mature antigen receptor, their further differentiation, survival, and programmed cell death (apoptosis) become antigen dependent.

Tolerance and Autoimmunity The random generation of a highly diverse repertoire of B- and T-cell receptors allows the adaptive immune system to recognize virtually any antigen. T cells recognizing autoantigens are termed self-reactive or autoreactive T cells, while B cells specific for autoantigens express autoantibodies on their surface. Tolerance is the process that eliminates (deletes) or downregulates such autoreactive cells. Consequently, a breakdown in this system can cause an autoimmune response or even autoimmune disease.73 B-Cell Tolerance The immune system has several mechanisms to filter autoreactive B lymphocytes out of the B-cell pool: (1) by clonal deletion of immature B cells in the bone marrow103; (2) by deletion of autoreactive B lymphocytes in the T-cell zones of secondary lymphoid organs, such as lymph nodes or the spleen110; (3) by induction of cellular anergy (“functional inactivation”); and (4) by a process called “receptor editing,” which changes the receptor specificity once an autoantigen has been encountered.103 At present it remains unclear to what extent these mechanisms are of relevance in maintaining tolerance and preventing autoimmune disorders in the human nervous system. T-Cell Tolerance The principle mechanism of T-cell tolerance is the deletion of self-reactive T lymphocytes during their development in the thymus.141 To delete all autoreactive cells, the presence of all autoantigens is required in the thymus. However, this is not the case.62,73,86 Some autoreactive T cells escape thymic education and enter the systemic immune compartment. In the mature immune system, several mechanisms are required to keep T cells in check and maintain peripheral tolerance, such as immunologic ignorance, peripheral deletion of specific T-cell clones, T-cell anergy, and suppression of T-cell activity. All these processes depend on the “transmission” of signals between T lymphocytes and adjacent APCs in addition to antigen recognition through a series of cell surface receptors that interact with their counterligands on the neighboring cell via cell-cell contact to provide co-stimulatory signals (see Experimental Treatments in EAN below). The site of cell-cell contact has also been termed the immunologic synapse. Breakdown of Tolerance If one of the regulatory mechanisms fails, a specific immune response is mounted against self-antigens. This

leads to expansion of autoreactive effector T cells and generation of autoantibodies through T-cell help, and may give rise to tissue damage, a scenario that Paul Ehrlich termed horror autotoxicus. Autoaggressive responses may initiate autoimmune disease. The autoimmune response may persist because the immune system is not able to remove the autoantigen from the body, in contrast to foreign antigens; even worse, new hitherto hidden autoantigens can be released to amplify the immune response and broaden its epitope specificity, a process termed epitope spreading.91 The effector mechanisms leading to tissue damage in autoimmune diseases are essentially the same as those operative in other inflammatory disorders. Various mechanisms are operative to promote or prevent tissue damage once an autoimmune reaction has been initiated. Recruitment of large numbers of monocytes/macrophages and T cells and the release of cytotoxic cytokines and chemokines would serve to augment the injurious tissue reaction. Autoantibodies are also active partners in tissue damage or dysfunction. Conversely, a suppressive/inhibitory local environment with only a few or not fully competent APCs and anti-inflammatory cytokines, inappropriate antibody concentrations, lack of complement activation, or merely anatomic barriers between autoreactive immune cells and target structures may all help to prevent autoimmune tissue destruction. Finally, induction of apoptosis in activated T cells may eliminate the autoreactive T cells before the full inflammatory cascade is set in motion.

The Local Immune Circuit The peripheral nervous system (PNS) has traditionally been considered as “immunologically privileged,” yet not as strictly as the central nervous system (CNS).152 This view has undergone revision within recent years, and now differences from other organs appear to be more quantitative than qualitative. The PNS is separated from the external environment by the blood-nerve barrier (BNB), which does restrict access of immune cells and soluble mediators to a certain degree; however, this restriction is not complete, either anatomically or functionally. The BNB is practically absent at nerve roots, in dorsal root ganglia, and at nerve terminals. Immune surveillance, as found in most organs, is present in the PNS as well: activated T and B lymphocytes can cross the BNB irrespective of their antigen specificity, and APCs, such as macrophages, can abundantly be detected in peripheral nerve tissue. Schwann cells also can function as “nonprofessional” APCs.153 Most of our present knowledge on the relevance of putative autoantigens and on the mechanisms involved in the pathogenesis of immune-mediated demyelination has been obtained in the animal model known as experimental autoimmune neuritis (EAN).

611

Experimental Autoimmune Neuritis

THE EAN MODELS Actively Induced EAN Mode of Immunization and Antigens Two decades after the description of experimental autoimmune encephalomyelitis (EAE), the first successful attempt to induce EAN was undertaken by Byron Waksman, who immunized rabbits with homogenized PNS tissue in adjuvant.145 Since then, investigators have studied EAN in various species using a panel of PNS autoantigens (see Chapter 26). As in EAE, many molecularly defined neuritogenic components of PNS myelin were identified, such as protein P2,71 the P2 peptides spanning the neuritogenic epitope (amino acids 61 to 70) of P2,106 protein P0,96 peripheral myelin protein 22 (PMP22),30 and myelin-associated glycoprotein (MAG).147 Susceptible animal species, including rabbit, monkey, rat, and mouse (Table 27–1), are inoculated subcutaneously with the neuritogen emulsified in incomplete Freund’s adjuvant

enriched with homogenized Mycobacterium tuberculosis (complete Freund’s adjuvant [CFA]). This mode of immunization was more specifically named actively induced EAN. The mycobacterial component of CFA with the included heat shock proteins may lead to severe arthritis,146 which is the basis of experimental arthritis models,155 but represents an unwanted side effect in EAN induction. Yet the use of CFA still represents the international standard in T-cell–mediated EAN and EAE models. Modern adjuvants such as Titermax (lacking mineral oil) are optimized mainly for B-cell responses and induce a weaker and more variable EAN (S. Jung and R. Gold, unpublished observations, 1998). In our experience in the rat, the following procedure minimizes unwanted side effects when using CFA: (1) preparation of the emulsion with 1 mg/mL or less final mycobacterial concentration, (2) subcutaneous injection of less than 50 ␮L of the immunogen at only one rat hind limb footpad, and (3) distribution of the remaining adjuvant at the tail base. In mouse models, injection at the footpad is avoided and the adjuvant is distributed along the flanks and the tail base. This strategy

Table 27–1. Animal Models of Immune-Mediated Neuropathies Animal Model Lewis Rat Active EAN

Adoptive transfer EAN

Antigen (First Description)

Comments

• PNS myelin, P2 (Kadlubowski and Hughes71) P0 (Linington et al.96) • ⫹ cyclosporine A (Pender et al.108) • PMP22 (Gabriel et al.30), MAG (Weerth et al.147) • P2 (Linington et al.95) P2 (aa 61–70) (Olee et al.106), P0 (aa 180–199) (Linington et al.96)

• Reliable model, commonly used • Chronic relapsing course, not robust • Only mild disease course • Homogeneous course, rapid onset, chronic relapsing after repeated T-cell transfer

Brown Norway rat

• P2 adoptive transfer (Linington et al.97)

Rabbits

• GalC (Saida et al.116) • Bovine brain ganglioside mixture, GM1 (Yuki et al.157)

• Chronic course • Axonal pathology

• P2 (Taylor and Hughes137) • P2 ⫹ PT ⫹ IL-12 (Calida15) • P0 aa 180–199 ⫹ PT ⫹ anti–CTLA-4 (Zhu et al.162)

• Mild disease • Severe course • Multimodal manipulation of immune system necessary

• MBP (18.5-, 21-KDa isoforms) (Abromsom-Leeman1)

• Peripheral and central demyelination

Knockout Mice Spontaneous inflammatory neuropathy

• NOD/B7-2 deficient (Salomon et al.117)

Superimposed neuropathy

• P0 ⫹/⫺ (Schmid120), CX32 ⫺/⫺ (Kobsar et al.83)

• Spontaneous chronic autoimmune peripheral neuropathy, mimics human CIDP • Macrophage mediated, secondary chronic immune response

Murine EAN Active SJL mice C57/B16 mice

Adoptive transfer Balb/C

aa ⫽ amino acids; CIDP ⫽ chronic inflammatory demyelinating neuropathy; CTLA-4 ⫽ cytotoxic T lymphocyte–associated antigen-4; CX32 ⫽ connexin 32; EAN ⫽ experimental autoimmune neuritis; GalC ⫽ galactocerebrosidase; IL-12 ⫽ interleukin-12; MAG ⫽ myelin-associated glycoprotein; MBP ⫽ myelin basic protein; PMP22 ⫽ peripheral myelin protein 22; PNS ⫽ peripheral nervous system; PT ⫽ pertussis toxin.

612

Neuroimmunology of the Peripheral Nervous System

guarantees a high efficacy of disease induction without producing concomitant severe arthritis. Ultimately these measures will also help to reduce the size of experimental groups. EAN in the Lewis rat is still the most commonly studied monophasic model because the disease course is predictable and relatively homogeneous. This allows for small group sizes of five to six rats in therapeutic or pathophysiologic studies. Typically, female rats at the age of 6 to 8 weeks are used for experimentation. Although the spectrum of available neuritogens has been expanded, immunization with bovine peripheral myelin (BPM) or the neuritogenic P2 peptide aa 53–7869 still gives the most reproducible results in the Lewis rat. Disease severity depends on the dose of the neuritogen inoculated, and severe disease may lead not only to demyelination, but also to axonal damage (Fig. 27–1). In this case, residual clinical signs such as gait ataxia may persist into remission. There are also differences in disease expression depending on the type and dosage of neuritogen. Whereas myelin immunization in the Lewis rat causes widespread perivascular demyelination resembling the pathology of Guillain-Barré syndrome (GBS) and chronic inflammatory demyelinating polyradiculoneuropathy (CIDP), with P2 peptide, wallerian degeneration indicative of axonal damage also is observed. With P0, and more so with PMP22 and MAG, the disease course is much milder at roughly equivalent molar doses. Clinical signs may even be lacking, and often inflammatory demyelination is only evident by histologic or electrophysiologic studies. In contrast to Lewis rats, the Brown Norway (BN) strain is resistant to active EAN. In principle EAN can also be studied in monkeys,26 yet primate models have been largely abandoned in EAN research. EAN has been induced in SJL mice137 and C57BL/6 mice,162 but the currently available models are characterized by rather mild disease in contrast to EAE. EAN in the mouse allows the use of genetically modified strains obtained by crossbreeding in various gene mutations. One example is knockout mice that lack P0,33 leading to dysmyelination and secondary autoimmunity.120 These mice were originally generated to better understand functional consequences of myelin deficiency in genetic models for Charcot-MarieTooth disease. By reverse transcription–polymerase chain reaction (RT-PCR) and Western blot, P0 was also found to be expressed in thymic epithelial cells.100,142 In P0 knockout mice on a C57BL /6 genetic background, tolerance to P0 is lacking and can be reconstituted by thymic transplants from wild-type mice100 or partly by generating bone marrow chimeras.142 Using the P0 knockout model, nontolerized or cryptic epitopes located in the extracellular P0 domain could be described, which may allow for development of better mouse EAN models in the future. All these models depend on immunization with a neuritogenic autoantigen. Recently, some mouse models have become available that develop spontaneous autoimmune diseases. Among these are genetically engineered mice or spon-

taneous mutants like the nonobese diabetic (NOD) mouse. In the NOD model, the co-stimulatory B7-1 and B7-2 molecules on immune cells have been shown to play distinct roles. Elimination of B7-2 expression by breeding NOD mice onto the B7-2–deficient background prevents diabetes, but leads to the development of a spontaneous autoimmune peripheral polyneuropathy (SAPP).118 Morphologic analysis revealed significant demyelination, with a mononuclear cellular infiltrate composed of dendritic cells, CD4⫹ T cells, and CD8⫹ T cells. Adoptive transfer of reactivated T cells from NOD mice with SAPP induced a severe peripheral neuropathy in the recipient crossbred NOD/severe combined immunodeficiency mice, thus formally fulfilling the traditional KochWitebsky’s postulates for an autoimmune pathogenesis of this mouse disease, as reviewed by Rose and Bona.113 Most EAN models feature combined demyelinating and axonal pathology. In rabbits, but not in rodent species, one recently described specific EAN model affects principally axons. On repetitive sensitization with a bovine brain ganglioside mixture, injected rabbits developed flaccid limb weakness of acute onset.157 Pathology showed predominant wallerian-like degeneration, with neither lymphocytic infiltration nor demyelination, indicating primarily axonal damage. The injected rabbits developed high titers of immunoglobulin G (IgG) antibodies to the monosialoganglioside GM1, and these antibodies were deposited on axons of the anterior roots and of peripheral nerves. Sensitization with purified GM1 also induced axonal neuropathy, indicating that GM1 was the likely immunogen in the ganglioside mixture. This model may help to clarify the molecular pathogenesis of axonal GBS and to develop appropriate treatment strategies. Disease Characteristics and Electrophysiology In EAN inflammation is multifocal, affecting nerve roots and peripheral nerves at multiple sites with, for unknown reasons, predominance in the lumbosacral region. Consequently, loss of tail tone occurs at disease onset. Disease progression is monitored by using a 10-point scale52,82 ranging from 0 ⫽ normal to 10 ⫽ death. Some investigators even prefer a 20-point EAN scale, which may be more sensitive in specific experimental settings.29 To obtain an objective scoring, the rater must be blinded as to the experimental groups. Grades 9 and 10 of 10 are usually not observed in EAN because, in contrast to EAE, the disease does not afflict the brainstem with its cardiovascular centers. All scoring is done on the same platform, which should be flat and not slippery. In case of painful adjuvant arthritis, scoring of gait is more difficult and prone to misclassification. In the Lewis rat, actively induced EAN typically starts around day 12 and then progresses for another 4 to 6 days to reach the peak of disease. At scores above 6, care must be taken that animals still have access to food and water. In full-blown EAN, concomitant weight loss of up to 15% of the initial weight is a common finding, probably reflecting

Experimental Autoimmune Neuritis

FIGURE 27–1 Antigen dose determines the severity of myelin-induced EAN. Lewis rats were immunized with increasing doses of bovine myelin in complete Freund’s adjuvant. Two representative cross sections taken at the S1 nerve root level are shown from a rat immunized with 3 mg of myelin (A) and from a rat immunized with 5 mg of myelin (B). Note predominant demyelination in a perivascular pattern in A versus widespread demyelination and admixed axonal changes in B. Bar: 20 ␮m. C, Representative electron micrograph with a demyelinated axon exhibiting mild degenerative changes adjacent to an axon with severe myelin disruption (long arrow). A ⫽ axon; M ⫽ mitochondria; SC ⫽ Schwann cell. (Electron micrograph provided by Dr. B. Schäfer.)

613

614

Neuroimmunology of the Peripheral Nervous System

cytokine and leptin stress responses similar to those in human disease.7 Rats with mild EAN may present without clinical signs, while histology clearly demonstrates inflammatory infiltrates in the PNS. Similar to studies in human patients, electrophysiologic studies can be performed under general anesthesia (neu-

roleptanalgesia) to define the functional consequences of PNS pathology. Thus a number of relevant nerve functions can be studied in serial electrophysiologic recordings (Fig. 27–2). As first signs of affliction, F-wave latencies become markedly prolonged and the responses are dispersed at disease onset.47,57 Progression of the disease

Day 0–3 Day 4

Time after transfer

15 ∞

18 ∞

1mV 2ms

21 ∞

24 ∞

Day 5 3 ∞

15 ∞ 0 2 4 6 8 10 ms

0 2 4 6 8 10 12 14 ms

Proximal

Distal

A

Time after transfer

5 weeks

50μV 5ms

6 weeks

100μV 5ms

7 weeks

500μV 2ms

9 weeks

1mV 2ms

10 weeks

1mV 2ms Proximal

Distal

B FIGURE 27–2 Representative recording of a serial nerve conduction study in AT-EAN. A, Time course of proximally and distally evoked compound muscle action potentials and F waves in a rat developing fulminant neuritis with complete paraplegia 5 days after injection of a high cell dose of P2-specific CD4⫹ T cells. B, Five weeks later, the first baseline deflections (compound muscle action potentials in statu nascendi) can be seen after distal stimulation, indicating earliest signs of regeneration. Over the following 10 weeks, functional restitution to virtually normal conduction occurred. (From Heininger, K., Stoll, G., Linington, C., et al.: Conduction failure and nerve conduction slowing in experimental allergic neuritis induced by P2-specific T-cell lines. Ann. Neurol. 19:44, 1986, with permission.)

615

Experimental Autoimmune Neuritis

is paralleled by slowing of motor nerve conduction velocity by up to 40%, decreased afferent nerve conduction velocity, and marked lowering of proximal and distal compound muscle action potential amplitudes. Conduction failure or conduction block may be observed in 80% to 100% of rats with EAN at disease maximum, and also F waves disappear in up to 80% of rats at this stage of the disease. Depending on disease severity and the degree of admixed axonal damage, these abnormalities may only partly improve during recovery, thus leading to irreversible electrophysiologic changes similar to those in patients with GBS.

revealed that they strip off myelin lamellae, induce vesicular disruption of the myelin sheath, and phagocytose damaged myelin.114 Interestingly, macrophages and not Schwann cells are the primary major histocompatibility complex (MHC) class II–expressing cell type in the inflamed PNS.121 Depending on the amount of myelin or neuritogenic peptide used for immunizations, and thus on the intensity of inflammation, an increasing degree of admixed axonal damage may occur.45,47,114 In particular, in T-cell line–mediated adoptive transfer (AT-) EAN (see below), the neuritis may take a fulminant course as a result of prominent endoneurial edema and ischemia that lead to severe axonal damage.

Pathology The inflammatory infiltrate in EAN is largely made up of lymphocytes and macrophages. In the rat, degranulated mast cells also are present. This leads to focal demyelination of nerves predominantly around venules. A semiquantitative and robust 4-point score has been proposed that facilitates quantification of inflammatory and demyelinating lesions in semithin sections52 (Fig. 27–3). As the inflammatory lesion proceeds, macrophages emerge as the predominant cells. These macrophages may be derived from blood-borne monocytes and endoneurial macrophages.76,102 Electron microscopic studies have

Cellular and Molecular Disease Mechanisms in EAN Entry of Inflammatory Cells: The Role of Cell Adhesion Molecules, Chemokines, and Matrix Metalloproteinases. In order to mediate a local immune response within the peripheral nerve, activated immunocompetent cells need to cross the BNB, an anatomically tight interface separating the systemic immune compartment from the PNS109 (Fig. 27–4). This mechanism of transendothelial migration is a multistep process occurring in an ordered sequence. In the first step, cellular adhesion molecules (CAMs) are expressed on leukocytes and vascular endothelium, resulting in a

FIGURE 27–3 Semiquantitative, categorical histologic score to assess inflammation and demyelination. Semithin sections are cut and stained with toluidine blue. At least 50 perivascular areas of cross sections are examined by a blinded observer as described by Hartung et al.52 A, Grade 0 ⫽ normal perivascular area. B, Grade 1 ⫽ mild cellular infiltration adjacent to a vessel. C, Grade 2 ⫽ cellular infiltrate plus demyelinating fibers adjacent to a vessel. D, Grade 3 ⫽ cellular infiltrate plus demyelination around a vessel and at more distant sites and minor axonal damage. (Magnification approximately ⫻450.)

A

B

C

D

616

Neuroimmunology of the Peripheral Nervous System

A

RBC 3

1 2

B

slowing and attachment of the circulating immune cells along the vessel wall. Normally, the flowing blood quickly dislodges cells that touch the vessel wall; thus adhesion molecules must act as mechanical anchors, but also function as tissue-specific recognition molecules. Based on structural differences, CAMs can be categorized into four groups: the immunoglobulin superfamily, selectins, integrins, and cadherins, all of which are involved in lymphocyte recruitment and extravasation.5 The specific function of individual CAMs has been elucidated by blocking their action with specific monoclonal antibodies in various animal models (see Transmigration and Early Effector Phase: Adhesion Molecules and MMPs below) or by generat-

FIGURE 27–4 Pathology at the blood-nerve barrier in EAN. Electron micrographs illustrating perivenular inflammation and endothelial morphology in rat sciatic nerve after immunization with the P2 aa 60–70 peptide of basic P2 protein. A, The perivenular interstitium contains many mononuclear cells and a single erythrocyte while an intravascular lymphoid cell is adherent to the adluminal endothelial surface. Intercellular tight junctions appear normal. (⫻7000.) B, Endothelial abnormalities in a vessel surrounded by inflammatory cells. Endothelial cell processes labeled 1, 2, and 3 exhibit intercellular dissociation with loss of tight junctions. A wide intercellular space has opened between endothelial cells 1 and 2, and a platelet overlies but does not block the space. Lower left inset, Electron-dense material appears to accumulate next to the basal lamina in the subendothelial space (arrowhead) (⫻14,600). Loss of tight junctions and endothelial separation appears between endothelial cells 2 and 3 (arrowhead). Lower right inset, Separation between endothelial cells 2 and 3 at higher magnification (⫻14,600). (From Powell, H. C., Olee, T., Brostoff, S. W., and Mizisin, A. P.: Comparative histopathology of experimental allergic neuritis induced with minimum length neuritogenic peptides by adoptive transfer with sensitized cells or direct sensitization. J. Neuropathol. Exp. Neurol. 50:658, 1991, with permission.)

ing knockout animals for the corresponding gene of a particular CAM. The extravasation of leukocytes into the CNS parenchyma is facilitated by the expression of CAMs and counterligands on both cerebral vascular endothelial cells and leukocytes, which strengthens their adhesive interaction.5 The initially unstable interaction provided by some CAMs is strengthened through the interaction of other CAMs, such as intracellular adhesion molecule-1 (ICAM-1) with lymphocyte function–associated antigen-1 (LFA-1), and vascular cell adhesion molecule-1 (VCAM-1) with very late antigen-4 (VLA-4), as demonstrated in EAN.3,24 Within the inflammatory process, expression of CAMs is upregulated, thus

Experimental Autoimmune Neuritis

augmenting attachment of immune cells. After interaction with the corresponding ligands, various CAMs are shed or cleaved from the cell surface.32 They circulate within body fluids and are thought to regulate cellular interactions and to promote de-adhesion. In a second step, chemokines come into play, providing signals to direct leukocyte migration into and within the extravascular space. Because lymphocytes must be positioned correctly to interact with other cells, the pattern and anatomic distribution of chemokines within the target tissue as well as the types of chemokine receptors expressed on the cell surface become critically important in orchestrating the ongoing immune responses.16,101 In EAN, a differential pattern of chemotactic signals with peak expression levels prior to or coincident with maximum clinical disease activity has been defined, involving different chemokines such as CXCL10 (IP-10), CCL2 (MCP-1), CCL3 (MIP-1␣), and CCL5 (RANTES).81 Corresponding chemokine receptors were detected in the inflamed PNS, with CCR1 and CCR5 primarily expressed by endoneurial macrophages, and CCR2, CCR4, and CXCR3 by invading T lymphocytes. Quantitative analysis revealed that CXCR3 is highest in infiltrating T cells compared to the other receptors. Its ligand CXCL10 mirrors the distribution of the cognate receptor within the inflamed PNS, and delineates endothelial cells as the primary cellular source of CXCL10, thus pointing to a pathogenic role for specific chemokine receptors and IP-10 in the generation of inflammatory demyelinating neuropathies.81 Finally, in a third step, matrix metalloproteinases (MMPs) are secreted by leukocytes in order to disrupt the BNB, thus facilitating transmigration of cells and of plasma-derived macromolecules (immunoglobulins and complement components) into the parenchyma of the nervous system. A differential expression pattern of various MMPs, specifically MMP-7 and MMP-9, can be depicted in the endoneurium and epineurium of the inflamed peripheral nerve (Fig. 27–5).79,92 The therapeutic potential of blocking chemokines and MMPs is described below (see Transmigration and Early Effector Phase: Adhesion Molecules and MMPs). Amplification and Termination of the Local Immune Response: The Role of Cytokines. Concomitant with cellular infiltration, a broad range of cytokines is induced in EAN nerves.34,77 Cytokines are mainly expressed by T cells and macrophages, but also to some extent by Schwann cells. At disease onset, T cells and a subpopulation of macrophages express interferon-␥ (IFN-␥), the prototype of a proinflammatory cytokine122 (Fig. 27–6). IFN-␥ is inducible by interleukin (IL)-18, a key mediator of innate immunity. In EAN, infiltrating macrophages strongly express IL-18. It is of note that patients with GBS may show increased IL-18 serum levels during the acute phase of their disease.65 IL-18 probably contributes to the

617

FIGURE 27–5 Immunohistochemistry for matrix metalloproteinases (MMPs) in the sciatic nerve of a Lewis rat with experimental autoimmune neuritis with clinically active disease. Arrowheads point to immunoreactivity for MMP-9, which can be localized around blood vessels, and to invading mononuclear cells, which can be identified as infiltrating T lymphocytes. Localization and distribution of the signal suggest that MMPs, such as MMP-9, are involved in the disruption of the blood-nerve barrier as well as in the process of migration of immunocompetent cells within the inflamed peripheral nerve. (Original magnification: ⫻200.) See Color Plate

amplification of the initial inflammatory response. In keeping with a key role in disease initiation, neutralization of IFN-␥ by antibodies attenuated clinical EAN53 (Fig. 27–7). IFN-␥ induces the expression of MHC class II molecules on monocytes/macrophages and dendritic cells, which are indispensable components of the trimolecular antigen complex in specific antigen recognition by autoreactive T cells. IFN-␥, moreover, activates macrophages to release tumor necrosis factor-␣ (TNF-␣) and other potentially myelinotoxic substrates. Indeed, in EAN macrophages adhering to nerve fibers expressed TNF-␣,130 thus providing the appropriate local milieu for myelin damage. Of note, in GBS patients, levels of circulating TNF-␣ correlated with electrophysiologic abnormalities.125 Injections of TNF-␣ into nerves may cause inflammation, demyelination, and damage to endothelial cells,111,115 although this mechanism has been debated by others.140 In further support of a detrimental role of TNF-␣ in EAN, experimental neutralization of TNF-␣ ameliorated demyelination.130 Seemingly in contradiction to this hypothesis, TNF-␣ also diminishes inflammation in nerves by inducing T-cell apoptosis. Apoptosis of T cells is an important mechanism for termination of immune responses (see below). Accordingly, neutralization of TNF-␣ by antibodies led to a prolonged persistence of T cells in

618

Neuroimmunology of the Peripheral Nervous System

FIGURE 27–6 Interferon-␥ (IFN-␥) immunoreactivity in EAN. Ventral roots of a rat with experimental autoimmune neuritis 12 days after active immunization. Cryosections of 1 ␮m thickness were labeled for IFN-␥ immunoreactivity by the DB-1 monoclonal antibody (A) and W3/13 antigen expressed on T cells and some neutrophils (B). Arrows denote cells staining positive for W3/13 and IFN-␥, respectively. ax ⫽ axon; my ⫽ myelin sheath; V ⫽ venule. Counterstaining was with hematoxylin. T cells, neutrophils, and macrophages are stained for IFN-␥. However, IFN-␥ was present in nerves only transiently between days 12 and 14 after immunization. Bar: 10 ␮m. (See Schmidt et al.122)

EAN nerves.151 At present, it appears that in EAN neutralization of TNF-␣–mediated myelin damage outweighs the ameliorating effects through T-cell removal, because animals treated with antibodies against TNF-␣ showed a modest overall clinical benefit.130 In addition to IL-18, IL-12 is a key cytokine of the innate immune system produced by type 1 T-helper (Th) cells and macrophages.139 IL-12 exerts proinflammatory actions as a p35/40 heterodimer, but is immunosuppressive in the form of its IL-12 p40 homodimers. IL-12 messenger RNA (mRNA)–positive cells have been described in EAN by in situ hybridization at onset of overt disease.160 However, IL12 p35 and IL-12 p40 have not been investigated separately. Another study using RT-PCR found no modification

of constitutive IL-12 p35 mRNA levels during EAN, but marked induction of IL-12 p40 mRNA that was delayed to the recovery phase.34 Functionally, IL-12 injected into healthy peripheral nerve provoked inflammation and caused marked demyelination.107 Thus it is conceivable that macrophage-derived IL-12 might promote inflammation at onset of EAN, while IL-12 p40 homodimers may have immunosuppressive effects during recovery. Cellular infiltration in EAN apparently has an intrinsic anti-inflammatory component in part mediated by downregulatory cytokines. IL-10, a potent downregulating Th2 cytokine, is already induced on a subpopulation of infiltrating macrophages at disease onset and parallels the initially strong expression of proinflammatory cytokines.34,64 When therapeutically administered from the start of immunization, IL-10 effectively suppressed and shortened clinical EAN. Even when given after day 12 postimmunization, after clinical EAN had been established, IL-10 also effectively suppressed the severity of EAN.6 Both prophylactic and therapeutic treatment with fusidin (4 mg/day intraperitoneally per rat) markedly ameliorated the clinical course of the disease compared to vehicle-treated animals. The beneficial effects were associated with profound modifications of the capacity of these rats to produce and release the aforementioned pro- and anti-inflammatory cytokines.21 Transforming growth factor-␤1 (TGF-␤1), another antiinflammatory cytokine, reached peak levels later, but shortly before clinical recovery.75 In support of a diseasemitigating effect, external administration of TGF-␤1 to EAN animals reduced inflammation and clinical severity.40 In conclusion, analysis of cytokine responses in autoimmune disorders such as EAN and GBS helped to disclose some of the regulatory mechanisms operative in immunemediated demyelination. Effector Mechanisms of Myelin Destruction. T-cell infiltration is the initial event in EAN. However, the mere presence of autoreactive T cells in nerves is not sufficient to induce clinical disease.54 Recruitment of macrophages is a crucial secondary step. The essential requirement of macrophages for disease induction in EAN has been highlighted by depletion experiments. Animals received intraperitoneal injections of silica quartz dust to divert monocytes to the peritoneal cavity. Macrophage depletion prevented all clinical, electrophysiologic, and histologic signs of EAN.19,52,58,136 There are two ways by which macrophages can destroy nerve tissue: by diffuse release of toxic factors or, more specifically, by adhering to and attacking the myelin sheath. MACROPHAGE-MEDIATED MECHANISMS: TOXIC FACTORS. Besides complement and cytokines, macrophages are a major source of toxic factors such as arachidonic acid metabolites, nitric oxide, oxygen radicals, and proteases. Their functional significance in disease progression has been

619

Experimental Autoimmune Neuritis

Day 0

Day 4

5mV 1ms

Day 8

SHAM

+anti-IFN γ

+IFN γ

FIGURE 27–7 Effect of IFN-␥ in EAN induced by adoptive transfer of P2-specific T-cell lines. Serial motor nerve conduction studies at the sciatic nerve. After T-cell transfer, animals received a monoclonal antibody to IFN-␥ (DB-1) on days 0, 3, and 7, leading to amelioration of disease activity. Conversely, rats treated with recombinant rat IFN-␥ showed more pronounced functional deficit. (From Hartung, H. P., Schäfer, B., Van der Meide, P. H., et al.: The role of interferon-gamma in the pathogenesis of experimental autoimmune disease of the peripheral nervous system. Ann. Neurol. 27:247, 1990, with permission.)

elucidated by pharmacologic blocking experiments (see Effector Phase: Macrophage- and Antibody-Directed Treatments below). Arachidonic acid metabolites, possibly synergizing with activated complement, could function as chemoattractants and secretagogues for inflammatory cells or could enhance vascular permeability and breach the BNB. Reactive oxygen species such as superoxide anion, hydrogen peroxide, and hydroxyl radicals could inflict peroxidation injury on myelin, but may also act by generating chemotactic signals and by exerting cytotoxic effects on endothelial cells.85 Macrophage-derived neutral proteases and phospholipases induce myelin damage in vitro. Accordingly, microinjection of a proteinase into rat sciatic nerve produced inflammatory demyelination, and treatment of EAN rats154 with proteinase inhibitors delayed development of clinical disease.119 B-CELL AND MACROPHAGE-MEDIATED MECHANISMS: ANTI-MYELIN ANTIBODIES AND COMPLEMENT. The hallmark of actively induced EAN and of the majority of human GBS cases is segmental demyelination that cannot be explained by diffuse release of toxic mediators alone. It is a basic morphologic observation that, in EAN and GBS, macrophages adhere to nerve fibers at an early stage of disease development and often normally appearing nerve

fibers are attacked.41,88 In contrast, after axotomy or nerve crush, macrophages selectively infiltrate nerve fibers undergoing wallerian degeneration, but they spare intact fibers.129 In EAN, macrophages strip off myelin lamellae from the axons, resulting in demyelination. The mechanisms by which initial T-cell infiltration leads to macrophage adherence on the surface of individual myelin sheaths are largely unknown. In analogy to human disease,11,156 it is likely that humoral autoantibodies to myelin components can play a role in the pathogenesis of EAN. These antibodies may block nerve excitation or neuromuscular transmission or may enhance T-cell–mediated demyelination by complement-mediated lysis and antibody-dependent cellular cytotoxicity effector mechanisms (see Chapter 26). Polyvalent antibodies to various components of the PNS have been detected in hyperimmune serum of rabbits and guinea pigs with EAN. However, these autoantibodies are unable to trigger demyelination without prior disruption of the BNB. To overcome this obstacle, the AT-EAN model can be used in co-transfer studies to examine the putative demyelinating activity of sera from patients with GBS or CIDP.42,54,138 Following an intravenous injection of a subneuritogenic cell dose, breakdown of the BNB is observed.43 Subsequent

620

Neuroimmunology of the Peripheral Nervous System

intravenous injection of purified IgG fractions from GBS/EAN then allows for examination of demyelinating activity by electrophysiology and/or histology (see above). In EAN increased antibody titers against peripheral myelin components could be measured in the blood stream.84,161 They can gain access to nerves by T-cell–mediated disruption of the BNB.126 Accordingly, a rapid increase in the passage of immunoglobulin into spinal roots, together with endoneurial infiltration of T lymphocytes and polymorphonuclear leukocytes, was demonstrated in AT-EAN. Accumulation of immunoglobulin was maximal during the worsening of neurologic deficit, and declined rapidly before the onset of neurologic recovery.43 In support of a decisive role of complement in the pathogenesis, decomplementation of animals with cobra venom factor partly suppressed EAN.27,144 Moreover, in myelin-induced EAN, macrophages were concentrated in areas of strong terminal complement complex (TCC) accumulation on myelin sheaths and Schwann cells131 (Fig. 27–8). Functionally, TCC deposition on the surface of myelin sheaths may promote an influx of calcium, which is considered an important initial step in myelin breakdown and cytotoxicity in general. Additionally, macrophages in culture respond to TCC by liberating chemotactically active and proinflammatory eicosanoids that have been found to contribute to the pathogenesis of EAN.47,52 Alternatively, TCC formation in myelin membranes may cause activation

of myelin-associated neutral proteinases with subsequent hydrolysis of myelin proteins. It is conceivable that TCCtargeted myelin is selectively attacked by macrophages. In support of this notion, Hafer-Macko and colleagues44 showed TCC deposits on the outer surface of myelin sheaths in nerve specimens from demyelinating GBS, while in axonal GBS, TCC was deposited axonally at paranodes. In both settings, macrophages adhered to nerve fibers at the location of TCC deposits. Schwann cells appear to escape TCC attacks, as shown in vitro by induction of cell cycle activation, proliferation, and rescue from apoptotic cell death.60 It was concluded from these in vitro studies that sublytic C5b-9 detected on Schwann cells in vivo during inflammatory neuropathies may even facilitate survival of Schwann cell to ensure remyelination. The Role of Mast Cells and Polymorphonuclear Leukocytes. Mast cells are part of inflammatory infiltrates in EAN,99 but they decrease in number in the course of EAN9 and degranulate possibly as a consequence of T-cell–mediated delayed-type hypersensitivity63; alternatively, their activation may be induced by myelin as shown in vitro. In turn, mast cell–derived proteases can degrade myelin in vitro. Mast cells are important sources of vasoactive amines (5-hydroxytryptamine, histamine) and arachidonic acid metabolites that both could augment vascular permeability and thereby induce a leaky BNB. In support of a pathogenic role, mast cell–stabilizing drugs such as reserpine and nedocromil prevented or attenuated the disease.10 Polymorphonuclear leukocytes that have an enormous proinflammatory potential can be detected in EAN lesions, but their pathophysiologic role is not well defined.43,88,122

Adoptive Transfer EAN

FIGURE 27–8 Complement deposition in EAN. Ventral root section from an animal with EAN 11 days after immunization (i.e., before overt clinical disease), immunostained for the terminal complement complex (C5b-9). The reaction product is localized on the surface of myelin sheaths (arrows) and on myelinating Schwann cells (asterisk). Staining appears preferentially around a venule (v), while other fibers are spared. Terminal complement complex is deposited only transiently for a few days. Bar: 10 ␮m. (See Stoll et al.131)

AT-EAN and Disease Course The generation of antigen-specific autoaggressive T-cell lines was an important step toward a better understanding of experimental paradigms in EAN. AT-EAN has also provided conclusive proof of the pivotal pathogenic role of T cells in experimental neuritis, thus fulfilling the Koch-Witebsky criteria for EAN as an (experimental) autoimmune disorder.113 Following the classic methods developed for EAE,8 T-cell lines can be established in vitro from lymph node cells of rats immunized with P2.95,96 These primary cell cultures are propagated by repetitive, antigen-specific activation cycles in turn with IL-2–mediated expansion to generate T-cell lines. Also, T-cell clones can be generated from immunized Lewis rats.89 By intravenous transfer of defined numbers of T lymphoblasts, ranging from 1 to 20 million cells, monophasic AT-EAN is induced. In analogy to actively induced EAN, T-cell lines

Experimental Autoimmune Neuritis

reactive with P2 and P0 were first found to transfer the disease.96 Epitope specificity of autoreactive T-cell lines that can produce disease has been examined and found to reside primarily in the amino acid sequences 61 to 70 of P2105 or 180 to 199 of P0.96 AT-EAN of the Lewis rat begins within 3 to 4 days after cell injection and typically achieves its maximum between days 6 and 10: High T-cell numbers cause earlier and more severe disease than do lower numbers.59 Disease incidence reaches up to 100% going in parallel with a very low interindividual variation of clinical disease; this is ideal to study smaller groups of animals, which is of value in therapeutic settings with restricted access to precious reagents. Although BN rats have a very low susceptibility to overt EAN induced by active immunization, T-cell lines specifically reactive to bovine P2 could be isolated from these normal-looking rats. Upon transfer to syngeneic BN recipients, these T cells were capable of evoking EAN.97 This indicates that autoreactive T-cell clones are also contained in the T-cell repertoire of this nonresponder strain. Non-neural antigens such as ovalbumin are capable of inducing disease by using another paradigm of generating AT-EAN. In a first step, T-cell lines are generated that react specifically to ovalbumin, a foreign (nonself) protein in the rat. Following intraneural microinjection of ovalbumin into one sciatic nerve, ovalbumin-specific T-cell lines were transferred that led to EAN-like inflammatory lesions in the nerve injected with ovalbumin, but not at the contralateral side that received casein as control antigen.54 These observations allow us to conclude that, in principle, any foreign protein achieving access to the PNS may become subject to immune surveillance. Applied to the situation in GBS, it is conceivable that preceding infections with bacteria or viruses may initiate the autoaggressive assault once a specific T-cell repertoire has been generated and antigens of the infectious agent are deposited in nerve tissue. Alternatively, an immune reaction may occur with infectious agents such as Campylobacter jejuni that show molecular similarities with nerve antigens (molecular mimicry).23 Pathology of AT-EAN In contrast to active EAN, the T-cell response is predominant in AT-EAN, with a negligible contribution of B-cell–derived autoantibodies. Macrophage-derived effector mechanisms are also operative in AT-EAN.58,63 Usually, pathogenic changes are initiated by vasogenic edema resulting from perivascular influx of autoaggressive lymphocytes, followed by increased endothelial permeability to serum proteins. In full-blown EAN, the BNB breaks down completely and allows rapid immigration of numerous inflammatory cells.58,63 Thus endoneurial fluid pressure and edema increase rapidly, adding superimposed ischemichypoxic damage. This is associated with rapid deterioration

621

of nerve conduction up to complete and long-lasting conduction failure in severe disease, indicating massive destruction of axons. The electrophysiologic findings resemble hyperacute GBS.59

Chronic (Relapsing) EAN: Mode of Immunization and Species-Specific Aspects Chronic EAN has been developed as a model for human CIDP. The only myelin antigen for which there is clear evidence for a purely demyelinating response is galactocerebroside (GalC; see Chapter 26). Rabbits immunized with GalC develop a slowly progressive disease starting with trembling and then progressing to tetraplegia (Fig. 27–9).116,132 High titers of circulating antibodies against GalC could be detected.132 Intraneural injections of antiserum from afflicted rabbits with high titers of these antibodies caused demyelination in recipient rat sciatic nerve in vivo87 and also could be seen in vitro in myelin preparations. This is remarkable because the antiserum recognizes epitopes involving the galactose moiety of the glycolipid and does not bind to micellar dispersions of the purified lipid. Surprisingly and still unexplained, purified IgG fractions containing high-titer anti-GalC antibodies do not cause demyelination upon passive transfer, nor can they inhibit normal (re-) myelination after experimental nerve damage in rabbits with ongoing GalC-EAN.128,138 In the absence of complement, binding of the anti-GalC serum to myelin causes swelling and separation of compacted myelin lamellae. The usefulness of GalC-EAN as a disease model is limited because the rabbit is the only species producing overt disease, while mice and rats are poor responders to an immune challenge with GalC. Chronic relapsing EAN can be induced in the Lewis rat by treatment with low-dose cyclosporine A after active immunization with myelin.108 First, cyclosporine A prevents the development of EAN. Upon its withdrawal, chronic relapsing EAN with severe inflammation and demyelination of nerve roots, ganglia, and spinal nerves is observed. This chronic EAN also goes along with signs of remyelination, yet this model is not very robust. DA rats show mild chronic EAN upon immunization with BPM, which makes this model more suitable for testing new strategies of therapeutic intervention. Repeated adoptive transfer of P2-specific T-cell lines has also been used to produce chronic relapsing EAN resembling relapsing GBS but not CIDP.90 It is plausible that in chronic EAN, similar to the situation in EAE, chronic and ongoing destruction of myelin causes sensitization to other components of the myelin sheath as well. This paradigm of “epitope spreading” as shown in EAE91 has not conclusively been demonstrated in EAN (see Table 27–1).

622

Neuroimmunology of the Peripheral Nervous System

Pre-immunization 2mV 2ms 4 months 10 months 18 months

C

Proximal

Distal

FIGURE 27–9 Galactocerebroside-induced chronic EAN in the rabbit. A, Dorsal root of a rabbit 20 weeks after primary immunization with galactocerebroside, methylated bovine serum albumin, and complete Freund’s adjuvant. Numerous axons (A) close to the venule (v) are completely demyelinated (arrows), and macrophages abound around the vessel. B, Electron micrograph from the same animal showing a macrophage with an activated nucleus (N) loaded with myelin debris and encircling demyelinated axons (A), one of which is enveloped by its Schwann cell cytoplasm. C, Representative example of a serial nerve conduction study. Compound muscle action potentials evoked after stimulation of the sciatic nerve at proximal and distal sites are recorded over the small foot muscles. Note the parallel decrease of potential amplitudes after proximal and distal stimulation at months 4 and 10 and the prolonged latencies of the M and F responses. After 18 months, following clinical recovery, nerve conduction velocities reached nearly preimmunization values. (From Stoll, G., Schwendemann, G., Heininger, K., et al.: Relation of clinical, serological, morphological, and electrophysiological findings in galactocerebrosideinduced experimental allergic neuritis. J. Neurol. Neurosurg. Psychiatry 49:258, 1986, with permission.)

623

Experimental Autoimmune Neuritis

result of downregulation of effector molecules and T-cell apoptosis (Table 27–2 and Fig. 27–10).

EXPERIMENTAL TREATMENTS IN EAN The immune response in EAN can be divided into three phases: induction, amplification, and effector phases. In the induction phase, the injected autoantigen is presented to “naive” T cells by professional APCs such as macrophages or dendritic cells, resulting in T-cell activation (see Immunologic Principles above). Two external signals are crucially required for effective T-cell activation by antigen presentation: the antigen-specific signal provided by the immunogenic peptide and presented in the context of MHC molecules on APCs, and the antigen-independent signal called co-stimulation (see Chapter 98). Co-stimulation is mediated by adhesion molecules of the integrin family, such as ␣L␤2 (LFA-1, CD11a/CD18) and ␣4␤1 (VLA-4, CD49d/CD29), and by adhesion molecules of the immunoglobulin superfamily, such as ICAM-1 (CD54), VCAM-1 (CD106),5 and, of special functional relevance, cytotoxic T lymphocyte–associated antigen-4 (CD152), B7-1 (CD80), and B7-2 (CD86), expressed on both T cells and APCs117 (see Chapter 26). Activated T cells then circulate in the blood, attach to the venular endothelium in the PNS, and penetrate the BNB through activation of MMPs.35 This transendothelial migration gives rise to the effector phase of the immune response in EAN. In the PNS, the autoantigen is then presented to T cells by macrophages. At the same time, the reactivated CD4⫹ T cells augment the immune response by recruiting further T cells and macrophages via chemokines and cytokines (amplification phase). The resulting breakdown of the BNB then allows the passage of circulating autoantibodies and complement components (a complementary pathway of the effector phase). Finally, the termination of the ongoing immune response is the

Induction Phase: The Role of Antigen Presentation and Co-stimulation The first important step in the pathogenesis of EAN is the presentation of autoantigen in the induction phase. This results in the physiologic activation of disease-inducing T cells. Several tools are available to modulate immunoregulatory mechanisms. Using the paradigm of myelinspecific oral tolerance,148 Gaupp et al. have shown that active EAN can be prevented or ongoing disease be ameliorated by feeding with oral myelin.31 This effect can be further modulated by adjuvants such as cholera toxin.67 Active suppression of effector T cells rather than generation of Th2-type T cells was identified as the putative underlying mechanism of oral tolerance. Therapy targeted against the ␣/␤ TCR by using the monoclonal antibody (mAb) R73 showed therapeutic and preventive potential in different EAN models.69 The underlying molecular mechanism was impairment of antigen recognition and T-cell function by TCR occupancy. Because the frequency of TCR-positive cells was transiently reduced by only 50% and returned to normal within 10 days as shown by fluorescence-activated cell sorter analysis, TCR-mediated T-cell downregulation rather than full depletion of lymphocytes appears to be decisive. A more selective way to inhibit proliferating T cells is aimed at blockade of the IL-2 receptor, which is expressed on this subset of activated T cells. In AT-EAN, prophylactic treatment with the mAb ART18, directed at the IL-2 receptor, mitigated disease severity when given

Table 27–2. Recent Experimental Treatments of Immune-Mediated Neuropathies Experimental Strategy

Target

Blockade of adhesion molecules

• • • • • • • • • • • • • • • •

Tissue invasion Induction of T-cell apoptosis Cytokine modulation Induction of T-cell anergy

Induction of Th2 response or regulatory T cells Mitigation of macrophage function

ICAM-1/LFA-1 VCAM-1/VLA-4 L-selectin Matrix metalloproteinases Antigen therapy Steroid pulse therapy IL-2 receptor IL-10 T-cell receptor ␣/␤ CD2 molecule T-cell activation Anti-CD28 Oral tolerance Cyclooxygenase inhibitors Oxygen radical scavengers Liposome-encapsulated dichlormethylene

Reference 4 24 2 61 150, 158 6, 50 69 70 127 122, 134 31 46, 51 68

624

Neuroimmunology of the Peripheral Nervous System

VENULE

PERIPHERAL NERVE

APC

BNB

MΦ T

TNF OH¯ PGE

Res. MΦ

T

TH

Axon B

C6

CAMs T

Chemokines T

Abs

MMPs T

PLT

NO TNF MMPs

LT TNF

C5b-9

MΦ T SC

CD40L MΦ

LTC4 MMPs ROS OH¯ IL-1

Apoptosis

MC

FIGURE 27–10 Simplified scheme depicting the hypothetical sequence of major immune mechanisms in EAN (see sections on Immunologic Principles and The EAN Models in text). Circulating lymphocytes are activated by an as yet unknown antigen (e.g., P2 or P0 in EAN). Activated lymphocytes adhere to the endothelial layer of intraneural venules and migrate intraneurally. Adhesion, rolling, and migration are governed by a number of cell adhesion molecules. Intraneurally, macrophages are activated and produce cytokines, reactive oxygen species, prostaglandins, and complement factors, which in turn activate further inflammatory cells that are resident or have meanwhile transmigrated through an open bloodnerve barrier. Here, metalloproteinases, leukotrienes, and interleukin-1 play a role. Moreover, mast cells may also induce increased vascular permeability and inflammatory reactions. Platelets (PLT) may stimulate antigen-presenting cells (APC) through the CD40 ligand pathway. Activated macrophages may also function as antigen-presenting cells and activate specific and bystander T-cells. Moreover, macrophages may directly damage myelin sheaths and axons by cytotoxic mechanisms including oxygen radicals. Transmigrated T-helper cells activate transmigrated B cells to produce myelin-specific antibodies that can attach to the myelin sheath or to the axon and damage nerve fibers with the mediation of complement. Macrophages may also be targeted to nerve fibers after being armed with circulating antibodies (not shown) following the mechanism of antibody-dependent cellular cytotoxicity. Finally, antibodies may directly block axonal excitation at the nerve terminal. It is not clear whether Schwann cells per se are also attacked by antibodies and T cells, and some may undergo programmed cell death (apoptosis). The inflammatory reaction can be terminated by T-cell apoptosis induced by repeated antigenic stimulation and by the Fas/Fas-ligand system, in which Schwann cells may also play an important part. Inset, Electron micrograph of an apoptotic Schwann cell. T-cells may also be downregulated by the cytokine TGF-␤ (not shown). Abs ⫽ antibodies; APC ⫽ antigen-presenting cell; BNB ⫽ blood-nerve barrier; CAM ⫽ cellular adhesion molecule; C ⫽ complement components; IL ⫽ interleukin; LTC ⫽ leukotriene; MC ⫽ mast cell; MMP ⫽ matrix metalloproteinases; M⌽ ⫽ macrophage; NO ⫽ nitric oxide; OH ⫽ hydroxyl radicals; PGE ⫽ prostaglandin E; PLT ⫽ platelet; SC ⫽ Schwann cell; Th ⫽ T-helper cell; TNF ⫽ tumor necrosis factor. See Color Plate

Experimental Autoimmune Neuritis

625

during the latent induction phase but not after onset of clinical symptoms.50 The most potent co-stimulatory signal known to date is mediated by the T-cell surface receptor CD28 interacting with CD80 (B7-1) or CD86 (B7-2) on APCs such as macrophages or dendritic cells. Normally, a CD28 signal alone does not lead to activation of resting T cells. In apparent contradiction to this paradigm, CD28-specific mAbs have been identified recently in several experimental models that induce proliferation and/or cytokine secretion of T cells in the absence of TCR engagement. In the rat system, the mAb JJ316 activates all resting T cells to proliferate without a TCR signal in vitro and in vivo133 and is therefore called “superagonistic.” In keeping with the Th2-promoting effect of CD28 signals in co-stimulation, it was shown that T-cell activation with the mAb JJ316 primed CD4⫹ T cells for Th2 differentiation in vitro, and induced IL-4 and IL-10 expression along with Th2dependent immunoglobulin isotypes in vivo. Treatment with JJ316 during the induction phase of active EAN and AT-EAN dramatically reduced disease severity and improved nerve function as revealed by electrophysiology. In addition, JJ316 given 1 week before immunization had a preventive effect.123 Morphologically, JJ316 markedly reduced T-cell infiltration of the sciatic nerve without induction of apoptosis. These results suggest that the superagonistic mAb preferentially activates regulatory (formerly called suppressor) T cells, thus providing a model for novel antibody-mediated treatments as an alternative to conventional immunosuppressive strategies.

ameliorating EAN24; this integrin seems to be the adhesion molecule that is most important in transendothelial migration of T cells in rodent EAN and hence the most promising target candidate for future therapeutic intervention in GBS. Surprisingly, blockade of VLA-4 and VCAM-1 blockade are also effective in enhancing T-cell apoptosis in the inflammatory lesion (see Termination of the Immune Response below).93 In vivo, inhibition of other coaccessory molecules such as CD2 turned out also to be effective in various EAN models.70 Since the anti-CD2 antibody did not exert its effect by inhibition of T-cell activation, induction of anergy, or depletion of T cells, the most probable mechanism is thought to be impairment of T-cell migration across the BNB. Transmigration of T cells through the BNB is mediated by upregulation of a repertoire of enzymes. Among these are MMPs, a family of calcium-dependent zinc endopeptidases that may be implicated in the pathogenesis of inflammatory demyelinating disorders.48 MMPs are also involved in the processing of TNF-␣. Among the MMPs, MMP-7 and MMP-9 are upregulated in inflamed muscle or nerve specimens in human nerve biopsies and in EAN nerves,18,20,78,80 as shown by immunocytochemistry, RT-PCR, and zymography. Indeed, the inhibition of these proinflammatory enzymes by the broad-spectrum MMP inhibitor BB-1101 exerted a strong preventive and also therapeutic effect in EAN.112

Transmigration and Early Effector Phase: Adhesion Molecules and MMPs

By transmigration and damage to the separating endothelial barrier, T cells breach the way for the passage of other humoral and cellular components of the inflammatory cascade from the blood stream to the nerve parenchyme. Intraneurally, CD4- or CD8-positive T cells meet their target antigen presented by MHC class II or class I molecules on infiltrating or resident macrophages102,121 or on nonprofessional APCs such as Schwann cells37 and support the amplification of the immune response. Recently, antigen-specific therapy has been developed as a novel tool for the treatment of experimental autoimmune disorders (Fig. 27–11). This is based on the observation that TCR reengagement at an appropriate stage of the T-cell cycle eventually leads to T-cell apoptosis by activationinduced cell death.94 In EAN, both actively induced and adoptively transferred, the intravenous administration of recombinant P2 prevented and ameliorated disease in the Lewis rat, and this was associated with a profound increase in T-cell apoptosis in peripheral immune organs as well as in the sciatic nerve.150 TNF-␣ has been identified as a principal mediator of the beneficial effects of antigen therapy in situ.151 The intracellular signaling mechanisms that are associated with high-dose antigen therapy have not yet been elucidated.

More than a decade ago it was shown that activated T lymphocytes are capable of traversing endothelial barriers that separate the nervous system from the blood stream.152 CAMs are essential for effective antigen presentation and also mediate cell-cell communication in vivo and in vitro in rodents and humans. Interaction of immune cells or extracellular matrix via these molecules profoundly influences a variety of biologic functions, including homing and transmigration of immune cells through the BNB. In the acute phase of EAN, upregulation of CAMs such as ICAM-1 and VCAM-1 at the endothelial tight junctions of lesion-associated blood vessels (i.e., part of the BNB) parallels disease activity and parenchymal infiltration.24 This points to a common pathogenic role of adhesion molecules in the initiation of tissue inflammation in EAN. This notion was reinforced by therapeutic manipulation with mAbs directed at these molecules in vivo. First, the involvement of ICAM-1 in transendothelial migration was delineated in EAN,3 followed by that of LFA-14 and L-selectin.2 This was later extended to VLA-4 and its ligand VCAM-1. Blockade of VLA-4 (␣4␤1) by mAbs in vivo was effective in

Effector Phase: T-Cell–Directed Treatments

626

Neuroimmunology of the Peripheral Nervous System

DISEASE COURSE

WEIGHT

10 200

9

6 5 4 3 2 1

160 140 120

Daily i.v. treatment

Daily i.v. treatment

100

0 0

A

180

Injection

Mean weight ± SD

7

Injection

Mean score ± SD

8

10

0

20

B

Time (days)

10

20

Time (days)

Ovalbumin control group 2x 500 μg recombinant P2 FIGURE 27–11 Preventive antigen-specific therapy of AT-EAN. In groups of six rats, AT-EAN was induced by intravenous transfer of neuritogenic P2-specific T cells. Starting on day 1, rats received either the specific antigen P2 (filled squares) or ovalbumin as control antigen (filled circles) by daily intravenous injection. Graphs show the typical disease course (A) and body weight (B); note that in B the y axis starts at 100 g. In the control group, disease started at day 3, reaching its maximum at day 7 with paraparesis (A) and 15% weight loss (B). In contrast, the group receiving high-dose antigen was completely protected. See Color Plate

To further modulate the P2-directed immune response in the antigen-specific treatment paradigm, a polypeptide oligomer (P2-16mer) harboring 16 repeats of the neuritogenic epitope (aa 58 to 73) of the P2 protein interspersed by peptides as spacers was designed and synthesized. T-cell epitope oligomers of this type have been shown to be particularly effective in the specific suppression of EAE by the mechanism of inducing high-zone tolerance. In contrast to other broadly immunosuppressive mechanisms, this new therapeutic approach for prevention and therapy of autoimmune diseases aimed more specifically at the antigenspecific elimination of autoreactive T cells. Not only were these “designer polypeptides” useful for prevention and therapy but, more importantly, a tolerizing effect was achieved (labeled “vaccination”) that lasted at least 4 months.127,151

Effector Phase: Macrophage- and Antibody-Directed Treatments In parallel with local T-cell activation, secretion of cytokines upregulates local chemokine expression, which leads to attraction and local activation of monocytic cells, thus resulting in amplification of the immune response52 (see Immunologic Principles above; see also Chapter 98).

Macrophages, the principal APCs in the PNS, are armed with a repertoire of secreted proinflammatory mediators, including free reactive oxygen metabolites (“oxygen radicals”). They also release free NO, which may further augment tissue damage and block impulse propagation, because NO radicals were shown to interfere with conduction properties of nerve fibers and may induce permanent nerve damage if applied ex vivo.112 In addition, macrophages bind via Fc receptors to sites of immunoglobulin and complement deposition. Thus macrophages not only serve as phagocytic cells to clear tissue debris, but in addition receive stimulatory signals that augment their proinflammatory function and thereby initiate a vicious circle of immune activation. These effector functions of macrophages can be inhibited by administration of steroids, cyclooxygenase inhibitors, or eicosanoids, ameliorating the disease course of EAN.49,51,52 (Fig. 27–12). Thus macrophage-derived proinflammatory arachidonic acid metabolites significantly contribute to functional and tissue damage in EAN. To more selectively eliminate phagocytic cells, liposomeencapsulated dichlormethylene diphosphate was studied in active EAN and AT-EAN.68 Efficient suppression of active EAN and of AT-EAN was observed, thus confirming the decisive role of macrophages as effector cells during EAN. Similar to experimental studies in EAE,124

627

Experimental Autoimmune Neuritis FIGURE 27–12 Serial motor nerve conduction studies in EAN: therapeutic effects of oxygen radical scavengers. Representative redrawn original recording of compound muscle action potentials obtained from the small foot muscles after electrical stimulation of the sciatic nerve distally (above the ankle; upper trace) and proximally (sciatic notch; lower trace). Lewis rats were immunized with 3 mg of myelin and treated with either catalase or superoxide dismutase (SOD) twice daily from day 7 after immunization until sacrifice. F-wave failure, conduction block, and conduction slowing evolve over time in shamtreated EAN rats, designated “myelin,” whereas only minor changes are seen in enzyme-treated animals. Arrows indicate stimulation artifacts. (From Hartung, H. P., Schäfer, B., Heininger, K., and Toyka, K. V.: Suppression of experimental autoimmune neuritis by the oxygen radical scavengers superoxide dismutase and catalase. Ann. Neurol. 23:453, 1988, with permission.)

Myelin

+ Catalase

+ Sod

Day 1

Day 13

Day 21

liposome-encapsulated glucocorticosteroids may also be of superior therapeutic efficacy in EAN because they specifically target inflammatory phagocytes and were able to reduce macrophage infiltration. Along with other proinflammatory mediators, NO may also be involved in orchestrating the local immune reaction. Interestingly, not only T cells but also glial cells as facultative APCs can produce NO in response to proinflammatory cytokines.38 Because NO seems to be involved in a complex network of pro- and anti-inflammatory signaling, it is not surprising that the net effect of NO synthase inhibition is often unpredictable and such inhibition has indeed been shown to ameliorate active EAN.72,163 As important players in the effector phase, antibodies and complement are other targets of experimental treatments. Plasma exchange, which is an established treatment in human inflammatory neuropathies, has also been shown to be therapeutically effective in EAN.56 Similar findings were obtained by complement depletion via cobra venom factor, which reduced inflammation and demyelination in EAN.143 Treatment of ongoing EAN was also achieved through neutralization of hyperimmune EAN serum containing polyvalent antibodies by plasma infusions from healthy animals,55 whereas in rat EAN, human intravenous immunoglobulin was only marginally 29 or not at all25 effective. In GBS, serum IgG antibodies directed to gangliosides (see Chapter 26) have been identified that functionally and sometimes irreversibly block neuromuscular transmission ex vivo.13 With EAN serum, this block is fulminant and irreversible.14 This blockade can be neutralized by polyvalent human IgG and its Fab fraction.12

Termination of the Immune Response The termination of ongoing PNS inflammation can be induced by continuous downregulation of the vicious circle

5mV 1ms

of the amplification-effector pathways. One such method is silencing of macrophages (see Effector Phase: Macrophageand Antibody-Directed Treatments above); another is apoptosis of autoimmune T cells, which occurs during the natural disease course of EAN.159 Glucocorticosteroids are among the most efficient antiinflammatory drugs and can eradicate infiltrating T cells through augmenting apoptosis (Fig. 27–13). Based on the use of high-dose glucocorticosteroids in human disease, steroid “pulse therapy” (10 mg/kg body weight) in rat EAN led to a four- to fivefold increase of T-cell apoptosis in the inflamed sciatic nerve.158 It is assumed that, at high doses, glucocorticosteroids can directly promote cell death, and this is thought to reflect a nongenomic, physicochemical action of glucocorticosteroids.36 Thus high-dose steroid hormones may accelerate the termination of the inflammatory assault and thereby mitigate the effector phase of the disease.

Survival Factors in the PNS As a counterbalance to the autoimmune attack, endogenous tissue components may serve to attenuate the proinflammatory and destructive properties of activated complement factors. For example, Schwann cells are endowed on their membrane with a number of regulatory complement proteins such as CR1 (CD35), decay-accelerating factor CD55, and membrane cofactor proteins CD46 and CD59.35 Another possibility is through the expression of neurotrophic factors, which have been shown to protect glial cells from the destructive effect of TNF-␣.98 To further support survival of local glial cells that undergo apoptosis during EAN,149 therapeutic studies with neurotrophins were undertaken. Neither administration of ciliary neurotrophic factor39 nor of brain cell–derived neurotrophic factor28 ameliorated the disease course. Part of the treatment failure may

628

Neuroimmunology of the Peripheral Nervous System

obstacles not the least because these trials are very costly and time consuming. There is an obvious bias toward newly developed drugs and compounds provided by the pharmaceutical industry that are successfully explored in experimental models, and then enter the stage of controlled clinical trials. To test existing drugs, in particular in combination with already available effective treatments, investigator-initiated trials are needed, which require funding by institutional grants and larger international consortiums such as the Immune Neuropathies Cause and Treatment (INCAT) Group in Europe or the intramurally funded trials at the National Institutes of Health in the United States.

A

REFERENCES

B FIGURE 27–13 T-cell apoptosis in EAN. A, Inflammatory infiltrate in the sciatic nerve of an EAN rat after glucocorticosteroid therapy. T cells are labeled in red, and apoptotic cells appear black. B, At higher magnification some apoptotic cells are still labeled red for a T-cell membrane antigen (arrows). At late stages of T-cell apoptosis, only the black signal for fragmented DNA is left, and the membrane signal has been lost (arrowheads). Bar in B: 20 ␮m. See Color Plate

be due to reduced bioavailability of the neurotrophin at the inflammatory site.22 In analogy to the CNS,74 immune cells themselves could participate in the downregulation or termination of the inflammatory process, by means of “protective” immunity. One mechanism that may be involved in this process is local release of neurotrophic factors.

Future Perspectives for Treatment In conclusion, the field of modern immunotherapy in inflammatory neuropathies is rapidly moving forward, especially as we recognize targets of T-cell signaling and as newly developed designer molecules become available as more specific tools. The enlarging gap between our deeper understanding of molecular mechanisms in the animal models and the increased demands on the scientific quality standards of treatment trials pose increasing practical

1. Abromson-Leeman, S., Bronson, R., and Dorf, M. E.: Experimental autoimmune peripheral neuritis induced in BALB/c mice by myelin basic protein-specific T cell clones. J. Exp. Med. 182:587, 1995. 2. Archelos, J. J., Fortwangler, T., and Hartung, H. P.: Attenuation of experimental autoimmune neuritis in the Lewis rat by treatment with an antibody to L-selectin. Neurosci. Lett. 235:9, 1997. 3. Archelos, J. J., Maurer, M., Jung, S., et al.: Suppression of experimental allergic neuritis by an antibody to the intracellular adhesion molecule ICAM-1. Brain 116:1043, 1993. 4. Archelos, J. J., Maurer, M., Jung, S., et al.: Inhibition of experimental autoimmune neuritis by an antibody to the lymphocyte function-associated antigen-1. Lab. Invest. 70:667, 1994. 5. Archelos, J. J., Previtali, S. C., and Hartung, H. P.: The role of integrins in immune-mediated diseases of the nervous system. Trends Neurosci. 22:30, 1999. 6. Bai, X. F., Zhu, J., Zhang, G. X., et al.: IL-10 suppresses experimental autoimmune neuritis and down- regulates TH1-type immune responses. Clin. Immunol. Immunopathol. 83:117, 1997. 7. Batocchi, A. P., Rotondi, M., Caggiula, M., et al.: Leptin as a marker of multiple sclerosis activity in patients treated with interferon-beta. J. Neuroimmunol. 139:150, 2003. 8. Ben-Nun, A., Wekerle, H., and Cohen, I. R.: The rapid isolation of clonable antigen-specific T lymphocyte lines capable of mediating autoimmune encephalomyelitis. Eur. J. Immunol. 11:195, 1981. 9. Brosnan, C. F., Lyman, W. D., Tansey, F. A., and Carter, T. H.: Quantitation of mast cells in experimental allergic neuritis. J. Neuropathol. Exp. Neurol. 44:196, 1985. 10. Brosnan, C. F., and Tansey, F. A.: Delayed onset of experimental allergic neuritis in rats treated with reserpine. J. Neuropathol. Exp. Neurol. 43:84, 1984. 11. Buchwald, B., Ahangari, R., and Toyka, K. V.: Differential blocking effects of the monoclonal anti-GQ1b IgM antibody and alpha-latrotoxin in the absence of complement at the mouse neuromuscular junction. Neurosci. Lett. 334:25, 2002.

Experimental Autoimmune Neuritis 12. Buchwald, B., Ahangari, R., Weishaupt, A., and Toyka, K. V.: Intravenous immunoglobulins neutralize blocking antibodies in Guillain-Barre syndrome. Ann. Neurol. 51:673, 2002. 13. Buchwald, B., Toyka, K. V., Zielasek, J., et al.: Neuromuscular blockade by IgG antibodies from patients with Guillain-Barre syndrome: a macro-patch-clamp study. Ann. Neurol. 44:913, 1998. 14. Buchwald, B., Weishaupt, A., Toyka, K. V., and Dudel, J.: Immunoglobulin G from a patient with Miller-Fisher syndrome rapidly and reversibly depresses evoked quantal release at the neuromuscular junction of mice. Neurosci. Lett. 201:263, 1995. 15. Calida, D. M., Kermlev, S. G., Fujioka, T., et al.: Experimental allergic neuritis in the SJL/J mouse: induction of severe and reproducible disease with bovine peripheral nerve myelin and pertussis toxin with or without interleukin-12. J. Neuroimmunol. 107:1, 2000. 16. Campbell, J. J., Hedrick, J., Zlotnik, A., et al.: Chemokines and the arrest of lymphocytes rolling under flow conditions. Science 279:381, 1998. 17. Carroll, M. C., and Prodeus, A. P.: Linkages of innate and adaptive immunity. Curr. Opin. Immunol. 10:36, 1998. 18. Choi, Y. C., and Dalakas, M. C.: Expression of matrix metalloproteinases in the muscle of patients with inflammatory myopathies. Neurology 54:65, 2000. 19. Craggs, R. I., Brosnan, J. V., King, R. H., and Thomas, P. K.: Chronic relapsing experimental allergic neuritis in Lewis rats: effects of thymectomy and splenectomy. Acta Neuropathol. (Berl.) 70:22, 1986. 20. Dalakas, M. C., and Quarles, R. H.: Autoimmune ataxic neuropathies (sensory ganglionopathies): are glycolipids the responsible autoantigens? Ann. Neurol. 39:419, 1996. 21. Di Marco, R., Khademi, M., Wallstrom, E., et al.: Amelioration of experimental allergic neuritis by sodium fusidate (fusidin): suppression of IFN-gamma and TNFalpha and enhancement of IL-10. J Autoimmun. 13:187, 1999. 22. Dittrich, F., Thoenen, H., and Sendtner, M.: Ciliary neurotrophic factor: pharmacokinetics and acute-phase response in the rat. Ann. Neurol. 35:151, 1994. 23. Enders, U., Karch, H., Toyka, K. V., et al.: The spectrum of immune responses to Campylobacter jejuni and glycoconjugates in Guillain-Barre syndrome and in other neuroimmunological disorders. Ann. Neurol. 34:136, 1993. 24. Enders, U., Lobb, R., Pepinsky, R. B., et al.: The role of the very late antigen-4 (VLA-4) and its counterligand vascular cell adhesion molecule-1 (VCAM-1) in the pathogenesis of experimental autoimmune neuritis (EAN) of the Lewis rat. Brain 121:1257, 1998. 25. Enders, U., Toyka, K. V., Hartung, H. P., and Gold, R.: Failure of intravenous immunoglobulin (IVIg) therapy in experimental autoimmune neuritis (EAN) of the Lewis rat. J. Neuroimmunol. 76:112, 1997. 26. Eylar, E. H., Toro Goyco, E., Kessler, M. J., and Szymanska, I.: Induction of allergic neuritis in rhesus monkeys. J. Neuroimmunol. 3:91, 1982. 27. Feasby, T. E., Gilbert, J. J., Hahn, A. F., and Neilson, M.: Complement depletion suppresses Lewis rat experimental allergic neuritis. Brain Res. 419:97, 1987.

629

28. Felts, P. A., Smith, K. J., Gregson, N. A., and Hughes, R. A. C.: Brain-derived neurotrophic factor in experimental autoimmune neuritis. J. Neuroimmunol. 124:62, 2002. 29. Gabriel, C. M., Gregson, N. A., Redford, E. J., et al.: Human immunoglobulin ameliorates rat experimental autoimmune neuritis. Brain 120:1533, 1997. 30. Gabriel, C. M., Hughes, R. A. C., Moore, S. E., et al.: Induction of experimental autoimmune neuritis with peripheral myelin protein-22. Brain 121:1895, 1998. 31. Gaupp, S., Hartung, H. P., Toyka, K., and Jung, S.: Modulation of experimental autoimmune neuritis in Lewis rats by oral application of myelin antigens. J. Neuroimmunol. 79:129, 1997. 32. Gearing, A. J. H., and Newman, W.: Circulating adhesion molecules in disease. Immunol. Today 14:506, 1993. 33. Giese, K. P., Martini, R., Lemke, G., et al.: Mouse P0 gene disruption leads to hypomyelination, abnormal expression of recognition molecules, and degeneration of myelin and axons. Cell 71:565, 1992. 34. Gillen, C., Jander, S., and Stoll, G.: Sequential expression of mRNA for proinflammatory cytokines and interleukin-10 in the rat peripheral nervous system: comparison between immune-mediated demyelination and wallerian degeneration. J. Neurosci. Res. 51:489, 1998. 35. Gold, R., Archelos, J. J., and Hartung, H. P.: Mechanisms of immune regulation in the peripheral nervous system. Brain Pathol. 9:343, 1999. 36. Gold, R., Buttgereit, F., and Toyka, K. V.: Mechanism of action of glucocorticosteroid hormones: possible implications for therapy of neuroimmunological disorders. J. Neuroimmunol. 117:1, 2001. 37. Gold, R., Toyka, K. V., and Hartung, H. P.: Synergistic effect of IFN-gamma and TNF-alpha on expression of immune molecules and antigen presentation by Schwann cells. Cell. Immunol. 165:65, 1995. 38. Gold, R., Zielasek, J., Kiefer, R., et al.: Secretion of nitrite by Schwann cells and its effect on T-cell activation in vitro. Cell. Immunol. 168:69, 1996. 39. Gold, R., Zielasek, J., Schroder, J. M., et al.: Treatment with ciliary neurotrophic factor does not improve regeneration in experimental autoimmune neuritis of the Lewis rat. Muscle Nerve 19:1177, 1996. 40. Gregorian, S. K., Lee, W. P., Beck, L. S., et al.: Regulation of experimental autoimmune neuritis by transforming growth factor-beta 1. Cell. Immunol. 156:102, 1994. 41. Griffin, J. W., Stoll, G., Li, C. Y., et al.: Macrophage responses in inflammatory demyelinating neuropathies. Ann. Neurol. 27(Suppl.):S64, 1990. 42. Hadden, R. D. M., Gregson, N. A., Gold, R., et al.: Guillain-Barre syndrome serum and anti-Campylobacter antibody do not exacerbate experimental autoimmune neuritis. J. Neuroimmunol. 119:306, 2001. 43. Hadden, R. D. M., Gregson, N. A., Gold, R., et al.: Accumulation of immunoglobulin across the ‘blood-nerve barrier’ in spinal roots in adoptive transfer experimental autoimmune neuritis. Neuropathol. Appl. Neurobiol. 28:489, 2002. 44. Hafer-Macko, C., Hsieh, S. T., Li, C. Y., et al.: Acute motor axonal neuropathy: an antibody-mediated attack on axolemma. Ann. Neurol. 40:635, 1996.

630

Neuroimmunology of the Peripheral Nervous System

45. Hahn, A. F., Feasby, T. E., Steele, A., et al.: Demyelination and axonal degeneration in Lewis rat experimental allergic neuritis depend on the myelin dosage. Lab. Invest. 59:115, 1988. 46. Hartung, H. P., Heininger, K., Schafer, B., and Toyka, K. V.: Substance P and astrocytes: stimulation of the cyclooxygenase pathway of arachidonic acid metabolism. FASEB J. 2:48, 1988. 47. Hartung, H. P., Heininger, K., Schäfer, B., et al.: Immune mechanisms in inflammatory polyneuropathy. Ann. N. Y. Acad. Sci. 540:122, 1988. 48. Hartung, H. P., and Kieseier, B. C.: The role of matrix metalloproteinases in autoimmune damage to the central and peripheral nervous system. J. Neuroimmunol. 107:140, 2000. 49. Hartung, H. P., Pollard, J. D., Harvey, G. K., and Toyka, K. V.: Immunopathogenesis and treatment of the Guillain-Barré syndrome. Part I. Muscle Nerve 18:137, 1995. 50. Hartung, H. P., Schäfer, B., Diamantstein, T., et al.: Suppression of P2-T cell line-mediated experimental autoimmune neuritis by interleukin-2 receptor targeted monoclonal antibody ART 18. Brain Res. 489:120, 1989. 51. Hartung, H. P., Schäfer, B., Heininger, K., and Toyka, K. V.: Suppression of experimental autoimmune neuritis by the oxygen radical scavengers superoxide dismutase and catalase. Ann. Neurol. 23:453, 1988. 52. Hartung, H. P., Schäfer, B., Heininger, K., et al.: The role of macrophages and eicosanoids in the pathogenesis of experimental allergic neuritis: serial clinical, electrophysiological, biochemical and morphological observations. Brain 111:1039, 1988. 53. Hartung, H. P., Schäfer, B., Van der Meide, P. H., et al.: The role of interferon-gamma in the pathogenesis of experimental autoimmune disease of the peripheral nervous system. Ann. Neurol. 27:247, 1990. 54. Harvey, G. K., Gold, R., Toyka, K. V., and Hartung, H. P.: Nonneural-specific T lymphocytes can orchestrate inflammatory peripheral neuropathy. Brain 118:1263, 1995. 55. Harvey, G. K., Pollard, J. D., Schindhelm, K., and McLeod, J. G.: Experimental allergic neuritis: effect of plasma infusions. Clin. Exp. Immunol. 76:452, 1989. 56. Harvey, G. K., Schindhelm, K., Antony, J. H., and Pollard, J. D.: Membrane plasma exchange in experimental allergic neuritis: effect on antibody levels and clinical course. J. Neurol. Sci. 88:207, 1988. 57. Heininger, K., Fierz, W., Schafer, B., et al.: Electrophysiological investigations in adoptively transferred experimental autoimmune encephalomyelitis in the Lewis rat. Brain 112:537, 1989. 58. Heininger, K., Schäfer, B., Hartung, H. P., et al.: The role of macrophages in experimental autoimmune neuritis induced by a P2-specific T-cell line. Ann. Neurol. 23:326, 1988. 59. Heininger, K., Stoll, G., Linington, C., et al.: Conduction failure and nerve conduction slowing in experimental allergic neuritis induced by P2-specific T-cell lines. Ann. Neurol. 19:44, 1986. 60. Hila, S., Soane, L., and Koski, C. L.: Sublytic C5b-9stimulated Schwann cell survival through PI 3-kinasemediated phosphorylation of BAD. Glia 36:58, 2001.

61. Hughes, P. M., Wells, G. M. A., Clements, J. M., et al.: Matrix metalloproteinase expression during experimental autoimmune neuritis. Brain 121:481, 1998. 62. Huseby, E. S., Sather, B., Huseby, P. G., and Goverman, J.: Age-dependent T cell tolerance and autoimmunity to myelin basic protein. Immunity 14:471, 2001. 63. Izumo, S., Linington, C., Wekerle, H., and Meyermann, R.: Morphologic study on experimental allergic neuritis mediated by T cell line specific for bovine P2 protein in Lewis rats. Lab. Invest. 53:209, 1985. 64. Jander, S., Pohl, J., Gillen, C., and Stoll, G.: Differential expression of interleukin-10 mRNA in wallerian degeneration and immune-mediated inflammation of the rat peripheral nervous system. J. Neurosci. Res. 43:254, 1996. 65. Jander, S., and Stoll, G.: Interleukin-18 is induced in acute inflammatory demyelinating polyneuropathy. J. Neuroimmunol. 114:253, 2001. 66. Janeway, C. A. J.: The immune system evolved to discriminate infectious nonself from noninfectious self. Immunol. Today 13:11, 1992. 67. Jung, S., Gaupp, S., Hartung, H. P., and Toyka, K. V.: Oral tolerance in experimental autoimmune neuritis (EAN) of the Lewis rat – II. Adjuvant effects and bystander suppression in P2 peptide-induced EAN. J. Neuroimmunol. 116:21, 2001. 68. Jung, S., Huitinga, I., Schmidt, B., et al.: Selective elimination of macrophages by dichlormethylene diphosphonatecontaining liposomes suppresses experimental autoimmune neuritis. J. Neurol. Sci. 119:195, 1993. 69. Jung, S., Kramer, S., Schluesener, H. J., et al.: Prevention and therapy of experimental autoimmune neuritis by an antibody against T cell receptors-alpha/beta. J. Immunol. 148:3768, 1992. 70. Jung, S., Toyka, K., and Hartung, H. P.: T cell directed immunotherapy of inflammatory demyelination in the peripheral nervous system—potent suppression of the effector phase of experimental autoimmune neuritis by anti-CD2 antibodies. Brain 119:1079, 1996. 71. Kadlubowski, M., and Hughes, R. A. C.: Identification of the neuritogen for experimental allergic neuritis. Nature 277:140, 1979. 72. Kahl, K. G., Zielasek, J., Uttenthal, L. O., et al.: Protective role of the cytokine-inducible isoform of nitric oxide synthase induction and nitrosative stress in experimental autoimmune encephalomyelitis of the DA rat. J. Neurosci. Res. 73:198, 2003. 73. Kamradt, T., and Mitchison, N. A.: Tolerance and autoimmunity. N. Engl. J. Med. 334:655, 2001. 74. Kerschensteiner, M., Gallmeier, E., Behrens, L., et al.: Activated human T cells, B cells, and monocytes produce brain-derived neurotrophic factor in vitro and in inflammatory brain lesions: a neuroprotective role of inflammation? J. Exp. Med. 189:865, 1999. 75. Kiefer, R., Funa, K., Schweitzer, T., et al.: Transforming growth factor-beta 1 in experimental autoimmune neuritis: cellular localization and time course. Am. J. Pathol. 148:211, 1996. 76. Kiefer, R., Kieseier, B. C., Brück, W., et al.: Macrophage differentiation antigens in acute and chronic autoimmune polyneuropathies. Brain 121:469, 1998.

Experimental Autoimmune Neuritis 77. Kiefer, R., Kieseier, B. C., Stoll, G., and Hartung, H. P.: The role of macrophages in immune-mediated damage to the peripheral nervous system. Prog. Neurobiol. 64:109, 2001. 78. Kieseier, B. C., Clements, J. M., Pischel, H. B., et al.: Matrix metalloproteinases MMP-9 and MMP-7 are expressed in experimental autoimmune neuritis and the Guillain-Barré syndrome. Ann. Neurol. 43:427, 1998. 79. Kieseier, B. C., Seifert, T., Giovannoni, G., and Hartung, H. P.: Matrix metalloproteinases in inflammatory demyelination— targets for treatment. Neurology 53:20, 1999. 80. Kieseier, B. C., Storch, M. K., Archelos, J. J., et al.: Effector pathways in immune mediated central nervous system demyelination. Curr. Opin. Neurol. 12:323, 1999. 81. Kieseier, B. C., Tani, M., Mahad, D., et al.: Chemokines and chemokine receptors in inflammatory demyelinating neuropathies: a central role for IP-10. Brain 125:823, 2002. 82. King, R. H. M., Craggs, R. I., Cross, M. L. P., and Thomas, P. K.: Effects of glucocorticoids on experimental allergic neuritis. Exp. Neurol. 87:9, 1985. 83. Kobsar, I., Berghoff, M., Samsam, M., et al.: Preserved myelin integrity and reduced axonopathy in connexin32-deficient mice lacking the recombination activating gene-1. Brain 126:804, 2003. 84. Koehler, N. K., Martin, R., and Wietholter, H.: The antibody repertoire in experimental allergic neuritis: evidence for PMP-22 as a novel neuritogen. J. Neuroimmunol. 71:179, 1996. 85. Konat, G. W., and Wiggins, R. C.: Effect of reactive oxygen species on myelin membrane proteins. J. Neurochem. 45:1113, 1985. 86. Kyewski, B., Derbinski, J., Gotter, J., and Klein, L.: Promiscuous gene expression and central T-cell tolerance: more than meets the eye. Trends Immunol. 23:364, 2002. 87. Lafontaine, S., Rasminsky, M., Saida, T., and Sumner, A. J.: Conduction block in rat myelinated fibres following acute exposure to anti-galactocerebroside serum. J. Physiol. (Lond.) 323:287, 1982. 88. Lampert, P. W.: Mechanism of demyelination in experimental allergic neuritis: electron microscopic studies. Lab. Invest. 20:127, 1969. 89. Lannes-Viera, J., Goudable, B., Drexler, K., et al.: Encephalitogenic, myelin basic protein-specific T cells from naive rat thymus: preferential use of the T cell receptor gene V beta 8.2 and expression of the CD4⫺CD8⫺ phenotype. Eur. J. Immunol. 25:611, 1995. 90. Lassmann, H., Fierz, W., Neuchrist, C., and Meyermann, R.: Chronic relapsing experimental allergic neuritis induced by repeated transfer of P2-protein reactive T cell lines. Brain 114:429, 1991. 91. Lehmann, P. V., Forsthuber, T., Miller, A., and Sercarz, E. E.: Spreading of T-cell autoimmunity to cryptic determinants of an autoantigen. Nature 358:155, 1992. 92. Leppert, D., Hughes, P., Huber, S., et al.: Matrix metalloproteinase upregulation in chronic inflammatory demyelinating polyneuropathy and nonsystemic vasculitic neuropathy. Neurology 53:62, 1999. 93. Leussink, V. I., Zettl, U. K., Jander, S., et al.: Blockade of signaling via the very late antigen (VLA-4) and its counterligand vascular cell adhesion molecule-1 (VCAM-1) causes increased T cell apoptosis in experimental autoimmune neuritis. Acta Neuropathol. (Berl.) 103:131, 2002.

631

94. Liblau, R., Tisch, R., Bercovici, N., and McDevitt, H. O.: Systemic antigen in the treatment of T-cell-mediated autoimmune diseases. Immunol. Today 18:599, 1997. 95. Linington, C., Izumo, S., Suzuki, M., et al.: A permanent rat T cell line that mediates experimental allergic neuritis in the Lewis rat in vivo. J. Immunol. 133:1946, 1984. 96. Linington, C., Lassmann, H., Ozawa, K., et al.: Cell adhesion molecules of the immunoglobulin supergene family as tissue-specific autoantigens: induction of experimental allergic neuritis (EAN) by P0 protein-specific T cell lines. Eur. J. Immunol. 22:1813, 1992. 97. Linington, C., Mann, A., Izumo, S., et al.: Induction of experimental allergic neuritis in the BN rat: P2 proteinspecific T cells overcome resistance to actively induced disease. J. Immunol. 137:3826, 1986. 98. Louis, J. C., Magal, E., Takayama, S., and Varon, S.: CNTF protection of oligodendrocytes against natural and tumor necrosis factor-induced death. Science 259:689, 1993. 99. McCombe, P. A., van der Kreek, S. A., and Pender, M. P.: Neuropathological findings in chronic relapsing experimental allergic neuritis induced in the Lewis rat by inoculation with intradural root myelin and treatment with low dose cyclosporin A. Neuropathol. Appl. Neurobiol. 18:171, 1992. 100. Miyamoto, K., Miyake, S., Schachner, M., and Yamamura, T.: Heterozygous null mutation of myelin P0 protein enhances susceptibility to autoimmune neuritis targeting P0 peptide. Eur. J. Immunol. 33:656, 2003. 101. Moser, B., and Loetscher, P.: Lymphocyte traffic control by chemokines. Nat. Immunol. 2:123, 2001. 102. Mueller, M., Wacker, K., Ringelstein, E. B., et al.: Rapid response of identified resident endoneurial macrophages to nerve injury. Am. J. Pathol. 159:2187, 2001. 103. Nemazee, D.: Receptor selection in B and T lymphocytes. Annu. Rev. Immunol. 18:19, 2000. 104. Ochsenbein, A. F., and Zinkernagel, R. M.: Natural antibodies and complement link innate and acquired immunity. Immunol. Today 21:624, 2000. 105. Olee, T., Powell, H. C., and Brostoff, S. W.: New minimum length requirement for a T cell epitope for experimental allergic neuritis. J. Neuroimmunol. 27:187, 1990. 106. Olee, T., Weise, M., Powers, J., and Brostoff, S.: A T cell epitope for experimental allergic neuritis is an amphipathic alpha-helical structure. J. Neuroimmunol. 21:235, 1989. 107. Pelidou, S. H., Deretzi, G., Zou, L. P., et al.: Inflammation and severe demyelination in the peripheral nervous system induced by the intraneural injection of recombinant mouse interleukin-12. Scand. J. Immunol. 50:39, 1999. 108. Pender, M. P., Stanley, G. P., Yoong, G., and Nguyen, K. B.: The neuropathology of chronic relapsing experimental allergic encephalomyelitis induced in the Lewis rat by inoculation with whole spinal cord and treatment with cyclosporin A. Acta Neuropathol. (Berl.) 80:172, 1990. 109. Powell, H. C., Olee, T., Brostoff, S. W., and Mizisin, A. P.: Comparative histopathology of experimental allergic neuritis induced with minimum length neuritogenic peptides by adoptive transfer with sensitized cells or direct sensitization. J. Neuropathol. Exp. Neurol. 50:658, 1991. 110. Rathmell, J. C., Townsend, S. E., Xu, J. C., et al.: Expansion or elimination of B cells in vivo: dual roles for. Cell 87:319, 1996.

632

Neuroimmunology of the Peripheral Nervous System

111. Redford, E. J., Hall, S. M., and Smith, K. J.: Vascular changes and demyelination induced by the intraneural injection of tumour necrosis factor. Brain 118:869, 1995. 112. Redford, E. J., Smith, K. J., Gregson, N. A., et al.: A combined inhibitor of matrix metalloproteinase activity and tumour necrosis factor-a processing attenuates experimental autoimmune neuritis. Brain 120:1895, 1997. 113. Rose, N. R., and Bona, C.: Defining criteria for autoimmune diseases (Witebsky’s postulates revisited). Immunol. Today 14:426, 1993. 114. Rosen, J. L., Brown, M. J., Hickey, W. F., and Rostami, A.: Early myelin lesions in experimental allergic neuritis. Muscle Nerve 13:629, 1990. 115. Said, G., and Hontebeyrie Joskowicz, M.: Nerve lesions induced by macrophage activation. Res. Immunol. 143:589, 1992. 116. Saida, K., Saida, T., Brown, M. J., et al.: Antiserum-mediated demyelination in vivo: a sequential study using intraneural injection of experimental allergic neuritis serum. Lab. Invest. 39:449, 1978. 117. Salomon, B., and Bluestone, J. A.: Complexities of CD28/B7: CTLA-4 costimulatory pathways in autoimmunity and transplantation. Annu. Rev. Immunol. 19:225, 2001. 118. Salomon, B., Rhee, L., Bour-Jordan, H., et al.: Development of spontaneous autoimmune peripheral polyneuropathy in B7-2-deficient NOD mice. J. Exp. Med. 194:677, 2001. 119. Schabet, M., Whitaker, J. N., Schott, K., et al.: The use of protease inhibitors in experimental allergic neuritis. J. Neuroimmunol. 31:265, 1991. 120. Schmid, C. D., Stienekemeier, M., Oehen, S., et al.: Immune deficiency in mouse models for inherited peripheral neuropathies leads to improved myelin maintenance. J. Neurosci. 20:729, 2000. 121. Schmidt, B., Stoll, G., Hartung, H. P., et al.: Macrophages but not Schwann cells express Ia antigen in experimental autoimmune neuritis. Ann. Neurol. 28:70, 1990. 122. Schmidt, B., Stoll, G., van der Meide, P., et al.: Transient cellular expression of gamma-interferon in myelin-induced and T-cell line-mediated experimental autoimmune neuritis. Brain 115:1633, 1992. 123. Schmidt, J., Elflein, K., Stienekemeier, M., et al.: Treatment and prevention of experimental autoimmune neuritis with superagonistic CD28-specific monoclonal antibodies. J. Neuroimmunol. 140:143, 2003. 124. Schmidt, J., Metselaar, J. M., Wauben, M. H., et al.: Drug targeting by long-circulating liposomal glucocorticosteroids increases therapeutic efficacy in a model of multiple sclerosis. Brain 126:1895, 2003. 125. Sharief, M. K., Ingram, D. A., and Swash, M.: Circulating tumor necrosis factor-alpha correlates with electrodiagnostic abnormalities in Guillain-Barre syndrome. Ann. Neurol. 42:68, 1997. 126. Spies, J. M., Pollard, J. D., Bonner, J. G., et al.: Synergy between antibody and P2-reactive T cells in experimental allergic neuritis. J. Neuroimmunol. 57:77, 1995. 127. Stienekemeier, M., Falk, K., Rotzschke, O., et al.: Vaccination, prevention, and treatment of experimental autoimmune neuritis (EAN) by an oligomerized T cell epitope. Proc. Natl. Acad. Sci. U. S. A. 98:13872, 2001.

128. Stoll, G., Reiners, K., Schwendemann, G., et al.: Normal myelination of regenerating peripheral nerve sprouts despite circulating antibodies to galactocerebroside in rabbits. Ann. Neurol. 19:189, 1986. 129. Stoll, G., Griffin, J. W., Li, C. Y., and Trapp, B. D.: Wallerian degeneration in the peripheral nervous system: participation of both Schwann cells and macrophages in myelin degradation. J. Neurocytol. 18:671, 1989. 130. Stoll, G., Jung, S., Jander, S., et al.: Tumor necrosis factoralpha in immune-mediated demyelination and wallerian degeneration of the rat peripheral nervous system. J. Neuroimmunol. 45:175, 1993. 131. Stoll, G., Schmidt, B., Jander, S., et al.: Presence of the terminal complement complex (C5b-9) precedes myelin degradation in immune-mediated demyelination of the rat peripheral nervous system. Ann. Neurol. 30:147, 1991. 132. Stoll, G., Schwendemann, G., Heininger, K., et al.: Relation of clinical, serological, morphological, and electrophysiological findings in galactocerebroside-induced experimental allergic neuritis. J. Neurol. Neurosurg. Psychiatry 49:258, 1986. 133. Tacke, M., Clark, G. J., Dallman, M. J., and Hünig, T.: Cellular distribution and costimulatory function of rat CD28: regulated expression during thymocyte maturation and induction of cyclosporin A sensitivity of costimulated T cell responses by phorbol ester. J. Immunol. 154:5121, 1995. 134. Tacke, M., Hanke, G., Hanke, T., and Hunig, T.: CD28mediated induction of proliferation in resting T cells in vitro and in vivo without engagement of the T cell receptor: evidence for functionally distinct forms of CD28. Eur. J. Immunol. 27:239, 1997. 135. Takeda, K., Kaisho, T., and Akira, S.: Toll-like receptors. Annu. Rev. Immunol. 21:335, 2003. 136. Tansey, F. A., and Brosnan, C. F.: Protection against experimental allergic neuritis with silica quartz dust. J. Neuroimmunol. 3:169, 1982. 137. Taylor, W. A., and Hughes, R. A.: Experimental allergic neuritis induced in SJL mice by bovine P2. J. Neuroimmunol. 8:153, 1985. 138. Toyka, K. V., and Heininger, K.: Humoral factors in peripheral nerve disease. Muscle Nerve 10:222, 1987. 139. Trinchieri, G.: Interleukin-12: a proinflammatory cytokine with immunoregulatory functions that bridge innate resistance and antigen-specific adaptive immunity. Annu. Rev. Immunol. 13:251, 1995. 140. Uncini, A., Di Muzio, A., Di Guglielmo, G., et al.: Effect of rhTNF-alpha injection into rat sciatic nerve. J. Neuroimmunol. 94:88, 1999. 141. Van Parijs, L., Biuckians, A., and Abbas, A. K.: Functional roles of Fas and Bcl-2-regulated apoptosis of T lymphocytes. J. Immunol. 160:2065, 1998. 142. Visan, L. A., Visan, A., Weishaupt, H. H., et al.: Tolerance induction by intrathymic expression of myelin P0 protein. J. Immunol. 172:1364, 2004. 143. Vriesendorp, F. J., Flynn, R. E., Malone, M. R., and Pappolla, M. A.: Systemic complement depletion reduces inflammation and demyelination in adoptive transfer experimental allergic neuritis. Acta Neuropathol. (Berl.) 95:297, 1998. 144. Vriesendorp, F. J., Flynn, R. E., Pappolla, M. A., and Koski, C. L.: Complement depletion affects demyelination and

Experimental Autoimmune Neuritis

145.

146.

147.

148.

149.

150.

151.

152.

153.

inflammation in experimental allergic neuritis. J. Neuroimmunol. 58:157, 1995. Waksman, B. H., and Adams, R. D.: Allergic neuritis: an experimental disease of rabbits induced by the injection of peripheral nervous tissue and adjuvants. J. Exp. Med. 102:213, 1955. Wauben, M. H. M., Hoedemaekers, A. C. W. E., Graus, Y. M. F., et al.: Coimmunization of MHC class II competitor peptides during experimental autoimmune myasthenia gravis induction resulted not only in a suppressed, but also in an altered immune response. Ann. N. Y. Acad. Sci. 841:338, 1998. Weerth, S., Berger, T., Lassmann, H., and Linington, C.: Encephalitogenic and neuritogenic T cell responses to the myelin-associated glycoprotein (MAG) in the Lewis rat. J. Neuroimmunol. 95:157, 1999. Weiner, H. L., Friedman, A., Miller, A., et al.: Oral tolerance: immunologic mechanisms and treatment of animal and human organ-specific autoimmune diseases by oral administration of autoantigens. Annu. Rev. Immunol. 12:809, 1994. Weishaupt, A., Bruck, W., Hartung, T., et al.: Schwann cell apoptosis in experimental autoimmune neuritis of the Lewis rat and the functional role of tumor necrosis factor-alpha. Neurosci. Lett. 306:77, 2001. Weishaupt, A., Gold, R., Gaupp, S., et al.: Antigen therapy eliminates T cell inflammation by apoptosis: effective treatment of experimental autoimmune neuritis with recombinant myelin protein P2. Proc. Natl. Acad. Sci. U. S. A. 94:1338, 1997. Weishaupt, A., Gold, R., Hartung, T., et al.: Role of TNFalpha in high-dose antigen therapy in experimental autoimmune neuritis: inhibition of TNF-alpha by neutralizing antibodies reduces T-cell apoptosis and prevents liver necrosis. J. Neuropathol. Exp. Neurol. 59:368, 2000. Wekerle, H., Linington, C., Lassmann, H., and Meyermann, R.: Cellular immune reactivity within the CNS. Trends Neurosci. 9:271, 1986. Wekerle, H., Schwab, M., Linington, C., and Meyermann, R.: Antigen presentation in the peripheral nervous system: Schwann cells present endogenous myelin autoantigens to lymphocytes. Eur. J. Immunol. 16:1551, 1986.

633

154. Westland, K., and Pollard, J. D.: Proteinase induced demyelination: an electrophysiological and histological study. J. Neurol. Sci. 82:41, 1987. 155. Whitehouse, D. J., Whitehouse, M. W., and Pearson, C. M.: Passive transfer of adjuvant-induced arthritis and allergic encephalomyelitis in rats using thoracic duct lymphocytes. Nature 224:1322, 1969. 156. Yan, W. X., Archelos, J. J., Hartung, H. P., and Pollard, J. D.: P0 protein is a target antigen in chronic inflammatory demyelinating polyradiculoneuropathy. Ann. Neurol. 50:286, 2001. 157. Yuki, N., Yamada, M., Koga, M., et al.: Animal model of axonal Guillain-Barre syndrome induced by sensitization with GM1 ganglioside. Ann. Neurol. 49:712, 2001. 158. Zettl, U. K., Gold, R., Toyka, K. V., and Hartung, H. P.: Intravenous glucocorticosteroid treatment augments apoptosis of inflammatory T cells in experimental autoimmune neuritis (EAN) of the Lewis rat. J. Neuropathol. Exp. Neurol. 54:540, 1995. 159. Zettl, U. K., Gold, R., Toyka, K. V., and Hartung, H. P.: In situ demonstration of T cell activation and elimination in the peripheral nervous system during experimental autoimmune neuritis in the Lewis rat. Acta Neuropathol. (Berl.) 91:360, 1996. 160. Zhu, J., Bai, X. F., Mix, E., and Link, H.: Cytokine dichotomy in peripheral nervous system influences the outcome of experimental allergic neuritis: dynamics of mRNA expression for IL-1b, IL-6, IL-10, IL-12, TNF-a, TNF-b, and cytolysin. Clin. Immunol. Immunopathol. 84:85, 1997. 161. Zhu, J., Link, H., Weerth, S., et al.: The B cell repertoire in experimental allergic neuritis involves multiple myelin proteins and GM1. J. Neurol. Sci. 125:132, 1994. 162. Zhu, Y., Ljunggren, H. G., Mix, E., et al.: CD28-B7 costimulation: a critical role for initiation and development of experimental autoimmune neuritis in C57BL/6 mice. J. Neuroimmunol. 114:114, 2001. 163. Zielasek, J., Jung, S., Gold, R., et al.: Administration of nitric oxide synthase inhibitors in experimental autoimmune neuritis and experimental autoimmune encephalomyelitis. J. Neuroimmunol. 58:81, 1995.

28 Principles of Immunotherapy JOHN D. POLLARD, HANS-PETER HARTUNG, AND RICHARD A. C. HUGHES

Immunosuppressive Agents Glucocorticoids Alkylating Agents

Immunophilin Binding Agents Azathioprine Mycophenolate Mofetil

Immunotherapy for immune-mediated disorders of the peripheral nervous system (PNS) is currently under rapid development. The aim of an effective therapy is to reduce severity and duration of clinical deterioration, as in monophasic diseases such as Guillain-Barré syndrome (GBS), and to reduce frequency of relapses and to prevent disability from disease progression, as in chronic diseases such as chronic inflammatory demyelinating polyradiculoneuropathy (CIDP). When choosing treatment with short- and/or long-term effects, benefits and risks of the therapeutic approach need to be weighed carefully in each individual patient. Unfortunately, our therapeutic armamentarium, consisting of glucocorticoids (GCs), alkylating agents, plasma exchange (PE), intravenous immunoglobulin (IVIg), and others, is not always effective enough to stop ongoing disease. In GBS, for example, approximately 50% to 60% of patients do respond to plasma exchange or high-dose IVIg, but up to 40% develop significant long-term disability (see Chapter 98). Thus there is a pressing need to provide more effective therapeutic strategies. Large efforts have been undertaken to further increase our current understanding of the pathogenetic role of various factors in immune-mediated peripheral neuropathies. New approaches including the use, for example, of more effective immunomodulators, as tested already in the animal model experimental autoimmune neuritis (EAN),67 or compounds that target the transmigration of immunocompetent cells across the blood-nerve barrier by blocking specific chemokine receptors, adhesion molecules, or proteases,4,63,64 are underway. However, whether these encouraging approaches, effective in the animal model, will

Immunomodulatory Agents Therapeutic Plasma Exchange Intravenous Immunoglobulin

translate into efficacious clinical therapy still needs to be evaluated. At present all therapeutic approaches are aimed to modulate or to suppress the immune system in order to reduce the amount of damage to the myelin sheath and/or the axon. Several lines of research are trying to identify strategies that promote remyelination and axonal regeneration. Neurotrophic factors, at least in theory, could represent a group of proteins that might promote regeneration. In a small, placebo-controlled pilot trial of 10 GBS patients, the potential role for subcutaneous brain-derived neurotrophic factor was examined. However, no difference in the outcome between the groups was observed.13 Similarly, no striking effects have been seen with different neurotrophins in EAN.38,71 Although present data are not encouraging, the favorable effects of neurotrophic factors on nerve regeneration define them as interesting molecules that might be clinically helpful in the near future. Furthermore, recent evidence indicates that immunocompetent cells can produce neurotrophins, suggesting that the inflammatory reaction might exhibit some beneficial or even neuroprotective effects.61 Nevertheless, a concept of a “neuroprotective immunity” cannot fully be established at present.59 It is unlikely that a single treatment will be effective in all patients as long as the pathogenesis of the underlying immune-mediated disease is not precisely understood. The principle of combination therapy has found utility in cancer treatment and is evolving as a strategic principle for immune-mediated peripheral neuropathies. Rationales for combining different mechanistic approaches in order to target a broader spectrum of critical immune mechanisms 635

636

Neuroimmunology of the Peripheral Nervous System

exist. Studies in animal models support the concept that different treatment strategies can act synergistically, as demonstrated with GCs and the administration of a specific antigen in EAN.117 Immunomodulation may find utility in conjunction with specific forms of therapy. As it becomes clearer which mechanistic strategies work best, more rational combinations can be designed. In chronic immune-mediated disorders a major problem for the treating neurologist is to predict the clinical course of the disease. In an individual patient it may be difficult to determine whether the chosen treatment regimen is having a therapeutic effect unless the patient improves rapidly during therapy. In some instances it becomes clear that a patient is either a responder or a nonresponder to a particular treatment; however, this clear dichotomy remains difficult to establish in many other patients. Thus surrogate markers that are linked to the underlying immunopathogenesis might be a valuable tool to objectively measure ongoing disease activity. Such markers are not fully established yet for immune-mediated neuropathies of the PNS; however, various markers are under critical investigation at present. It may be of great practical value to have useful surrogate markers in predicting the clinical response of patients at hand, because it may be helpful to stratify affected patients into different therapeutic regimens. Moreover, such markers may facilitate the administration of certain treatments at critical checkpoints during the clinical course of the disease, optimizing therapeutic effects and minimizing side effects. At present, the treating neurologist must make decisions on therapy based on clinical assessment and accumulation of disability. Moreover, he or she has to define whether a patient has responded to therapy or additional treatment should be given. This chapter discusses the available therapeutic arsenal for immunotherapy in disorders of the PNS.

IMMUNOSUPPRESSIVE AGENTS Glucocorticoids Glucocorticoids have powerful anti-inflammatory and immunosuppressive effects that have led to their use in a large number of autoimmune disorders and in transplant rejection. They are known to be associated with serious side effects that depend on dose and duration of treatment. Mechanism of Action Glucocorticoids act directly and indirectly by affecting gene transcription. They diffuse across all membranes by means of their high lipid solubility. Within the cytoplasm they bind to specific glucocorticoid receptors (GCRs). The cytosolic GCR is associated with immunophilin and heat shock proteins and, upon ligand binding, a conformational

change occurs with disassociation of the GC-GCR complex from the associated proteins. The complex is then transported from the cytoplasm to the nucleus, where it binds to specific regions of DNA known as glucocorticoid response elements (GREs). Genes with GREs can be directly influenced by GCs.8 Other genes are affected indirectly by means of GC effects on transcription factors. For instance, GC-GCR complexes may bind to nuclear factor-␬B (NF-␬B), which mediates upregulation of cytokine gene transcription. Binding of GC-GCR to NF-␬B inhibits the latter from binding to its DNA binding site, thus inhibiting pro-inflammatory gene transcription.1,8 Glucocorticoids can also alter gene transcription by upregulating the inhibitory molecule I␬B␣, which inactivates NF-␬B, binding to it before it can enter the nucleus.6 Glucocorticoids have also been shown to alter posttranscriptional mechanisms, including messenger RNA translation and protein synthesis. At the cellular level GCs have been shown to inhibit the production of a number of pro-inflammatory cytokines, including interleukin (IL)-1, IL-2, IL-3, IL-4, IL-5, IL-6, IL-13, granulocyte-macrophage colony-stimulating factor, and tumor necrosis factor (TNF)-␣. Glucocorticoids decrease the number of inflammatory cells entering a site of inflammation, and this effect is thought to be an indirect one because upregulation of those adhesion molecules involved in transmigration (e.g., vascular cell adhesion molecule-1, intracellular adhesion molecule [ICAM]-1) is dependent on cytokine production.107 Glucocorticoids inhibit phagocytosis by the downregulation of Fc␥ receptors on macrophages and monocytes5 and can inhibit the release of lysosomal enzymes from monocytes. B lymphocytes are relatively resistant to the effects of GCs, but immunoglobulin levels may be reduced by highdose therapy. Glucocorticoids can also inhibit lipid mediators of inflammation, including the cyclooxygenase metabolites prostaglandin D2, thromboxane B2, prostaglandin F2, and the leukotrienes. Thus GCs influence the movement and distribution of leukocytes, their functional properties, the synthesis and secretion of cytokines and other immune mediators, and microvascular permeability. Pharmacokinetics Glucocorticoids are lipophilic and rapidly and relatively completely absorbed when given orally, but because they may undergo presystemic metabolism, they may be incompletely bioavailable by the oral route. Intravenous steroids may be given as the water-soluble esters (e.g., succinates and phosphates), and these esters are rapidly hydrolyzed in the circulation so that the steroid is present as the corresponding free alcohol.12 Glucocorticoids bind to plasma proteins (␣-globulin), and it is the free steroid that is responsible for the biochemical and physiologic effects. Elimination occurs almost completely by way of metabolism. The II-oxo steroids cortisone and prednisone are reduced reversibly to their II-hydroxy

637

Principles of Immunotherapy

analogues, cortisol and prednisolone, their biologically active forms. The phase I metabolites undergo glucuronide or sulfate conjugation before secretion in the urine. Peak plasma concentrations occur between 1 and 2 hours following prednisone ingestion. The half-life of the GC is relatively short, in the range of 2 to 4 hours.90 Corticosteroid metabolism is altered by numerous other drugs, particularly the anticonvulsants phenytoin, carbamazepine, and phenobarbital, which induce the monooxygenase system. These agents increase the elimination of prednisone and methylprednisolone to a very significant degree.10 Ketoconazole, the macrolide antibiotics, and oral contraceptives can delay the elimination of the glucocorticoids. Adverse Effects Glucocorticoids are used therapeutically mainly for their anti-inflammatory actions. However, they have diverse metabolic effects that may constitute unwanted and adverse reactions. Because all nucleated cells have the same GCR, all cells are susceptible to the development of adverse events. The metabolic effects of steroids are mediated largely by their binding to GREs on DNA, whereas the anti-inflammatory effects result mainly from inhibitors of transcription factors. It may therefore be possible ultimately to design a GC with anti-inflammatory effects but without metabolic complications. The adverse effects of steroids are well known and include endocrine (adrenal suppression, growth suppression in children, increased weight and cushingoid appearance, hyperglycemia, glycosuria, hypokalemia, hyperlipidemia); musculoskeletal (myopathy, osteoporosis, aseptic necrosis of bone); gastrointestinal (dyspepsia, peptic ulceration); psychological (mood swings, psychosis, insomnia); ophthalmologic (cataract formation, glaucoma); cardiovascular (hypertension, arteriosclerosis); dermatologic (increased fragility of skin, ecchymoses, hirsutism, acne, striae); and immunologic (lymphopenia, susceptibility to opportunistic infection, delayed wound healing). The most common of these in patients on long-term therapy are osteoporosis, hypertension, diabetes, ulcer formation, cataracts, and increased weight. Use in Neuropathy Based on the mechanisms of action described earlier and current understanding of the pathogenesis of the various immune-mediated neuropathies, the efficacy of GCs in these neuropathies is fairly predictable. Glucocorticoids do not improve patients with GBS even when given by high-dose pulse therapy.52,54,113 Pathogenic antibodies to various gangliosides have been demonstrated at disease presentation in certain subtypes (Miller Fisher syndrome, acute motor axonal neuropathy), and these would not be altered by GCs. Similarly, neuropathies associated with immunoglobulin M (IgM) antibodies (i.e., anti–myelinassociated glycoprotein and anti-GM1 antibodies), such as

multifocal motor neuropathy (MMN), do not benefit from GCs.76 IgM antibodies tend to be T-cell independent, and B cells are relatively resistant to GCs. However, CIDP probably does respond to GCs.33,50,80 Corticosteroids are usually given in a daily oral dose but have been administered as pulsed doses of intravenous methylprednisone given over 5 days each month.68 The difference in treatment response suggests significant differences in pathogenic mechanisms in these acute and chronic disorders. The use of GCs in CIDP has largely been relegated to supportive treatment to IVIg and plasmapheresis because of their adverse affects. Similarly, even in the absence of randomized trials, experts regard GCs as an essential part of the therapy for vasculitic neuropathy,39 a disorder in which T cells are prominent in the pathogenesis. In those cases associated with systemic vasculitis, GCs are usually combined with cyclophosphamide.36 Glucocorticoids are also indicated, according to expert opinion, in high oral doses (60 to 100 mg/day) in the treatment of type 1 and 2 reactions of leprosy, which are characterized by an increase in cell-mediated immunoreactivity.95 GCs have not been found beneficial in other forms of generalized peripheral neuropathy. Anecdotal reports of their successful use in paraneoplastic neuropathy and subacute sensory neuronopathy are the exception rather than the rule, and randomized trials have not been performed. Their use in Bell’s palsy is controversial because of a lack of adequate randomized trials.96 The injection of GCs into the wrist of patients with carpal tunnel syndrome significantly ameliorates symptoms for up to 1 month, but the duration of the effect and the extent to which such injections avoid or delay the need for operation are unclear.79

Alkylating Agents Cyclophosphamide, chlorambucil, and melphalan are the most commonly used alkylating agents, the first of which to be used in clinical practice was nitrogen mustard. Although their use has greatly improved the outlook for patients with certain disorders such as vasculitis and lupus nephritis, severe toxicity has restricted their more general use. Mechanism of Action Alkylating agents act largely by cross-linking DNA and RNA and hence inhibit transcription and translation in both proliferating and resting cells. Cyclophosphamide and chlorambucil are rapidly metabolized to their active forms, phosphoramide mustard and phenylacetic acid mustard, respectively. Cyclophosphamide affects both B and T cells, and hence reduces both cell-mediated and antibodymediated immune reactions.123 The immunomodulatory effect depends on the dose and duration of treatment. Low-dose effects are attributed to deletion of suppressor cell precursors.45 Moderate doses deplete B cells without significant effects on T-cell–mediated immune responses.

638

Neuroimmunology of the Peripheral Nervous System

Higher doses inhibit T helper cells, and very high doses reduce cytotoxic T cells and natural killer cells.88 Pharmacokinetics Cyclophosphamide can be given orally or intravenously. It is metabolized in the liver to its active metabolites, and about 95% is excreted by the kidney. Chlorambucil is usually given orally and is also metabolized to its active metabolite, phenylacetic acid mustard, in the liver. Excretion is exclusively by the kidney. Adverse Effects Unwanted side effects are often dose related. Common adverse effects include alopecia, nausea, and vomiting, but the more important ones are bone marrow suppression, bladder toxicity, cancer, and infertility. Suppression of bone marrow is dose dependent. Following intravenous pulse treatment, minimum white blood cell (WBC) counts occur at between 1 and 2 weeks, and dose change is necessary if the leukocyte count is less than 1.5 ⫻ 106 WBCs/L. Regular monitoring of the WBC count is needed in patients on daily oral therapy, and the dose should be changed if the count falls below 3 ⫻ 106 WBCs/L. Hemorrhagic cystitis is more common in patients given daily oral cyclophosphamide, probably because of continued exposure to the toxic metabolite acrolein; however, its incidence has been reduced by encouraging adequate hydration. Patients given pulse intravenous therapy should be given pre- and postinfusion hydration and coadministration of mesna (2-mercaptoethane sulfonate), which has been reported to bind and neutralize the toxic metabolite acrolein.56 There is a greater risk of bladder and hematologic malignancy (leukemia and lymphoma) in patients given a total dose greater than 80 g. Monthly urinalysis is indicated during the period of therapy and at 6-month intervals thereafter. Cystoscopy is required in any patient found to have hematuria. The incidence of infertility increases with cumulative dose and patient age. The adverse effects of chlorambucil are similar to those of cyclophosphamide except that it does not produce bladder toxicity and is more prone to induce hematologic malignancy. Use in Neuropathy Despite the absence of randomized trials, we regard cyclophosphamide, given usually in association with corticosteroids, as the treatment of choice in patients with neuropathy caused by systemic vasculitis. It may be given by daily oral therapy at a dose of 2 mg/kg/day, with dose adjustments as necessary to keep the leukocyte count greater than 3.0 ⫻ 106 WBCs/L, together with prednisone 1 mg/kg/day with subsequent tapering. Hemorrhagic cystitis is less common when the drug is given by monthly intravenous therapy.56 Pulses are given in a dose of 0.5 to 1 g/m2 body surface area monthly for 6 months and thereafter at

3-month intervals for 1 to 2 years. Adequate hydration with frequent urination is particularly important for patients on daily therapy. Cyclophosphamide has also been recommended in patients with MMN. Although most of these patients respond to IVIg, they may continue to slowly deteriorate, and cyclophosphamide may allow disease arrest and even regression. Clinical improvement in association with reduction of anti-GM1 antibodies has been reported following oral cyclophosphamide.37 Pestronk and colleagues89 reported a superior response after high-dose intravenous induction therapy followed by intermittent monthly courses. A supplemental dose of oral cyclophosphamide (1 to 3 mg/kg) decreased the required frequency of IVIg in a further study.82 Randomized trials have not been performed.110 Both cyclophosphamide and chlorambucil have been used in patients with progressive neuropathy associated with IgM paraproteinemia85 but only in uncontrolled studies in small numbers of patients.76 Immunosuppression with cyclophosphamide in association with PE has been suggested by others. Cyclophosphamide was reported to benefit 80% of CIDP cases in an uncontrolled series when given with or without prednisone as pulse therapy in a dose of 1 g/m2 each month for 3 to 6 months. However, 30% of the patients relapsed and needed further treatment in 6 to 36 months.40 Once again, randomized trials have not been performed.53

Immunophilin Binding Agents The immunophilins are a large family of highly conserved, ubiquitously expressed proteins initially characterized by their ability to bind a number of immunosuppressive agents, including cyclosporin A (CsA), tacrolimus (FK506), and sirolimus (rapamycin). Three groups are recognized: cyclophilins, which bind CsA; FK506 binding proteins (FKBPs), which bind tacrolimus and sirolimus; and the parvulins. It has become clear that immunophilins bind a number of other molecules besides immunosuppressive agents, including heat shock proteins and cell cycle regulatory factors. They share the capacity to catalyze the isomerization of protein residues between their cis and trans configurations and are thought to play a role in protein folding.106 Immunophilins are currently considered to play a role in immunosuppression, protein folding, calcium homeostasis, and the regulation of RNA splicing. Mechanism of Action The immunosuppressive capacity of CsA and FK506 depends on the ability of the complex formed between CsA and its cytophilin and that between tacrolimus and its FKBP to inhibit calcineurin. Calcineurin is a calcium/calmodulin serine/threonine phosphatase that is an essential intermediate in T-cell signaling from the T-cell

639

Principles of Immunotherapy

antigen receptor.75 Calcineurin activity is necessary for transcriptional activation of a number of cytokine and other genes. Calcineurin binds to and dephosphorylates the transcriptional factor nuclear factor of activated T cell (NFAT) in the cytoplasm. Dephosphorylation of NFAT results in its translocation to the nucleus, where it binds to its target sequence on the promotor region of many genes, including the cytokine genes IL-2, IL-3, IL-4, IL-12, and TNF, resulting in their activation.93 NF-␬B is also activated by calcineurin.108 Hence, following stimulation of T cells via the T-cell receptor, a number of intracellular signaling events are initiated, including the activation of protein kinase C in addition to an increase in intracellular calcium levels. Calcium and calmodulin are recruited to activate calcineurin. In the presence of a complex of CsA with its cytophilin or tacrolimus with its FKBP, the activation of calcineurin is inhibited and hence also cytokine activation. The effects of cyclosporin are seen mainly on T cells, with B cells and macrophages apparently unaffected. Cyclosporin and calcineurin also regulate the production of transforming growth factor-␤, the increased production of which correlates with renal fibrosis and nephrotoxicity.24,62 Although sirolimus (rapamycin) resembles tacrolimus in structure and binds to the same family of immunophilins, it has a different molecular target. It has been shown to function only when bound to an FKBP in an intracellular complex that does not bind calcineurin. This target protein is termed the mammalian target of rapamycin (sirolimus) (mTOR), and is a member of the DNA PK family.44 Sirolimus is more effective than CsA in inhibiting T-cell proliferation, although it does not alter IL-2 or IL-2 receptor expression.31 Sirolimus, unlike CsA and FK506, has been shown to inhibit IL-2–dependent T-cell proliferation following binding of the cytokine to its receptor. The mTOR appears to play an important role in the control of translation and amino acid transport for a wide variety of cells. Binding of the sirolimus complex to mTOR inhibits the phosphorylation of p70 S6 kinase, the subsequent phosphorylation of S6 ribosomal protein, and protein synthesis.15 Amino acid transport and an inhibitor of translational initiation (4E-BP1) are also sensitive to sirolimus.14 Cyclosporin and tacrolimus are approved for use in organ and stem cell transplantation and sirolimus in the prevention of acute renal transplant rejection. Their use in autoimmune disorders is currently being explored, and there is already considerable experience for the use of cyclosporin. Adverse Effects The unwanted side effects of cyclosporin and tacrolimus are very similar and consist most commonly of nephrotoxicity, hypertension, and neurotoxicity. Nephrotoxicity is dose dependent, and it is important to monitor plasma CsA and serum creatinine levels during therapy. Should

there be any rise in creatinine levels, creatinine clearance should be measured. Less severe side effects include hepatotoxicity, lethargy, anorexia, hirsutism, and gum hypertrophy. Neurotoxicity may be manifested by headache, tremor, sleep disturbance, or a reversible, mainly posterior encephalopathy. These complications indicate the need for dose reduction or the drug’s discontinuation. As may be predicted from the differences described earlier between molecular targets of sirolimus and cyclosporin, the toxicity profile of sirolimus differs from that of tacrolimus and cyclosporin. Nephrotoxicity, hypertension, and neurotoxicity are not prominent but hyperlipidemia, leukopenia, and thrombocytopenia are more frequent. Use in Neuropathy There are no controlled trials of these agents in neuropathy.53,110 Cyclosporin has been used in refractory CIDP9,46 and vasculitis of nerve, but the side effect profile prevents its use as a first-line agent. The efficacy of tacrolimus and sirolimus in human autoimmune neuropathy has yet to be explored in appropriate trials. Beneficial response to tacrolimus has been reported in two patients with CIDP.2,84 Tacrolimus and sirolimus and their derivatives have been reported to stimulate nerve regeneration, a property that would be of value in the treatment of immune-mediated neuropathy.104

Azathioprine Azathioprine (AZA) was first synthesized in 1957, and has been an effective immunosuppressive agent in the treatment of autoimmune and neuroimmunologic disorders but particularly in immunomodulation of transplant recipients. Azathioprine has particular efficacy in the treatment of myasthenia gravis. Mechanism of Action Azathioprine is converted to 6-mercaptopurine (6MP), a purine analogue of hypoxanthine, and guanine, which in turn are converted to the active metabolites 6-thionosinic acid and 6-thioguanilic acid. These metabolites interfere with purine biosynthesis and reduce RNA and DNA synthesis. Azathioprine decreases T and B lymphocytes and antibody production and B cell proliferation.3,49 Natural killer cell activity and neutrophil migration are also decreased. Pharmacokinetics AZA is absorbed from the alimentary tract and distributed throughout the body, 30% being bound to plasma protein. It has a half-life of 4 to 5 hours, whereas that of 6MP is only 20 to 50 minutes. The majority of AZA is converted by xanthine oxidase to the inactive metabolite 6-thiouric acid, which is excreted by the kidney. Patients receiving treatment

640

Neuroimmunology of the Peripheral Nervous System

with allopurinol are at increased risk of toxicity because it suppresses xanthine oxidase, resulting in increased levels of AZA and its metabolites. In such patients the dose of AZA should be reduced by 50% to 70% and the blood count carefully monitored. Adverse Effects The incidence of serious adverse events with AZA is relatively low. The most frequent side effects reported in a group of 104 patients with severe myasthenia were reversible marrow depression, with leukopenia, gastrointestinal complications, infections, and transient elevation of liver enzymes.47 In 546 patients with rheumatoid arthritis, 60% of therapy terminations were due to gastrointestinal problems.103 These include nausea, vomiting, diarrhea, and abdominal pain. Hepatotoxicity manifests by cholestasis, hypersensitivity, or hepatic necrosis and has been reported in up to 10% of cases, but pancreatitis is a relatively rare event. Cessation of therapy is indicated in the presence of hepatotoxicity or pancreatitis. Bone marrow suppression, although more common, is dose dependent and usually responds to dose reduction or cessation of therapy. About 10% of the population is deficient in the enzyme thiopurine5-methyltransferase, with resultant reduced AZA metabolism and increased toxicity. One in 300 is homozygous for deficiency of this enzyme and at severe risk of major side effects from a single dose. Levels of this enzyme can now be measured, and measurement before starting treatment is helpful for selecting the correct dose.48 Mild intestinal symptoms are relatively common and can usually be minimized by dividing the dose and taking it after meals. Elevation of liver enzymes of mild degree, up to three times normal, is also commonly seen and may respond to dose reduction. Leukopenia induced by AZA increases the risk of opportunistic infections. There is an increased risk of lymphoma in renal transplant patients treated with AZA and corticosteroids, and also an increased risk of skin cancers. The latter are seen in patients treated for myasthenia gravis, but the risk of other malignancies appears to be less than in transplant patients.118 Use in Neuropathy Azathioprine is usually given in a dose of 2.5 mg/kg daily. In the treatment of myasthenia, a stepwise dose reduction to a maintenance dose of 0.8 mg/kg/day is usually advised once improvement has been maintained for 6 to 9 months. There is no information in the neuropathy literature on such changes to maintenance dosage. Complete blood counts and liver function tests should be monitored every 2 weeks for the first 8 weeks and then once every 4 weeks. If the WBC count falls below 3.0 ⫻ 106 cells/L, AZA should be discontinued for a few days and treatment introduced at a lower dose when the count rises above 3.5 ⫻ 106.

Azathioprine is widely used in patients with CIDP28 to lessen the need for IVIg and PE,42,92 although its efficacy has never been proven by randomized controlled trials. In the only controlled study reported, AZA (2.0 mg/kg) was not shown to be of additional benefit when combined with corticosteroid in a small study of CIDP.34,53

Mycophenolate Mofetil This relatively recently introduced compound was isolated from Penicillium mold cultures in 1913, but was developed for use in transplantation only in the 1980s. It appears to be marginally more effective than AZA in transplantation medicine.74 Mechanism of Action Mycophenolate mofetil is hydrolyzed in the liver to its active parent compound, mycophenolic acid, which is a noncompetitive inhibitor of inositol monophosphate dehydrogenase, an enzyme central to the synthesis of guanosine nucleotides and hence purines. Proliferating lymphocytes depend primarily on the de novo pathway for purine synthesis, whereas most other cells can use both the de novo and salvage pathways. Thus mycophenolate has a selective immunosuppressive effect on proliferating lymphocytes. Mycophenolate inhibits T-cell and B-cell proliferation to both T-cell–dependent and T-cell–independent mitogens.18 The incorporation of mannose and fucose into membrane glycoproteins such as adhesion molecules appears to be dependent on guanosine levels, which may explain the demonstrated effect of mycophenolate to reduce the adherence of leukocytes to endothelial cells. Pharmacokinetics Following absorption from the gastrointestinal tract, mycophenolate is completely hydrolyzed to the active form, mycophenolic acid. Peak serum concentration occurs at 1 hour, with a secondary peak 8 hours later. The biologic half-life is 17 hours and the bioavailability 94%. Mycophenolic acid is metabolized by the liver to mycophenolic glucuronide, which is excreted in the urine. Adverse Effects The most common side effects are gastrointestinal, including nausea, vomiting, diarrhea, cramping, and abdominal pain. Gastrointestinal hemorrhage, pancreatitis, leukopenia, and thrombocytopenia have been reported. Hepatotoxicity appears to be low compared to AZA.74 Side effects are mostly dose dependent and respond to lowering the dose or stopping the drug. From relatively brief experience with the use of this drug in rheumatoid arthritis and

Principles of Immunotherapy

transplantation, no increased incidence of carcinoma has been reported. Use in Neuropathy Mycophenolate is usually commenced in a daily divided dose of 2 g, but may be increased to 4 or 5 g. The dose should be reduced or the drug discontinued in the presence of neutropenia. Mycophenolate was shown to be slightly superior to AZA in a multicenter randomized trial in heart transplant patients in reducing mortality and acute rejection.65 However, there have been no controlled trials to date of mycophenolate in autoimmune neurologic disease, including neuropathy. Small case series have claimed both benefit and the lack of it when given to patients with CIDP.19,53,84 Blood counts should be monitored weekly for the first month, then twice monthly for 2 months, and then monthly.

IMMUNOMODULATORY AGENTS Therapeutic Plasma Exchange Therapeutic PE (plasmapheresis) was first shown to have clinical benefit in 1960 when Schwab and Fahey demonstrated alleviation of the symptoms of hyperviscosity in patients with Waldenström’s macroglobulinemia. Application to autoimmune disorders characterized by the presence of pathogenic antibodies or other serum components followed in the next two decades. The common method of PE utilizes cell separators, which are automated centrifugation devices. Blood is withdrawn from the patient, separated into its components, and returned following removal of one or more of its components. Later techniques employed a membrane filtration system to remove plasma from whole blood during extracorporeal perfusion. Most recently, selective immunoabsorption techniques utilizing tryptophanlinked polyvinyl alcohol have been developed. Whereas the former techniques require protein replacement (i.e., plasma or albumen), this is not required with selective immunoabsorption. The schedule of PE varies among different centers and often within a single center for different diseases. If the amount of a substance catabolized during exchange is equal to the sum of the amount synthesized and the amount shifted from the extravascular to the intravascular compartment, each procedure in which the volume exchanged approaches the patient’s plasma volume will remove approximately 50% to 60% of an intravascular substance. It is commonly believed that four to five such exchanges over 7 to 10 days constitutes adequate short-term therapy.102 Hence, according to a GBS study group in 1985, in the major GBS trials a total of 200 to 250 mL of plasma per kilogram of body weight was exchanged over 7 to 10 days.94 Because the efficiency of removing a particular serum

641

factor is greatest when the serum level is highest, some centers prefer to exchange smaller volumes on a more frequent basis (i.e., 2-L exchanges performed 10 times over 10 to 14 days).91 Unless the pathogenic factor is known, no ideal regimen of PE can be devised because plasma levels cannot be measured. PE may be of therapeutic benefit in those conditions in which a plasma constituent is responsible for disease manifestations. Such factors include autoantibodies (i.e., antibody to the acetylcholine receptor), toxins (i.e., phytanic acid in Refsum’s disease), and monoclonal antibodies (contributing to hyperviscosity in Waldenström’s macroglobulinemia). Other plasma components that may contribute to the disease pathogenesis, including cytokines and complement, are also removed by PE. The first application of PE in neurologic disease was in myasthenia gravis, an autoimmune disorder in which pathogenic antibodies play a central role. There is now substantial evidence of autoantibody involvement in several neuropathies, including GBS and its subsets, CIDP, IgM paraproteinemic-associated neuropathies, and possibly MMN (see Chapters 98, 99, and 101). However, the presence of pathogenic antibodies does not guarantee a response to PE, because antibodies may induce irreversible damage (i.e., neuronal loss). Adverse Events Potential adverse events identified by the National Institutes of Health Consensus Development Conference Statement (1960) include • Allergic reactions leading to anaphylaxis • Replacement with fluids depleted of coagulation factors, proteins, or electrolytes • Citrate-induced hypocalcemia • Replacement fluids containing plasma, which have the potential to transmit infection (e.g., human immunodeficiency virus [HIV] or hepatitis). Most centers now employ albumin or artificial plasma replacement to avoid risks associated with plasma. • Hemorrhage secondary to the use of systemic anticoagulants or dilution of coagulation factors • Activation of coagulation, complement, the fibrinolytic cascade, and/or the aggregation of platelets. Hence thrombophlebitis, thromboses, and pulmonary embolism have been observed. • Fluid imbalance • Vessel perforation and, in patients given arteriovenous shunts and central venous catheters, thrombosis and infection (bacterial endocarditis) may occur. Elderly patients with vascular instability carry an increased risk of severe complications. However, in centers expert in its use, PE is a safe procedure with fewer long-term side effects than corticosteroids and immunosuppressive agents. Complications were reported in 17% of 381 procedures in one series.22

642

Neuroimmunology of the Peripheral Nervous System

Use in Neuropathy Three large multicenter randomized controlled trials and a systematic review have shown that PE is effective therapy in GBS if started within the first 4 weeks of neurologic symptoms94 (see Chapter 98). The efficacy of PE in CIDP has also been shown by randomized trials32,43 and case series.20 Although the cause of GBS and CIDP remains uncertain, there is mounting evidence of a role for antibody and complement in certain subtypes of GBS (Miller Fisher syndrome and acute motor axonal neuropathy; see Chapter 26) and in some cases of CIDP,120 thus providing a rationale for PE therapy. PE has not proven of value in MMN, and it may aggravate this condition.86 Patients with neuropathy associated with immunoglobulin A (IgA) and immunoglobulin G (IgG) paraproteins respond to treatment with PE, whereas those with neuropathies associated with IgM paraproteins did not benefit in the one controlled study of PE in neuropathy associated with monoclonal gammopathy of uncertain significance (MGUS).32 Several uncontrolled studies have shown improvement when the IgM level is decreased by 50%,101 but because IgM levels rise rapidly following PE, it needs to be accompanied by some form of immunosuppression.85 PE together with dietary control remains essential in the maintenance treatment of Refsum’s disease (see Chapter 79). PE therapy has largely been replaced by IVIg therapy in patients with inflammatory demyelinating neuropathy on the grounds of convenience, avoidance of the necessity for specialized units and staff, greater ease of administration, and lesser occurrence of adverse events.

Intravenous Immunoglobulin The response of coincidental thrombocytopenia to treatment with IVIg being given for immune deficiency led to the introduction of IVIg for autoimmune thrombocytopenia.57 By a similar serendipitous observation, IVIg given for thrombocytopenia benefited a patient with CIDP115 and was introduced for the treatment of this and other potentially autoimmune diseases. The neurologic and neuromuscular disease indications for the use of IVIg have been reviewed by experts and expert panels25,70,119 and national guidelines have been promulgated.7 The immunoglobulin used in IVIg is prepared from the pooled plasma of several thousand donors.25,60 The method of preparation varies from product to product. The basic principle is that the gamma globulin is precipitated from the plasma by cold ethanol. The precipitate is gently treated at low pH with enzymes to avoid aggregation and remove impurities while preserving antibody activity. The final product consists of IgG monomer with some dimer and variable but small amounts of IgA (less than 2.5%) and traces of IgM. Some brands have particularly low levels of IgA. Viral safety is secured by screening the donors for risk

factors and testing their plasma for viral antigens and antibodies to eliminate hepatitis B, HIV, and other viruses. Viral elimination steps are also built into the manufacturing process so that the transmission of viral infection, although still a theoretical risk and the subject of active surveillance, has not been reported with current products. The final product is presented freeze dried or already dissolved as a solution with a stabilizer such as sucrose, maltose, or glycine. The distribution of IgG subclasses is approximately the same as that in plasma, and the full range of antibody activities is preserved. The inclusion of such a large donor population is thought to enhance the likelihood of the pool containing therapeutically useful anti-idiotypic antibodies that will inhibit the action of harmful autoantibodies. Administration The standard course first used for autoimmune thrombocytopenia was 0.4 g/kg daily for 5 days, and this has been adopted in neuromuscular diseases. However, for recurrent courses we and many others use 2.0 g/kg as a continuous infusion or as infusions of 1.0 g/kg on two consecutive days, thus avoiding the need for an overnight stay. We adhere to the manufacturers’ guidelines for the rate of infusion. In a recently reported series, patients preferred a faster rate of infusion (up to a maximum of 800 mL/hr instead of the usual 200 mL/hr) despite the increased risk of adverse events.41 If the patient is known to develop possibly allergic side effects during IVIg infusion and, as an exception, it has been decided to give IVIg again, we give chlorpheniramine and hydrocortisone intravenously 30 minutes before starting the infusion. During the infusion the patient should be under observation, especially during the first 30 minutes when the temperature, pulse, respiratory rate, and blood pressure should be monitored. Observations should be maintained throughout the infusion. If the patient develops side effects, we stop the infusion until the side effects subside and it may be possible to restart at a slower rate. Provision for treating anaphylaxis should be immediately available. Because many of the neurologic uses of IVIg are unlicensed, we always obtain signed informed consent before administering the treatment. Because of the theoretical risks, we also obtain such consent from patients with GBS, for whom the treatment is licensed. Our information sheet mentions the theoretical risks of transmission of viral and other infectious agents, including Creutzfeldt-Jakob disease. Mechanisms of Action Many different mechanisms of action have been proposed for IVIg (Table 28–1). It is likely that the actual mechanisms are multiple and the principal mechanism varies from disease to disease.60 Most of the evidence for these mechanisms comes from non-neurologic diseases, and only a few mechanisms have been investigated in peripheral nerve disorders.26

Principles of Immunotherapy

Table 28–1. Mechanisms of Action of Intravenous Immunoglobulin Autoantibodies Anti-idiotype blockade Increased breakdown of autoantibodies by blockade of endothelial FcRn transport receptors Suppression of antibody production Fc receptor Blockade of Fc␥ receptor III Stimulation of inhibitory Fc␥ receptor II Complement Prevention of membrane attack complex formation Cytokines Suppression of cytokine release Neutralization of circulating cytokines T cells Modulation of T-cell function Remyelination Possible enhancement of remyelination

Anti-idiotypic binding of IgG to autoantibodies has been shown to inhibit the action of autoantibodies against several systemic autoantigens and also against ganglioside GM1 in MMN,77,122 GQ1b in Miller Fisher syndrome, and neuroblastoma cells in CIDP.112 Immunoglobulin has also been shown to inhibit the neuromuscular blockade produced by GBS serum in a mouse phrenic nerve–diaphragm motor end-plate macropatch-clamp preparation.16 In a negative feedback loop, IVIg may engage Fc receptors on the B-cell surface and downregulate B-cell activation.29 Additionally, IVIg may saturate a protective endosomal receptor, FcRn, in endothelial cells that normally sequesters and protects IgG. If the receptor sites are blocked by the excess of IgG, then autoantibodies may be more exposed to catabolism in the lysosomes.121 IVIg is effective in idiopathic thrombocytopenia by blockade of the Fc receptor on macrophages. Antibodies to the Fc␥ receptor III or the Fc fraction of IgG are also effective. In addition, IgG has been shown to block the Fc receptors on human monocytes in vitro.69 In a murine model of autoimmune thrombocytopenia, the efficacy of IVIg in treating the lowered platelet count was shown in a knockout model to depend on stimulation of an inhibitory Fc␥ receptor II.97 However Fc␥ receptor III blockade on macrophages would inhibit macrophage-associated demyelination and provide a plausible explanation for the efficacy of IVIg in inflammatory neuropathy. Consistent with this hypothesis, there is a modulation of Fc␥II/Fc␥III receptors on the surface of circulating monocytes following IVIg23 such that the proportion of inhibitory Fc␥II receptors is increased relative to the active Fc␥III receptors. In dermatomyositis IVIg prevents the formation of complement membrane attack complex, possibly by binding of C3b to the Fc on IgG.11 Because complement-fixing

643

antibodies to myelin and deposition of complement at the myelin surface have been found in acute inflammatory demyelinating polyradiculoneuropathy and CIDP, this property of IVIg may also be important in these conditions. Human immune globulin contains antibodies against many cytokines.105 These may be partly responsible for the fall in levels of circulating cytokines after the administration of IVIg.26,60 In the acute stage of GBS the serum concentrations of TNF-␣ are increased but they fall following the administration of IVIg.99,100 There is also evidence that IVIg regulates T-cell function. Human immune globulin contains antibodies against the T-cell receptor that may be partly responsible.78 It also contains antibodies to superantigens that directly stimulate T cells. Binding to staphylococcal superantigen is the probable mechanism of action in Kawasaki disease.72 The expression of the important adhesion molecule ICAM on the surface of T cells is downregulated following IVIg, which may reduce their passage into the endoneurium.23 Indications Guillain-Barré Syndrome. According to a Cochrane systematic review, there are no adequate trials comparing IVIg with placebo, but IVIg has equivalent efficacy to PE in hastening recovery from severe GBS when treated within 2 weeks from onset of the disease.50 PE costs about the same, has more side effects, and is obviously less convenient. This led the American Academy of Neurology Practice Parameter Group to recommend consideration of IVIg for patients within 2 weeks and probably within 4 weeks after the onset of symptoms.51 It is not known whether a second course of IVIg is effective in patients who remain severely affected 2 or 3 weeks after the first. It is also not known whether IVIg is effective in children and in patients who have atypical forms of GBS, such as Miller Fisher syndrome, although it is often used in these situations. About 10% of patients relapse between 2 and 10 weeks after their first course of IVIg, an important consideration when patients are discharged to rehabilitation, and a second course is often given. Chronic Inflammatory Demyelinating Polyradiculoneuropathy. A Cochrane systematic review concluded that randomized controlled trials show that IVIg improves disability for at least 2 to 6 weeks compared with placebo.114 This is consistent with the general experience that between 60% and 80% of patients respond to IVIg.42,81 One trial showed that it has similar efficacy to PE,32 and another that it has similar efficacy to oral prednisolone.50 Because these three treatments seem to be equally effective, it is uncertain which of them should be the first choice. The disadvantages of IVIg are that it has to be repeated and 60% of patients require treatment for more than 6 months and 50% for more than a year. The frequency of repeated treatments ranges from 2 to 12 weeks. Despite the strength of the evidence and

644

Neuroimmunology of the Peripheral Nervous System

the conclusions of the Cochrane review, IVIg is not currently licensed for use in CIDP in the United States or the United Kingdom. Multifocal Motor Neuropathy or Pure Motor CIDP. Three small randomized trials have shown that MMN responds to treatment with IVIg in terms of an increase in strength, but it is not clear that there is a commensurate improvement in disability.111 Because steroids and PE are ineffective and immunosuppressant agents unproven, IVIg is the current treatment of choice. About two thirds of patients respond and require treatment every 2 to 12 weeks, which is inconvenient and expensive. The pure motor form of CIDP responds to IVIg and may be made worse by steroids,30 so IVIg is also the preferred first-line treatment. Paraproteinemic Demyelinating Neuropathy. Patients with IgA or IgG MGUS and demyelinating neuropathy often resemble those with CIDP and show a similar response to IVIg, although no randomized trial has ever been done. In two randomized controlled trials only a minority of patients with IgM MGUS and demyelinating neuropathy responded to IVIg.76 Other Neuropathies. There are anecdotal reports of improvement following IVIg in vasculitic neuropathy,58 diabetic proximal motor neuropathy, acute sensory neuronopathy,83 and paraneoplastic sensory neuropathy.109 No controlled trials or substantial series have been reported, and it is likely that such responses are rare and possible that they are coincidental. Adverse Events Minor side effects are common but serious adverse events are rare (Table 28–2).25,35,55,66,70,119 During standard rates of infusion, at least 10% of patients experience headache, back and limb pain, chest pain, flushing, or nausea, but these symptoms are usually tolerable or respond to slowing the infusion rate. The treatment may trigger migraine that can sometimes be prevented with propranolol,21 and in one series of 54 patients aseptic meningitis was reported in Table 28–2. Risks of Intravenous Immunoglobulin Common (5%–25%)

Uncommon (⬍5%)

Rare (⬍1%)

Very rare (⬍0.01%)

Headache, malaise, fever, myalgia, hypotension, lymphopenia, increased liver enzymes Meningism, urticaria, pompholyx, eczema (especially on the hands), neutropenia, renal failure Cerebral vasospasm, stroke, myocardial infarction, pulmonary embolism, pancytopenia, hemolytic anemia, alopecia Anaphylaxis, hepatic veno-occlusive disease, multiorgan failure

11%.98 The frequency of aseptic meningitis was not related to the brand of IVIg but was greater in those patients with a history of migraine. It subsided spontaneously after between 3 and 5 days. In one case aseptic meningitis was associated with reversible cerebral vasospasm causing ischemia116; the cause is not clear. With faster rates of infusion (up to 800 mL/hr), adverse effects are more common. They occurred in 38 of 50 consecutive patients (26% of 341 infusions) treated with a 5% solution given at up to 800 mL/hr.41 Eleven (22%) had major events but all recovered completely. Various skin reactions occur with IVIg and affected 7 (6%) of 120 patients in one series. During the infusion, flushing and urticaria may occur. These reactions may be related to particular batches. After the infusion, a vesicular rash on the hands and eczema have been reported. Renal failure has also been described.25 It has been attributed to sucrose in the IVIg preparation causing osmotic damage to the renal tubules. However, 8 (6.7%) of 119 patients (287 courses) developed renal impairment, irreversible in 2, and the occurrence of this side effect was independent of the sucrose content of the preparation.73 Patients should be checked for renal impairment before treatment and IVIg should be withheld if present. If it is given despite the presence of renal impairment, a non–sucrose-containing n brand should be used and renal function carefully monitored. Rare cases of stroke, myocardial infarction, and pulmonary embolism have been reported after IVIg. Their cause might be related to the increase in serum viscosity induced by IVIg.17,25,27,87 These occurrences have led to recommendations that extra caution should be exercised for patients who are suspected to have raised serum viscosity or cardiovascular or cerebrovascular disease or who are at risk of pulmonary embolism, but there is no consensus about the definition of these risk factors. Dalakas and Clark27 recommend caution in the presence of paraproteinemia, hyperfibrinogenemia, and hypercholesterolemia. Common sense dictates the use of a slower infusion rate, such as 0.4 g/kg over 5 days. Measuring plasma viscosity to identify patients who are at risk and lowering the viscosity if it is raised has been proposed.27 The use of prophylactic antiplatelet agents or low-dose heparin in patients with a history of stroke, myocardial infarction, or deep vein thrombosis has not been evaluated.

REFERENCES 1. Adcock, I. M., Brown, C. R., Gelder, C. M., et al.: Effects of glucocorticoids on transcription factor activation in human peripheral blood mononuclear cells. Am. J. Physiol. Cell Physiol. 268:C331, 1995. 2. Ahlmen, J., Andersen, O., Hallgren, G., and Peilot, B.: Positive effects of tacrolimus in a case of CIDP. Transplant. Proc. 30:4194, 1998.

Principles of Immunotherapy 3. Anstey, A., and Lear, J. T.: Azathioprine: Clinical pharmacology and current indications in autoimmune disorders. Biodrugs 33, 1998. 4. Archelos, J. J., Previtali, S. C., and Hartung, H. P.: The role of integrins in immune-mediated diseases of the nervous system. Trends Neurosci. 22:30, 1999. 5. Atkinson, J. P., and Frank, M. M.: Complement-independent clearance of IgG-sensitized erythrocytes: inhibition by cortisone. Blood 44:629, 1974. 6. Auphan, N., DiDonato, J. A., Rosette, C., et al.: Immunosuppression by glucocorticoids: inhibition of NF-kappa B activity through induction of I kappa B synthesis. Science 270:286, 1995. 7. Barnes, P., Hughes, R., Lecky, B., et al.: Association of British Neurologists Guidelines for the Use of Intravenous Immunoglobulin in Neurological Diseases. London, Association of British Neurologists, 2002. 8. Barnes, P. J., Greening, A. P., and Crompton, G. K.: Glucocorticoid resistance in asthma. Am. J. Respir. Crit. Care Med. 152:S125, 1995. 9. Barnett, M. H., Pollard, J. D., Davies, L., and McLeod, J. G.: Cyclosporin A in resistant chronic inflammatory demyelinating polyradiculoneuropathy. Muscle Nerve 21:454, 1998. 10. Bartoszek, M., Brenner, A. M., and Szefler, S. J.: Prednisolone and methylprednisolone kinetics in children receiving anticonvulsant therapy. Clin. Pharmacol. Ther. 42:424, 1987. 11. Basta, M., and Dalakas, M. C.: High-dose intravenous immunoglobulin exerts its beneficial effect in patients with dermatomyositis by blocking endomysial deposition of activated complement fragments. J. Clin. Invest. 94:1729, 1994. 12. Begg, E. J., Atkinson, H. C., and Gianarakis, N.: The pharmacokinetics of corticosteroid agents. Med. J. Aust. 146:37, 1987. 13. Bensa, S., Hadden, R. D., Hahn, A., et al.: Randomized controlled trial of brain-derived neurotrophic factor in Guillain-Barre syndrome: a pilot study. Eur. J. Neurol. 7:423, 2000. 14. Berretta, L., Gingras, A. C., Svithin, Y. V., et al.: Rapamycin blocks the phosphorylation of 4E-BPI and inhibits capdependent initiation of translation. EMBO J. 15:658, 1996. 15. Brown, E. J., Beal, P. A., Keith, C. T., et al.: Control of p70 s6 kinase by kinase activity of FRAP in vivo. Nature 377:441, 1995. [erratum appears in Nature 378:644, 1995.] 16. Buchwald, B., Ahangari, R., Weishaupt, A., and Toyka, K. V.: Intravenous immunoglobulins neutralize blocking antibodies in Guillain-Barre syndrome. Ann. Neurol. 51:673, 2002. 17. Caress, J. B., Cartwright, M. S., Donofrio, P. D., and Peacock, J. E. Jr.: The clinical features of 16 cases of stroke associated with administration of IVIg. Neurology 60:1822, 2003. 18. Chang, C. C., Aversa, G., Punnonen, J., et al.: Brequinar sodium, mycophenolic acid, and cyclosporin A inhibit different stages of IL-4- or IL-13-induced human IgG4 and IgE production in vitro. Ann. N. Y. Acad. Sci. 696:108, 1993. 19. Chaudhry, V., Cornblath, D. R., Griffin, J. W., et al.: Mycophenolate mofetil: a safe and promising immunosuppressant in neuromuscular diseases. Neurology 56:94, 2001. 20. Choudhary, P. P., and Hughes, R. A.: Long-term treatment of chronic inflammatory demyelinating polyradiculoneuropathy with plasma exchange or intravenous immunoglobulin. QJM 88:493, 1995.

645

21. Constantinescu, C. S., Chang, A. P., and McCluskey, L. F.: Recurrent migraine and intravenous immune globulin therapy. N. Engl. J. Med. 329:583, 1993. 22. Couriel, D., and Weinstein, R.: Complications of therapeutic plasma exchange: a recent assessment. J. Clin. Apheresis 9:1, 1994. 23. Creange, A., Gregson, N. A., and Hughes, R. A.: Intravenous immunoglobulin modulates lymphocyte CD54 and monocyte FcgammaRII expression in patients with chronic inflammatory neuropathies. J. Neuroimmunol. 135:91, 2003. 24. Cuhaci, B., Kumar, M. S., Bloom, R. D., et al.: Transforming growth factor-beta levels in human allograft chronic fibrosis correlate with rate of decline in renal function. Transplantation 68:785, 1999. 25. Dalakas, M. C.: Intravenous immunoglobulin in the treatment of autoimmune neuromuscular diseases: present status and practical therapeutic guidelines. Muscle Nerve 22:1479, 1999. 26. Dalakas, M. C.: Mechanisms of action of IVIg and therapeutic considerations in the treatment of acute and chronic demyelinating neuropathies. Neurology 59:S13, 2002. 27. Dalakas, M. C., and Clark, W. M.: Strokes, thromboembolic events, and IVIg: rare incidents blemish an excellent safety record. Neurology 60:1736, 2003. 28. Dalakas, M. C., and Engel, W. K.: Chronic relapsing (dysimmune) polyneuropathy: pathogenesis and treatment. Ann. Neurol. 9(Suppl.):134, 1981. 29. Diegel, M. L., Rankin, B. M., Bolen, J. B., et al.: Cross-linking of Fc gamma receptor to surface immunoglobulin on B cells provides an inhibitory signal that closes the plasma membrane calcium channel. J. Biol. Chem. 269:11409, 1994. 30. Donaghy, M., Mills, K. R., Boniface, S. J., et al.: Pure motor demyelinating neuropathy deterioration after steroid therapy and improvement with intravenous immunoglobulin. J. Neurol. Neurosurg. Psychiatry 57:778, 1994. 31. Dumont, F. J., Staruch, M. J., Koprak, S. L., et al.: Distinct mechanisms of suppression of murine T cell activation by the related macrolides FK-506 and rapamycin. J. Immunol. 144:251, 1990. 32. Dyck, P. J., Low, P. A., Windebank, A. J., et al.: Plasma exchange in polyneuropathy associated with monoclonal gammopathy of undetermined significance. N. Engl. J. Med. 325:1482, 1991. 33. Dyck, P. J., O’Brien, P. C., Oviatt, K. F., et al.: Prednisone improves chronic inflammatory demyelinating polyradiculoneuropathy more than no treatment. Ann. Neurol. 11:136, 1982. 34. Dyck, P. J., O’Brien, P. C., Swanson, C., et al.: Combined azathioprine and prednisone in chronic inflammatorydemyelinating polyneuropathy. Neurology 35:1173, 1985. 35. Eijkhout, H. W., and van Aken, W. G.: Blood, blood components and plasma products. In Side Effects of Drugs Annual. Amsterdam, Elsevier, p. 398, 2004. 36. Fauci, A. S., Katz, P., Haynes, B. F., and Wolff, S. M.: Cyclophosphamide therapy of severe systemic necrotizing vasculitis. N. Engl. J. Med. 301:235, 1979. 37. Feldman, E. L., Bromberg, M. B., Albers, J. W., and Pestronk, A.: Immunosuppressive treatment in multifocal motor neuropathy. Ann. Neurol. 30:397, 1991.

646

Neuroimmunology of the Peripheral Nervous System

38. Felts, P. A., Smith, K. J., Gregson, N. A., and Hughes, R. A.: Brain-derived neurotrophic factor in experimental autoimmune neuritis. J. Neuroimmunol. 124:62, 2002. 39. Frohnert, P. P., and Sheps, S. G.: Long-term follow-up study of periarteritis nodosa. Am. J. Med. 43:8, 1967. 40. Good, J. L., Chehrenama, M., Mayer, R. F., and Koski, C. L.: Pulse cyclophosphamide therapy in chronic inflammatory demyelinating polyneuropathy. Neurology 51:1735, 1998. 41. Grillo, J. A., Gorson, K. C., Ropper, A. H., et al.: Rapid infusion of intravenous immune globulin in patients with neuromuscular disorders. Neurology 57:1699, 2001. 42. Hahn, A. F., Bolton, C. F., Pillay, N., et al.: Plasma-exchange therapy in chronic inflammatory demyelinating polyneuropathy: a double-blind, sham-controlled, cross-over study. Brain 119:1055, 1996. 43. Hahn, A. F., Bolton, C. F., Zochodne, D., and Feasby, T. E.: Intravenous immunoglobulin treatment in chronic inflammatory demyelinating polyneuropathy: a double-blind, placebocontrolled, cross-over study. Brain 119:1067, 1996. 44. Heitman, J., Movva, N. R., and Hall, M. N.: Targets for cell cycle arrest by the immunosuppressant rapamycin in yeast. Science 253:905, 1991. 45. Hengst, J. C., and Kempf, R. A.: Immunomodulation by cyclophosphamide. Clin. Immunol. Allergy 4:199, 1984. 46. Hodgkinson, S. J., Pollard, J. D., and McLeod, J. G.: Cyclosporin A in the treatment of chronic demyelinating polyradiculoneuropathy. J. Neurol. Neurosurg. Psychiatry 53:327, 1990. 47. Hohlfeld, R., Michels, M., Heininger, K., et al.: Azathioprine toxicity during long-term immunosuppression of generalized myasthenia gravis. Neurology 38:258, 1988. 48. Holme, S. A., Duley, J. A., Sanderson, J., et al.: Erythrocyte thiopurine methyl transferase assessment prior to azathioprine use in the UK. QJM 95:439, 2002. 49. Hortelano, S., and Bosca, L.: 6-Mercaptopurine decreases the Bcl-2/Bax ratio and induces apoptosis in activated splenic B lymphocytes. Mol. Pharmacol. 51:414, 1997. 50. Hughes, R., Bensa, S., Willison, H., et al.: Randomized controlled trial of intravenous immunoglobulin versus oral prednisolone in chronic inflammatory demyelinating polyradiculoneuropathy. Ann. Neurol. 50:195, 2001. 51. Hughes, R., Wijdicks, E., Barohn, R. J., et al.: Practice parameter: immunotherapy for Guillain-Barre syndrome. Report of the Quality Standards Subcommittee of the American Academy of Neurology. Neurology 61:736, 2003. 52. Hughes, R. A.: Treatment of Guillain-Barre syndrome with corticosteroids: lack of benefit? Lancet 363:181, 2004. 53. Hughes, R. A., Swan, A. V., and van Doorn, P. A.: Cytotoxic drugs and interferons for chronic inflammatory demyelinating polyradiculoneuropathy. Cochrane Database Syst. Rev. CD003280, 2003. 54. Hughes, R. A., and van der Meche, F. G.: Corticosteroids for treating Guillain-Barre syndrome. Cochrane Database Syst. Rev. CD001446, 2000. [update in Cochrane Database Syst. Rev. 3:CD001446; 2000.] 55. Hughes, R. A. C.: Guidelines for the Use of Beta Interferons and Glatiramer Acetate in Multiple Sclerosis. London, Association of British Neurologists, p. 1, 2001. 56. Hull, K. M., and Boumpas, D. T.: Immunomodulating pharmaceuticals. In Rich, R. R., Fleisher, T. A., Shearer, W. T.,

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68. 69.

70.

71.

72.

73.

et al. (eds.): Clinical Immunology Principles and Practice, 2nd ed. St. Louis, C. V. Mosby, p. 1, 2001. Imbach, P., Barandun, S., d’Apuzzo, V., et al.: High-dose intravenous gammaglobulin for idiopathic thrombocytopenic purpura in childhood. Lancet 1:1228, 1981. Jayne, D. R., Davies, M. J., Fox, C. J., et al.: Treatment of systemic vasculitis with pooled intravenous immunoglobulin. Lancet 337:1137, 1991. Jones, T. B., Ankeny, D. P., Guan, Z., et al.: Passive or active immunization with myelin basic protein impairs neurological function and exacerbates neuropathology after spinal cord injury in rats. J. Neurosci. 24:3752, 2004. Kazatchkine, M. D., and Kaveri, S. V.: Immunomodulation of autoimmune and inflammatory diseases with intravenous immune globulin. N. Engl. J. Med. 345:747, 2001. Kerschensteiner, M., Stadelmann, C., Dechant, G., et al.: Neurotrophic cross-talk between the nervous and immune systems: implications for neurological diseases. Ann. Neurol. 53:292, 2003. Khanna, A. K., Cairns, V. R., Becker, C. G., and Hosenpud, J. D.: Transforming growth factor (TGF)-beta mimics and anti-TGF-beta antibody abrogates the in vivo effects of cyclosporine: demonstration of a direct role of TGF-beta in immunosuppression and nephrotoxicity of cyclosporine. Transplantation 67:882, 1999. Kieseier, B. C., Seifert, T., Giovannoni, G., and Hartung, H. P.: Matrix metalloproteinases in inflammatory demyelination: targets for treatment. Neurology 53:20, 1999. Kieseier, B. C., Tani, M., Mahad, D., et al.: Chemokines and chemokine receptors in inflammatory demyelinating neuropathies: a central role for IP-10. Brain 125:823, 2002. Kobashigawa, J., Miller, L., Renlund, D., et al.: A randomized active-controlled trial of mycophenolate mofetil in heart transplant recipients. Mycophenolate Mofetil Investigators. Transplantation 66:507, 1998. Koch, C.: Blood, blood components, plasma and plasma products. In Dukes, M. N. G., and Aronson, J. K. (eds.): Meyler’s Side Effects of Drugs. Amsterdam, Elsevier, p. 1123, 2000. Korn, T., Toyka, K., Hartung, H. P., and Jung, S.: Suppression of experimental autoimmune neuritis by leflunomide. Brain 124:1791, 2001. Koski, C. L.: Therapy of CIDP and related immune-mediated neuropathies. Neurology 59:S22, 2002. Kurlander, R. J.: Reversible and irreversible loss of Fc receptor function of human monocytes as a consequence of interaction with immunoglobulin G. J. Clin. Invest. 66:773, 1980. Latov, N., Chaudhry, V., Koski, C. L., et al.: Use of intravenous gamma globulins in neuroimmunologic diseases. J. Allergy Clin. Immunol. 108:S126, 2001. Laura, M., Gregson, N. A., Curmi, Y., and Hughes, R. A.: Efficacy of leukemia inhibitory factor in experimental autoimmune neuritis. J. Neuroimmunol. 133:56, 2002. Leung, D. Y., Burns, J. C., Newburger, J. W., and Geha, R. S.: Reversal of lymphocyte activation in vivo in the Kawasaki syndrome by intravenous gammaglobulin. J. Clin. Invest. 79:468, 1987. Levy, J. B., and Pusey, C. D.: Nephrotoxicity of intravenous immunoglobulin. QJM 93:751, 2000.

Principles of Immunotherapy 74. Lipsky, J. J.: Mycophenolate mofetil. Lancet 348:1357, 1996. 75. Liu, J., Farmer, J. D. Jr., Lane, W. S., et al.: Calcineurin is a common target of cyclophilin-cyclosporin A and FKBPFK506 complexes. Cell 66:807, 1991. 76. Lunn, M. P., and Nobile-Orazio, E.: Immunotherapy for IgM anti-myelin-associated glycoprotein paraproteinassociated peripheral neuropathies. Cochrane Database Syst. Rev. CD002827, 2003. 77. Malik, U., Oleksowicz, L., Latov, N., and Cardo, L. J.: Intravenous gamma-globulin inhibits binding of anti-GM1 to its target antigen. Ann. Neurol. 39:136, 1996. 78. Marchalonis, J. J., Kaymaz, H., Dedeoglu, F., et al.: Human autoantibodies reactive with synthetic autoantigens from T-cell receptor beta chain. Proc. Natl. Acad. Sci. U. S. A. 89:3325, 1992. 79. Marshall, S., Tardif, G., and Ashworth, N.: Local corticosteroid injection for carpal tunnel syndrome. Cochrane Database Syst. Rev. CD001554, 2002. [update in Cochrane Database Syst. Rev. 4:CD001554, 2000.] 80. Mehndiratta, M. M., and Hughes, R. A.: Corticosteroids for chronic inflammatory demyelinating polyradiculoneuropathy. Cochrane Database Syst. Rev. CD002062, 2002. [update in Cochrane Database Syst. Rev. 3:CD002062, 2001.] 81. Mendell, J. R., Barohn, R. J., Freimer, M. L., et al.: Randomized controlled trial of IVIg in untreated chronic inflammatory demyelinating polyradiculoneuropathy. Neurology 56:445, 2001. 82. Meucci, N., Cappellari, A., Barbieri, S., et al.: Long term effect of intravenous immunoglobulins and oral cyclophosphamide in multifocal motor neuropathy. J. Neurol. Neurosurg. Psychiatry 63:765, 1997. 83. Molina, J. A., Benito-Leon, J., Bermejo, F., et al.: Intravenous immunoglobulin therapy in sensory neuropathy associated with Sjogren’s syndrome. J. Neurol. Neurosurg. Psychiatry 60:699, 1996. 84. Mowzoon, N., Sussman, A., and Bradley, W. G.: Mycophenolate (CellCept) treatment of myasthenia gravis, chronic inflammatory polyneuropathy and inclusion body myositis. J. Neurol. Sci. 185:119, 2001. 85. Nobile-Orazio, E.: Neuropathies associated with anti MAG antibodies and IgM monoclonal gammopathies. In Latov, N., Wokke, J. H., and Kelly, J. I. (eds.): Immunological and Infectious Diseases of the Peripheral Nerves. Cambridge, UK, Cambridge University Press, p. 168, 1998. 86. Nobile-Orazio, E.: Multifocal motor neuropathy. J. Neuroimmunol. 115:4, 2001. 87. Okuda, D., Flaster, M., Frey, J., and Sivakumar, K.: Arterial thrombosis induced by IVIg and its treatment with tPA. Neurology 60:1825, 2003. 88. Ozer, H., Cowens, J. W., Colvin, M., et al.: In vitro effects of 4-hydroperoxycyclophosphamide on human immunoregulatory T subset function. I. Selective effects on lymphocyte function in T-B cell collaboration. J. Exp. Med. 155:276, 1982. 89. Pestronk, A., Cornblath, D. R., Ilyas, A. A., et al.: A treatable multifocal motor neuropathy with antibodies to GM1 ganglioside. Ann. Neurol. 24:73, 1988. 90. Pickup, M. E.: Clinical pharmacokinetics of prednisone and prednisolone. Clin. Pharmacokinet. 4:111, 1979.

647

91. Pollard, J. D.: Chronic inflammatory demyelinating polyradiculoneuropathy. Baillieres Clin. Neurol. 3:107, 1994. 92. Pollard, J. D., and Spies, J. M.: The Guillain Barre syndrome. In Asbury, A. K., McKhain, G. M., McDonald, W. I., et al. (eds.): Diseases of the Nervous System. Cambridge, UK, Cambridge University Press, p. 1110, 2002. 93. Rao, A., Luo, C., and Hogan, P. G.: Transcription factors of the NFAT family: regulation and function. Ann. Rev. Immunol. 15:707, 1997. 94. Raphael, J. C., Chevret, S., Hughes, R. A., and Annane, D.: Plasma exchange for Guillain-Barre syndrome. Cochrane Database Syst. Rev. CD001798, 2001. [update in Cochrane Database Syst. Rev. 2:CD001798, 2002.] 95. Sabin, T. D., Swift, T. R., and Jacobson, T. R.: Leprosy. In Dyck, P. J., Thomas, P. K., Griffin, J., et al. (eds.): Peripheral Neuropathy, 3rd ed. Philadelphia, W. B. Saunders, p. 1354, 1994. 96. Salinas, R. A., Alvarez, G., Alvarez, M. I., and Ferreira, J.: Corticosteroids for Bell’s palsy (idiopathic facial paralysis). Cochrane Database Syst. Rev. CD001942, 2002. 97. Samuelsson, A., Towers, T. L., and Ravetch, J. V.: Antiinflammatory activity of IVIG mediated through the inhibitory Fc receptor. Science 291:484, 2001. 98. Sekul, E. A., Cupler, E. J., and Dalakas, M. C.: Aseptic meningitis associated with high-dose intravenous immunoglobulin therapy: frequency and risk factors. Ann. Intern. Med. 121:259, 1994. 99. Sharief, M. K., Ingram, D. A., Swash, M., and Thompson, E. J.: IV immunoglobulin reduces circulating proinflammatory cytokines in Guillain-Barre syndrome. Neurology 52:1833, 1999. 100. Sharief, M. K., McLean, B., and Thompson, E. J.: Elevated serum levels of tumor necrosis factor-alpha in Guillain-Barre syndrome. Ann. Neurol. 33:591, 1993. 101. Sherman, W. H., Olarte, M. R., McKiernan, G., et al.: Plasma exchange treatment of peripheral neuropathy associated with plasma cell dyscrasia. J. Neurol. Neurosurg. Psychiatry 47:813, 1984. 102. Shumak, K. H., and Rock, G. A.: Therapeutic plasma exchange. N. Engl. J. Med. 310:762, 1984. 103. Singh, G., Fries, J. F., Spitz, P., and Williams, C. A.: Toxic effects of azathioprine in rheumatoid arthritis: a national post-marketing perspective. Arthritis Rheum. 32:837, 1989. 104. Steiner, J. P., Connolly, M. A., Valentine, H. L., et al.: Neurotrophic actions of nonimmunosuppressive analogues of immunosuppressive drugs FK506, rapamycin and cyclosporin A. Nat. Med. 3:421, 1997. 105. Svenson, M., Hansen, M. B., Ross, C., et al.: Antibody to granulocyte-macrophage colony-stimulating factor is a dominant anti-cytokine activity in human IgG preparations. Blood 91:2054, 1998. 106. Takahashi, N., Hayano, T., and Suzuki, M.: Peptidyl-prolyl cis-trans isomerase is the cyclosporin A-binding protein cyclophilin. Nature 337:473, 1989. 107. Thornhill, M. H., Kyan-Aung, U., and Haskard, D. O.: IL-4 increases human endothelial cell adhesiveness for T cells but not for neutrophils. J. Immunol. 144:3060, 1990. 108. Trushin, S. A., Pennington, K. N., Algeciras-Schimnich, A., and Paya, C. V.: Protein kinase C and calcineurin synergize to

648

109.

110.

111.

112.

113.

114.

115.

Neuroimmunology of the Peripheral Nervous System activate IkappaB kinase and NF-kappaB in T lymphocytes. J. Biol. Chem. 274:22923, 1999. Uchuya, M., Graus, F., Vega, F., et al.: Intravenous immunoglobulin treatment in paraneoplastic neurological syndromes with antineuronal autoantibodies. J. Neurol. Neurosurg. Psychiatry 60:388, 1996. Umapathi, T., Hughes, R. A., Nobile-Orazio, E., and Leger, J. M.: Immunosuppressive treatment for multifocal motor neuropathy. Cochrane Database Syst. Rev. CD003217, 2002. Van den Berg-Vos, R. M., Franssen, H., Wokke, J. H., and Van den Berg, L. H.: Multifocal motor neuropathy: long-term clinical and electrophysiological assessment of intravenous immunoglobulin maintenance treatment. Brain 125:1875, 2002. van Doorn, P. A., Brand, A., and Vermeulen, M.: Antineuroblastoma cell line antibodies in inflammatory demyelinating polyneuropathy: inhibition in vitro and in vivo by IV immunoglobulin. Neurology 38:1592, 1988. van Koningsveld, R., Schmitz, P. I., Meche, F. G., et al.: Effect of methylprednisolone when added to standard treatment with intravenous immunoglobulin for GuillainBarre syndrome: randomised trial. Lancet 363:181, 2004. van Schaik, I. S., Winer, J. B., de Haan, R., and Vermeulen, M.: Intravenous immunoglobulin for chronic inflammatory demyelinating polyneuropathy (Cochrane Review). In The Cochrane Library. Chichester, UK, John Wiley, Issue 2, 2003. Vermeulen, M., van der Meche, F. G., Speelman, J. D., et al.: Plasma and gamma-globulin infusion in chronic inflammatory polyneuropathy. J. Neurol. Sci. 70:317, 1985.

116. Voltz, R., Rosen, F. V., Yousry, T., et al.: Reversible encephalopathy with cerebral vasospasm in a Guillain-Barre syndrome patient treated with intravenous immunoglobulin. Neurology 46:250, 1996. 117. Weishaupt, A., Schonrock, L. M., Stienekemeier, M., et al.: Glucocorticosteroids modulate antigen-induced T cell apoptosis in experimental autoimmune neuritis and cause T cell proliferation in situ. Acta Neuropathol. (Berl.) 102:75, 2001. 118. Wessel, G., Abendroth, K., and Wisheit, M.: Malignant transformation during immunosuppressive therapy (azathioprine) of rheumatoid arthritis and systemic lupus erythematosus: a retrospective study. Scand. J. Rheumatol. Suppl. 67:73, 1987. 119. Wiles, C. M., Brown, P., Chapel, H., et al.: Intravenous immunoglobulin in neurological disease: a specialist review. J. Neurol. Neurosurg. Psychiatry 72:440, 2002. 120. Yan, W. X., Archelos, J. J., Hartung, H. P., and Pollard, J. D.: P0 protein is a target antigen in chronic inflammatory demyelinating polyradiculoneuropathy. Ann. Neurol. 50:286, 2001. 121. Yu, Z., and Lennon, V. A.: Mechanism of intravenous immune globulin therapy in antibody-mediated autoimmune diseases. N. Engl. J. Med. 340:227, 1999. 122. Yuki, N., and Miyagi, F.: Possible mechanism of intravenous immunoglobulin treatment on anti-GM1 antibody-mediated neuropathies. J. Neurol. Sci. 139:160, 1996. 123. Zhu, L. P., Cupps, T. R., Whalen, G., and Fauci, A. S.: Selective effects of cyclophosphamide therapy on activation, proliferation, and differentiation of human B cells. J. Clin. Invest. 79:1082, 1987.

29 Blood-Nerve Interface and Endoneurial Homeostasis ANANDA WEERASURIYA

Permeability of the BNI Dynamics of Blood-Nerve Exchange Relative Contributions of Perineurium and Microvessels Two Routes of Transcapillary Permeation Route of Transperineurial Permeation Transporters at the BNI The Perineurium: An Epithelium? The Epineurium: Not a Diffusion Barrier Assessment of Blood-Nerve Exchange of Solutes Endoneurial Fluid Dynamics Continuity between Cerebrospinal Fluid and Endoneurial Fluid

Convective Endoneurial Fluid Flow Driven by a Proximodistal Gradient Driving Force for Proximodistal Fluid Flow Daily Endoneurial Fluid Turnover (About 30%) Effect of Endoneurial Protein Concentration on EHP Decrease in Perineurial Compliance with Age Implications for Entrapment Neuropathies Endoneurial Homeostasis Controlled Variables of Endoneurial Homeostasis

Vertebrate peripheral axons and their associated glial cells function within a specialized milieu intérieur: the endoneurial microenvironment. Exchange of material between this intrafascicular physiologic space (accounting for about 20% to 25% of the fascicular volume) and the general extracellular space is restricted and regulated by the blood-nerve interface (BNI). This interface, also known as the bloodnerve barrier (BNB), consists of the endoneurial vascular endothelium and the multilayered investing sheath, the perineurium. The relative impermeability of the BNI to bloodborne material protects the endoneurial microenvironment from potentially harmful plasma constituents and rapid fluctuations of plasma solute concentrations that influence glial and axonal functions. The absence of lymphatic drainage in the endoneurial space79 further emphasizes the protective nature of the BNI. The composition of the endoneurial fluid and its physical properties are regulated by homeostatic mechanisms located at the BNI and in the endoneurial cellular elements. Highlighting the need to regulate these variables, the concept of the endoneurial microenvironment

Endoneurial Homeostasis in Lead (Pb) Neuropathy Endoneurial Homeostasis in Nerve Degeneration and Regeneration Endoneurial Homeostasis during Development Factors Regulating BNI Permeability Neurovascular Unit in Nerve Cellular Control of BNI Permeability Molecular Basis of BNI Permeability Regulation Therapeutic Implications for Peripheral Neuropathies Conclusion

as a regulated physiologic space was first proposed about 25 years ago.101,102 Early morphologic studies with several tracers described the permeability of the BNI as either leaky or impermeant and thus downplayed the BNI’s graded permeability to solutes of different sizes, which has been shown by subsequent physiologic studies. In reporting the results of morphologic studies of BNI alterations associated with neuropathies, it is unfortunately still common for authors to use terms such as functional/nonfunctional, competent/noncompetent, integrity/lack of integrity, and intact/breakdown to describe a change in permeability of endoneurial microvessels to morphologic tracers. Characterization of these permeability changes in these judgmental terms is not supported by the available literature. Such terms have limited usefulness in describing changes in endoneurial endothelial cells caused by crush, other trauma, severe toxic insults, and the like. In these cases, there is compelling evidence of endothelial cell injury or death. Therefore, when interpreting results 651

652

Functional Compartments and Nerve-Fiber Environment

from morphologic studies on BNI permeability, it is more prudent to limit the description to what is observed (e.g., a change in permeability to a particular tracer) rather than to misleadingly postulate a “breakdown” of the BNI, or BNB. Numerous studies have demonstrated increases in endoneurial microvessel permeability without any accompanying ultrastructural evidence of endothelial cell damage or death. At the same time, physiologic studies have measured initial increases and subsequent decreases in endoneurial microvessel and perineurial permeability. Taken together, these results reflect a functional complexity of the BNI much greater than that of a simple and passive restrictive barrier impeding the access of hydrophilic solutes to the endoneurial microenvironment. In fact, they are consistent with the dynamic aspects of a “blood-nerve interface” and reflect the operation of endoneurial homeostatic mechanisms, much as in the blood-tissue interfaces of other mammalian organs. For these reasons, and in anticipation of the increasing significance of emerging concepts such as the neurovascular unit and immunologically driven dynamic permeability changes, it is proposed that the BNI represents a more powerful and rigorous description of the regulatory dynamics governing the blood-nerve exchange of various solutes and water.

PERMEABILITY OF THE BNI Morphologic tracers (Evans blue–albumin, horseradish peroxidase [HRP], ferritin, etc.) have not been detected in the endoneurial space following intravascular or perifascicular administration of these substances.51,83 Furthermore, lymphatics have not been described in the endoneurium.79 This has led to the belief that the endoneurial extracellular space, like the cerebral microenvironment, is a protected space relatively free of plasma proteins. Subsequent quantitative studies with radiotracers confirmed the limited permeability of the endoneurial endothelium and perineurium.8,62,94,100,101,106 The morphologic correlates of this restricted blood-nerve exchange of small solutes as well as plasma macromolecules are the belts of tight junctions (zonulae occludentes); these surround the cells of the endoneurial vascular endothelium and inner layers of the perineurium.4,52,61,72 The calculated permeability coefficients demonstrate that only the cerebral capillaries are less permeable than the endoneurial counterparts. These studies have clearly and definitively established the specialized and regulated nature of the endoneurial microenvironment. The perineurium, which is the less permeable component of the BNI,24,64,100,101 defines and circumscribes the endoneurial microenvironment and separates it from the epineurial perifascicular fluid. This passive and nonvectorial role for the perineurium is consistent with its nonpolarized cytologic ultrastructure: basal membrane on both sides and

absence of a histologic distinction between apical and basal membranes,51,72 and the absence of transperineurial transport of ions102 and L-phenylalanine.87 The endoneurial vascular endothelium, in contrast, exerts an active role in regulating the immediate environs of axons and glia through its direct mediation of blood-nerve exchange. It exhibits clear cytologic polarization,37 and demonstrates facilitated transport of D-glucose62 and L-phenylalanine.87 Endoneurial capillaries, though leakier than the very “tight” epithelia,56 are nevertheless less permeable than all other capillaries examined so far except cerebral capillaries.10,11,48

DYNAMICS OF BLOOD-NERVE EXCHANGE Two routes are available for blood-nerve exchange of material. One pathway, the direct one, is across the endothelium of the endoneurial microvasculature. The other is an indirect route requiring passage of material through a third compartment interposed between the vascular space and the endoneurial extracellular space. This compartment is the perifascicular fluid that exchanges directly with the vascular compartment through the relatively leaky epineurial capillaries. Evidence from various physiologic and morphologic studies strongly indicates that blood-nerve exchange occurs predominantly via endoneurial capillaries and that transperineurial exchange is a minor contributor to this process. Hence, the terms BNI permeability and endoneurial capillary permeability are used interchangeably. Strictly speaking, BNI permeability to a given solute is slightly higher than the endoneurial capillary permeability to that same solute. The difference is due to the minor contribution of the relatively impermeable perineurium to BNI permeability.

Relative Contributions of Perineurium and Microvessels Theoretically, blood-borne substances can reach the endoneurial extracellular space either by traversing the endoneurial vascular endothelium or by crossing the multilayered perineurium after gaining access to the perineurial extracellular space. The following evidence favors the former pathway as the major route of blood-nerve exchange: 1. The perineurium is impermeable to ionic lanthanum, whereas the endoneurial capillaries are not.32,72 2. Intravascular perfusion with a hyperkalemic solution inactivates peripheral nerve much more rapidly than when the nerve is bathed by the same hyperkalemic solution.24 3. Histamine increases the permeability of endoneurial capillaries to macromolecular tracers, whereas it has no such effect on the perineurium.49,76

Blood-Nerve Interface and Endoneurial Homeostasis

4. In leprosy, the endoneurial blood vessels become permeable to ferritin, whereas the perineurium remains impermeable to this tracer.7 5. In the frog sciatic nerve, where perineurial permeability has been measured independently, the endoneurial capillaries are more permeable.89,100–102 This relationship has also been confirmed for the rat tibial63 and sciatic nerve.91 6. During the second to sixth week of wallerian degeneration, while the perineurial permeability increases about fourfold, the permeability coefficient–surface area product (PS) of the frog sciatic nerve decreases by more than 60%, reflecting the greater sensitivity to PS to permeability (P) of the capillaries than that of the perineurium.90 7. Calculations from the morphometric studies of Bell and Weddell6 reveal that the perimeter of perineurium in rat sciatic nerve cross section is approximately equal to the sum of the perimeters of all the endoneurial vessels. 8. PS of the adult rat sciatic nerve perineurium to [125I]albumin was measured to be 1.48 ⫾ 0.28 ⫻ 10⫺7 mL ⭈ g⫺1 ⭈ s⫺1 (n ⫽ 6) (A. Weerasuriya, unpublished observations). This is about two orders of magnitude less than the corresponding value for the endoneurial vessels.94 All these studies clearly emphasize the much more restrictive diffusion barrier properties of the perineurium. As far as solute concentrations in the endoneurial fluid are concerned, permeability of the BNI to a solute is an accurate and adequate measure of the blood-nerve exchange of that solute.

Two Routes of Transcapillary Permeation Before adequate structural evidence was available, physiologists, on the basis of permeability measurement to hydrophilic solutes of various sizes, postulated that the capillary endothelium contained a set of large and another set of small hydrophilic pores.19,53 The postulated small pores could be either cylindrical channels with a diameter of 7 to 9 nm or slitlike pores with a width of about 6 nm. They are considered to occupy about 0.01% of the capillary surface area, and are permeable to ions, amino acids, glucose, and the like but not to macromolecules such as albumin. These bigger molecules traverse the capillary wall via the larger pores, which are postulated to have a diameter of 70 to 90 nm. As would be expected, these large pores are rare, and the ratio of large to small pores, though variable, seems to be in the range of about 1:30,000.22,66 Evidence from electron microscopic studies suggest that the small pores are the interendothelial clefts, which have a width of about 20 nm and occupy about 0.4% of the capillary surface area.9 Because of their rarity, the identity of large pores is less certain. Potential candidates are (1) widened intercellular

653

junctions, (2) transendothelial channels formed by the fusion of plasmalemmal invaginations, and (3) fenestrae. Properties of the small- and large-pore pathways are exhaustively reviewed by Crone and Levitt,12 Taylor and Granger,80 and Rippe and Haraldsson.67 Endoneurial capillaries also have separate small- and large-pore pathways for hydrophilic solutes of different sizes.

Route of Transperineurial Permeation The perineurium is a multilayered structure of flattened cells derived from fibroblasts. Each layer of the perineurium is invested by basement membrane on both sides. Hence, in contrast to classic epithelial tissue, the perineurium does not display an apicobasal polarity. The major route for transperineurial passage of material, then, is a paracellular pathway. This paracellular pathway consists of a number of belts of intercellular tight junctions arranged in series due to the multilayered nature of the perineurium. The relative impermeance of this route is illustrated by the fact that local anesthetics administered for nerve blocks are injected via a needle that penetrates the perineurial sheath. The number of layers in the perineurium decreases proximodistally. Thus the fine terminals of sensory and motor nerves have only one or two layers of perineurial cells. In comparison to the central nervous system, the structure of the perineurium combines features of the dura mater in terms of mechanical strength and the arachnoid with respect to impermeance. A recent report of the presence of vascular endothelial cadherin in perineurial cells73 suggests that perineurial permeability is regulated by mechanisms similar to those operating on the vascular endothelial cells of the endoneurium. This possibility is further strengthened by the presence of tight junction proteins (occludin and zonula occludens-1) in perineurial and nerve endothelial cells.85 Attempts to examine the consequences of perineurial removal are confounded by a compromise of the endoneurial vasculature, which receives nutrient branches from transperineurial arteries.40,44,46,75,81 However, this remains an important question especially with regard to perineurial regeneration associated with nerve repair. Thomas and Bhagat,82 studying perineurial regeneration and reorganization following removal of endoneurial contents, suggested that “neural structures may be responsible for the development and maintenance of the structural organization of the perineurium.”

Transporters at the BNI Given the relative impermeability of the perineurium and endoneurial capillary wall, it is not surprising that bloodnerve exchange of several solutes depends on the presence of specific transporter molecules straddling the interface. This condition is analogous to that of the cerebral microvasculature, which is a dynamic capillary interface with the

654

Functional Compartments and Nerve-Fiber Environment

least known permeability and minimal pinocytotic traffic coupled with a whole host of transporters and sensors responsible for the moment-to-moment, active and immediate regulation of the cerebral microenvironment. The endoneurial microenvironment, unlike the cerebral microenvironment, does not require a moment-to-moment regulation of nutrients and oxygen. Several transporters designed to meet the metabolic requirements of endoneurial constituents have been described. The earlier radioisotopic demonstration of facilitated transport of D-glucose62 has been complemented by reports on the presence of GLUT-1 in endothelial cells and perineurial cells.15,16,33 In keeping with the earlier postulate that the perineurium is a specialized connective tissue, these studies did not demonstrate an apicobasal polarity of GLUT-1 transporters in perineurial cells. Hence the role of perineurial GLUT-1 transporters appears to be the nourishment of perineurial cells.

The Perineurium: An Epithelium? The earliest ultrastructural studies of the perineurium concluded that it was an epithelium.71 It was postulated that perineurial cells with active metabolically driven pumps contributed to the composition of the fluid surrounding axons and glial cells. An analogy was drawn to the arachnoid epithelium investing the central nervous system. Subsequent ultrastructural studies have not confirmed these suppositions, and in fact a somewhat different picture has emerged. Frank Low and colleagues20,21,27,35 concluded that the perineurium is a specialized type of connective tissue, and the subsequent demonstration of the embryologic origin of the perineurium from fibroblasts has validated this conclusion. These studies showed that perineurial cells were thin, flattened, interconnected cells and that the layers of perineurial cells had basement membrane on both sides. This is a clear distinction from layers of epithelial cells, which have basement membrane only on one side—the basolateral side. The other surface of the epithelial layer constitutes the apical side and is not covered by basement membrane. A cardinal feature of epithelial layers is the apicobasal polarity of functional organization. This is reflected in the asymmetrical distribution of transporters, channels, and pumps across the apical and basolateral membranes. In addition, the functional correlate is a transepithelial flux of solutes and most often a transepithelial potential difference. Despite several attempts, active transperineurial fluxes of solutes and a transperineurial potential difference have not been discovered.45,102,104 Thus the current consensus is that the fibroblastderived perineurium is a specialized laminated connective tissue.27,39 Even though perineurial cells express GLUT-1, transperineurial transport of glucose is yet to be demonstrated. There is as yet no evidence of sodium-driven transperineurial cotransport of glucose and amino acids.

Whether perineurial cells participating in microfasciculation associated with nerve regeneration acquire specialized functions remains an open and intriguing question.

The Epineurium: Not a Diffusion Barrier Several fascicles are loosely bound together by the epineurium to form a peripheral nerve. The function of the epineurium is that of a connective tissue. Separation of the bundles by disruption of the epineurium does not seem to disturb the integrity of the nerve fascicles. Morphologic tracer studies52,72 demonstrated the easy penetrance of tracer macromolecules through the epineurium and even some of the outer layers of the perineurium. It is only the innermost layers of the perineurium that are responsible for the diffusion barrier properties to the nerve sheath.

Assessment of Blood-Nerve Exchange of Solutes Concentrations of endoneurial constituents are affected by exchange across endoneurial capillary interface and the barrier imposed the inner layers of the perineurium, and the metabolic processes of endoneurial cells. Permeabilities of the two exchange sites are assessed by morphologic and physiologic methods. Both techniques have their advantages and limitations. Historically, morphologic techniques establish the initial broad parameters and then physiologic methods provide quantitative measurements of blood-nerve transfer coefficients. The major advantage of morphologic techniques is the localization of barriers to the penetrance of markers and the array of tracers of varying molecular weights, charges, and sizes. Conversely, these methods are limited by the sensitivity of the histologic staining techniques. For example, histologic methods have not detected blood-nerve transfer of macromolecules, but physiologic methods, with radiotracers, estimate blood-nerve transfer of albumin at about 10⫺5 cm/s. Thus physiologic methods not only allow a quantification of transfer rates, but also are more sensitive, especially to macromolecules and other larger species. Furthermore, only transfer rates of amino acids, sugars, and other small molecules can be studied by physiologic radioisotopic methods; these techniques do not allow localization of the sites of transfer or transport. Immunostaining techniques and in situ hybridization complement these transport studies by localizing the putative transporters and diffusion sites. Technical details of measurement of blood-nerve transfer rates are described elsewhere.64,94 The theoretical foundations of blood-tissue transfer studies and their limitations are elegantly described by Smith and Allen.74

655

Blood-Nerve Interface and Endoneurial Homeostasis

The extracellular space within a nerve fascicle is about 20% to 25% of the total intrafascicular volume. The endoneurial hydrostatic pressure (EHP) exerted by this fluid is about 2 to 3 mm Hg. In terms of fluid dynamics, the endoneurial contents consist of a noncompressible aqueous solution, and a somewhat compressible cellular content constrained by an elastic sheath, the perineurium. Hence, the EHP will be determined by the volume of the endoneurial contents and the compliance of the perineurial sheath. Fluid enters and leaves a given segment of the endoneurial space across the walls of the endoneurial microvessels, and by convective endoneurial proximodistal fluid flow. Given the slightly positive tissue hydrostatic pressure of the endoneurial space, fluid does not normally enter this space across the perineurium. Endoneurial extracellular fluid exchanges material with blood directly across the endoneurial vascular endothelium, and indirectly across the multilayered perineurium, and is turned over by convective endoneurial fluid flow (EFF). The kinetics of this exchange are governed by (1) the hydrostatic and osmotic pressures in the different compartments, (2) the ion concentrations in plasma and endoneurial fluid, (3) the distribution volumes of the relevant solutes in the endoneurium, (4) the PSs of the BNI to these solutes, and (5) the rate of EFF.

Continuity between Cerebrospinal Fluid and Endoneurial Fluid The meninges of the central nervous system are continuous with sheaths of peripheral nerve.21 Whereas the epineurial connective tissue layers are continuous with the dura mater at the central ends (dorsal and ventral roots) of peripheral nerves, the relationship between the other meninges (pia mater and arachnoid layer) and perineurium is more complex. At the subarachnoid angle of nerve roots, there appears to be continuity of the subarachnoid space and endoneurial space, thus providing for continuity between cerebrospinal fluid (CSF) and endoneurial fluid. This, then, is the conduit through which material passes from CSF to endoneurial fluid.69 The embryologic aspects of this continuity do not appear to have been investigated.

injections of dyes, crystals, and radioactive mineral salts. Their conclusions were based on comparisons of proximodistal spread of the indicators in dead and living tissue. The two major limitations of that study were (1) the use of injection volumes that were no smaller than 100 ␮L and (2) the use of small hydrophilic tracers that could have migrated down the nerve by entering the vascular compartment and later reentering the endoneurial space. Therefore, they were unable to calculate a precise rate of convective fluid flow from their data and suggested an approximate rate of 3 mm/hr. This was supported by Mellick and Cavanagh36 from the results of their studies on traumatized chicken sciatic nerves. Low,28 injecting 10 ␮L of tetrodotoxin into the endoneurium and monitoring the rate and spread of inactivation, concluded that convective fluid flow is about 4 to 8 mm/hr. Morphologic studies21,35 demonstrated a continuity between the spinal subarachnoid space and the endoneurial space of spinal nerve roots; this strongly supports the possibility that CSF contributes to endoneurial fluid. Earlier, Waksman88 had suggested that diphtheria toxin enters the central nervous system through peripheral nerves, probably exploiting this pathway. It is not unlikely that, in diabetic neuropathy, disturbances in convective EFF play a role in the greater fiber loss and demyelination seen more distally than proximally. The data in Figure 29–1 are consistent with a convective flow of endoneurial fluid. The shift of the curve to the right (proximal to distal) from the first to the second hour is clearcut, but that from the second to the fourth hour is somewhat more subtle, although nevertheless evident upon closer inspection of the two curves. Whereas both curves seem to

900 800 Activity (cpm) per mg of nerve

ENDONEURIAL FLUID DYNAMICS

700

1 hour 2 hours 4 hours

600 500 400 300 200 100

Proximal

Distal

0 −20

−15

−10

−5

0

5

10

15

20

Distance along nerve (mm)

Convective Endoneurial Fluid Flow Driven by a Proximodistal Gradient In addition to the blood-nerve exchange discussed earlier, another putative source of input/output to the endoneurium is convective EFF. In an elegant series of experiments, Weiss and colleagues105 demonstrated the presence of a proximodistal flow of fluid in rat and guinea pig sciatic nerve. As indices of fluid movement, they used endoneurial

FIGURE 29–1 Rate of endoneurial convective fluid flow of 22Na as a function of time. In all three experiments, 70 nL of saline with 22Na were microinjected into a rat sciatic nerve. Then, 1, 2, or 4 hours later, the nerves were harvested, cut into 3-mm segments, counted for 22 Na activity, dried, and weighed. Negative numbers indicate distances proximal to the site of injection and positive numbers represent distances distal to the site of injection. (A. Weerasuriya, unpublished data.)

656

Functional Compartments and Nerve-Fiber Environment

peak at the fifth millimeter, the 2-hour curve has a bigger shoulder at lengths less than 5 mm, and the 4-hour curve has a shoulder at lengths greater than 5 mm. Thus the pattern of 22 Na distribution along the length of the nerve at the three different survival times clearly demonstrates a progressive asymmetrical proximodistal movement of the isotope.

Driving Force for Proximodistal Fluid Flow There does not appear to have been any mechanistic studies addressing the issue of the driving force for proximodistal fluid flow. The descriptive studies on EHP57 are consistent with the CSF pressure in the spinal cord being the pressure head for the proximodistal flow of endoneurial fluid. The hydrostatic pressure in the cord is about 10 mm Hg, that in the peripheral nerves about 1 to 2 mm Hg, and that in the dorsal root ganglia about 3 to 5 mm Hg.43 The presence of this pressure drop from the interstitial spaces of the spinal cord to the endoneurial interstitium of peripheral nerve supports the contention that the CSF pressure is the main propulsive force of the endoneurial fluid. Unfortunately, this postulate is not without theoretical limitations. For example, what would be the fate of this CSF pressure–driven endoneurial fluid when it reaches the distal ends of sensory nerves, which do not have open perineurial sleeves?27 Is the perineurium in such nerves more permeable distally than proximally to allow for a transperineurial dissipation of endoneurial fluid? If not, is there a slower turnover of endoneurial fluid at the distal end of a sensory nerve? Would this then contribute to the greater vulnerability of sensory nerves to pyridoxine toxicity?

Daily Endoneurial Fluid Turnover (About 30%) The albumin content of desheathed human sural nerve was reported to be 8.7 ␮g/mg of dry weight.54 If it is assumed that (1) the endoneurial wet/dry weight ratio is 3.0, (2) endoneurial albumin is extracellular and free, and (3) endoneurial extracellular space is about 25%, then this value can be converted to an albumin concentration of 11.6 mg/mL in the endoneurial fluid. Plasma albumin concentration is 33.1 mg/mL.54 From these two concentration terms and the PS to albumin,94 the calculated rate of blood-nerve albumin transfer is about 1.2 mg ⭈ g⫺1 ⭈ day⫺1. At this rate of transfer, and assuming relatively constant albumin concentrations in endoneurium and plasma, about 30% of the endoneurial albumin is turned over each day. By comparison, CSF (and its constituents) is turned over about four times each day60 and, in anesthetized rats, daily turnover of interstitial albumin in hind limb muscles and skin is about 65%.65 The latter value increases about fourfold in awake, freely moving rats. The rate of removal of albumin from the endoneurium is also about 1.2 mg ⭈ g⫺1 ⭈ day⫺1. In muscle, skin, gut, and

similar tissues, lymphatic drainage plays a role in clearing interstitial albumin, and in the central nervous system the cerebrospinal fluid and perivascular drainage are the sinks for extracellular albumin.3 In peripheral nerve, in the absence of both lymphatics and an active CSF circulation, the route of removal of albumin and other macromolecules remains to be identified. The metabolic breakdown of albumin to supply amino acids to axons and glia is one possibility; another is the long-suspected proximodistal convective flow of endoneurial fluid.

Effect of Endoneurial Protein Concentration on EHP From the equation proposed by Landis and Pappenheimer,53 endoneurial albumin can be expected to exert an interstitial oncotic pressure of about 3 mm Hg. The recorded EHP of 2 to 3 mm Hg will oppose the endoneurial oncotic pressure and thus minimize net fluid filtration from the endoneurial vasculature. However, this balance of forces can be disturbed if the capillary permeability to albumin increases slightly, allowing the endoneurial albumin concentration to rise and thus draw more fluid from the vascular compartment into the endoneurial interstitium. The resulting endoneurial edema together with the low compliance and hydraulic conductivity of the perineurium will elevate EHP. The net fluid gain by the endoneurial space will cease when a new equilibrium is established among the hydrostatic and oncotic pressures of the endoneurial and intravascular compartments. It is quite likely that this is the mechanism of edema formation observed in experimental diabetic and lead neuropathies, as well as early wallerian degeneration. In contrast, the edema present in galactose-induced neuropathy is likely to be due to the presence of an excess of non–vascularly derived osmolytes in the endoneurium, corroborated by the absence of change in the PS of the BNI to sucrose in galactose-intoxicated rats.38

Decrease in Perineurial Compliance with Age Developmental increase of EHP has two plateaus.13 The first one is from about 3 to 13 weeks of age, and the second one is from 6 months onward. Elevations of EHP can be produced either by increased EFF or by reduced perineurial compliance. The results of the study by Crowley and colleagues13 demonstrated that reduced perineurial compliance with aging contributes to the second increase in EHP.

Implications for Entrapment Neuropathies Extrafascicular mechanical compression and the resultant ischemia certainly contribute to the symptomatology of entrapment neuropathies. But, are there other factors more closely tied to the endoneurial microenvironment that affect the susceptibility and evolution of the nerve injury? It is

Blood-Nerve Interface and Endoneurial Homeostasis

hypothesized that the initial event in the evolution of an entrapment neuropathy is the limitation and reduction of EFF as a result of externally applied mechanical forces. Second, elevation of EHP as a result of obstruction of EFF and continued application of these external forces leads to endoneurial ischemia and its attendant pathology. Additionally, the reduced compliance of the aged nerve13,99 makes it more susceptible to externally applied mechanical pressures causing an elevation of EHP. This hypothesis is consistent with (1) the paucity of carpal tunnel syndrome (CTS) in teenagers who play arcade games, (2) the higher incidence of CTS in conditions with increased tissue water content or edema, such as pregnancy and diabetes, and (3) the presence of symptoms quite proximal to the site of entrapment. However, this hypothesis does not explain the higher incidence of CTS in females or the varying patterns of recovery seen after resection of the flexor retinaculum.

ENDONEURIAL HOMEOSTASIS Much as in other tissues, the various compartments of the nervous system employ homeostatic mechanisms to regulate their respective internal environments. The relevant physiologic functional unit for a vertebrate peripheral nerve is a fasciculus. This cylindrical structural entity, encompassing axons, glial cells, and other attendant satellite cells, has its outer limits defined by the perineurium. This physical limit is also a physiologic barrier and limits the exchange of material between the endoneurial space and the general extracellular space surrounding a peripheral nerve.

Controlled Variables of Endoneurial Homeostasis The specialized microenvironment of peripheral nerve fibers is maintained with the assistance of the BNI. Regulated blood-nerve exchange across the BNI and turnover of endoneurial fluid by convective fluid flow are vital for the maintenance of endoneurial physiologic parameters (blood flow, oxygen tension, pH, oncotic pressure, hydrostatic pressure, ion concentrations, etc.) within the normal range necessary for the proper functioning of nerve fibers. There are independent transendothelial pathways for the movement of ions and macromolecules. Some of these physiologic parameters have been measured: PS to [14C]sucrose,63,70 to [14C]glucose,62 to [125I]albumin,94 and to 22 Na,91 as well as endoneurial blood flow,31,68 EHP,30,41 and ion concentration in endoneurial fluid.28,93 Other parameters have been described only qualitatively or not at all (endoneurial concentration of hydrogen and calcium, rate of convective EFF; volume of endoneurial extracellular space, perineurial permeability to macromolecules, tortuosity of the endoneurial extracellular contents, etc.). To establish that a particular variable is the controlled variable of a

657

homeostatic loop, it is necessary to establish the relative stability of that variable within a physiologic range in the face of environmental stressors and perturbations. This list of variables is not exhaustive, nor does it relate to all the pathophysiologic alterations associated with the endoneurial microenvironment. However, it is hoped that discussions regarding these few variables with available data in a few clinical scenarios will emphasize the relevance of considering pathophysiologic alterations as perturbations of endoneurial homeostasis. This promotes the view that therapeutic strategies are attempts to assist endoneurial mechanisms to restore the normal homeostatic balance of the endoneurium.

Endoneurial Homeostasis in Lead (Pb) Neuropathy The major effects of Pb on peripheral nerve are endoneurial edema, nuclear inclusion bodies, demyelination, elevation of EHP, and increased permeability of the BNI.108 Of these alterations, the nuclear inclusion bodies are observed first, followed by endoneurial edema. It had been suggested earlier29,42 that an increase in the permeability of endoneurial capillaries was the primary pathologic event leading to subsequent Schwann cell damage and segmental demyelination. Later studies on nerve pathology and with morphologic tracers indicated that accumulation of lead in the endoneurium and nerve edema precede qualitative changes in the permeability of endoneurial capillaries.59,109 The BNI index to albumin (a measure of the rate of albumin entry and removal from the endoneurium) starts to increase only at 6 weeks in lead-intoxicated rats, and suggests that the change in BNI permeability is subsequent to the direct toxic effect of lead on Schwann cells.47 A subsequent study96 provided clear, quantitative evidence that endoneurial pathology precedes an increase in permeability of the BNI, and furthermore, that the increase in permeability is about threefold. It is further implied that the increase of BNI permeability is not a consequence of disruption of its barrier properties, but more in the nature of an adaptive response to aid in the clearing of myelin debris from the endoneurium. The following scheme (Fig. 29–2), which is an extension of earlier hypotheses, attempts to delineate the causal relationships among the changes in various components of the endoneurium in leadinduced neuropathy. The primary event is the increase of plasma lead concentration. From the calculated PS of the BNI to ions and small nonelectrolytes, the PS of the BNI to lead is predicted to be around 1 ⫻ 10⫺4 mL ⭈ g⫺1 ⭈ s⫺1 and the half-time to be around 30 minutes. Therefore, in the absence of rapid fluctuations of blood lead levels, there is likely to be approximate equilibration between the lead concentrations in plasma and endoneurial fluid. Endoneurial lead accumulates among the myelin lamellae and in nuclear inclusion bodies in Schwann cells. The sequestration of lead in intranuclear bodies is a

658

Functional Compartments and Nerve-Fiber Environment

↑ Endoneurial Pb Schwann cells

↑ Permeability of BNI

Demyelination Endothelium

↑ Endoneurial albumin content

Endoneurial Homeostasis in Nerve Degeneration and Regeneration

Toxic effects on Schwann cells

↑ Endoneurial osmolytes ↑ Endoneurial water content and edema

Relatively unchanged perineurial permeability ↑ Endoneurial hydrostatic pressure

FIGURE 29–2 Proposed sequence of events leading to demyelination and endoneurial edema in lead neuropathy. (Data from Weerasuriya, A., Curran, G. L., and Poduslo, J. F.: Physiological changes in the sciatic nerve endoneurium of lead intoxicated rats: a model of endoneurial homeostasis. Brain Res. 517:1, 1990.)

protective mechanism to minimize elevations of cytosolic lead concentration. Myelin by-products increase the osmolality of the endoneurial fluid and cause a net movement of water from the vascular lumen to the endoneurium, leading to endoneurial edema. In the presence of a relatively unchanged permeability of the perineurium, the increased water content will cause an elevation of EHP. Endothelial cells also accumulate lead in intranuclear bodies, and the increased permeability of the BNI could be a consequence of endothelial damage, a Schwann cell–initiated mechanism to clear myelin debris from the endoneurium, or both. In either case, because interendothelial HRP leakage is not profuse and endothelial cell changes are more of a reactive nature rather than an expression of outright injury, it is suggested that the change in BNI permeability is an adaptive response of the vascular endothelium to restore/maintain the homeostasis of the endoneurial microenvironment. The occurrence of remyelination in chronically leadintoxicated rats supports the previous contention of an endoneurial microenvironment conducive for repair and regeneration. An intriguing but unresolved clinical issue is that lead toxicity predominantly leads to an encephalopathy in children, whereas in adults the major manifestation is a neuropathy. A greater vulnerability of the juvenile cerebral neurovascular unit to lead is a likely but as yet unproven hypothesis.

Wallerian degeneration is probably the most drastic reorganization of the endoneurial architecture.77,84 The fact that perineurial permeability increases only about twofold under these conditions97 argues against a “breakdown” of the barrier properties of this structure. An alternative interpretation of a dynamic response of the perineurium to maintain endoneurial homeostasis seems more plausible. Morphologically, perineurial cells in degenerating nerve show no evidence of disruption.25,107 On the contrary, they proliferate, hypertrophy, and increase their organelle content.25 The initial peak of perineurial permeability, lasting only a few days,97 is most likely related to the acute inflammatory response triggered by the trauma of nerve section, and the proliferative response of the perineurium25 (Fig. 29–3). Sectioned degenerating nerves display an increased number of mast cells.14,25,57 Degranulation of these mast cells releases histamine, a biogenic amine with potent inflammatory properties. The later sustained increase is probably a component of endoneurial homeostasis related to prevention of elevation of EHP (Fig. 29–4) and clearance of myelin debris from the endoneurium. The increased permeability of both components of the BNI is likely to facilitate the entry of monocytes into the endoneurium for myelin phagocytosis.78 Although lipid droplets and proteinaceous material are present among perineurial layers in degenerating nerves, the cells themselves are not disrupted.25,107 Also, it is unlikely that the postulated diffusible factor plays a role in the short-lasting acute increase of perineurial permeability. 175 Change from normal (%)

↑ Blood Pb >0.8 ug/mL

Perin. perm.

150 125 100

Endon. volume

75 50

Perin. area

25 0 0

2

4

6

8

10

12

14

16

Time (wk)

FIGURE 29–3 Relative changes in perineurial permeability to 22Na, endoneurial volume, and perineurial area during wallerian degeneration. Endon. Volume ⫽ endoneurial volume; Perin. Area ⫽ perineurial area; Perin. perm. ⫽ perineurial permeability. (Data from Weerasuriya, A., and Hockman, C. H.: Perineurial permeability to sodium during wallerian degeneration in rat sciatic nerve. Brain Res. 587:327, 1992.)

659

10

150 Pressure 125

8

100

6

75 Permeability

4

50

2

25

0

0 0

4

8

12

20

16

21

28

Time (d)

Perineural permeability—change from normal (%)

Endoneurial pressure (cm H2O)

Blood-Nerve Interface and Endoneurial Homeostasis

FIGURE 29–4 Endoneurial hydrostatic pressure and perineurial permeability to 22 Na during the first 4 weeks of wallerian degeneration. (Data from Weerasuriya, A., and Hockman, C. H.: Perineurial permeability to sodium during wallerian degeneration in rat sciatic nerve. Brain Res. 587:327, 1992.)

Endoneurial Homeostasis during Development During normal development, results from studies employing morphologic tracers indicate that the BNI of juvenile animals is more permeable to macromolecules than that of adult animals.23,34 This decrease in the PS of the BNI during development has been quantified with radiotracer studies.95 An intriguing issue arising out of these observations is how developing nerve, with its highly permeable BNI, avoids the

Wet weight to dry weight ratio

BNI PS to 22Na (mL . g−1 . sec−1) ⫻ 104

Delineation of the properties of the BNI during nerve regeneration provides a more comprehensive picture.92 From fourth day up to the eighth week after the crush lesion, BNI PS to 22Na was significantly greater than the normal value, but by the 18th week after the crush, PS was not significantly different from the normal value. The data in Figure 29–5 demonstrate the easier access of water into the endoneurial compartment for up to at least 8 weeks after the crush lesion. This increased ability of water to move into the endoneurial space is compared with the

measured endoneurial water content at the same time points (Fig. 29–6). The endoneurial water content, calculated as the wet weight–to–dry weight ratio, increased from the fourth day and remained elevated during the entire experimental period. Its peak at the second week postcrush corresponds to a period of increased PS of the BNI to 22Na. However, at 18 weeks postcrush, when the BNI PS to 22Na is normal, the nerve is still edematous. What, then, is the driving force for the elevated endoneurial water content? Although an increased BNI PS to plasma macromolecules could elevate the osmolality of the endoneurial fluid and thus draw water into this space, it is contended that a decrease in the endoneurial fluid turnover is also a factor in maintaining the increased water content of the regenerated nerve. Conversely, the elevated water content is not associated with nerve edema; at 12 weeks after a crush injury the EHP of the regenerating nerve has returned to normal.98 It is postulated that an altered compliance of the regenerated nerve helps to maintain EHP in the normal range.

14 12 10 8 6 4 2 0

7 6 5 4 3 0

0

2

4

6

8

10

12

14

16

18

Time (wk)

FIGURE 29–5 BNI permeability to 22Na during regeneration after a crush lesion. Time after the lesion is represented on the x axis. There were four animals at the 4-day and 1-week time points, and five animals at all other time points. (Data from Weerasuriya, A.: Changes in the permeability coefficient-surface area product [PS] of the bloodnerve interface [BNI] to sodium in degenerating and regenerating rat sciatic nerve. Paper presented at the 11th Biennial Meeting of the Peripheral Nerve Study Group, Aachen, Germany, July 1993.)

2

4

6

8

10

12

14

16

18

Time (wk)

FIGURE 29–6 Changes in endoneurial wet weight-to-dry weight ratio in desheathed rat sciatic nerve following a crush injury. Data are presented as mean ⫾ SEM. The SEMs of normal controls and of injured nerve at 6, 8, and 18 weeks are smaller than the size of the symbol. (Data from Weerasuriya, A.: Changes in the permeability coefficient-surface area product [PS] of the blood-nerve interface [BNI] to sodium in degenerating and regenerating rat sciatic nerve. Paper presented at the 11th Biennial Meeting of the Peripheral Nerve Study Group, Aachen, Germany, July 1993.)

660

Functional Compartments and Nerve-Fiber Environment

pathology observed in an adult nerve with a comparably permeable BNI. For example, in the adult an elevated PS causing BNI extravasation of plasma proteins inevitably leads to endoneurial edema and increased EHP. The primary event in that pathology is the increase of endoneurial albumin content elevating the oncotic pressure in the endoneurial interstitial space. Because of the operation of Starling forces, water is drawn into the endoneurial interstitium from the vascular compartment. This leads to an elevated endoneurial water content and, together with the low hydraulic conductivity and compliance of the perineurium, causes an increase in the EHP. This elevated EHP can compromise the intrafascicular microcirculation, leading to ischemia and its accompanying pathology. By contrast, a highly permeable BNI during the first weeks of development does not lead to an elevated EHP and edema because the endoneurial accumulation of fluids and osmolytes is prevented by a more permeable perineurium, allowing for the clearance of this material by the epineurial lymphatics. Additionally, a higher compliance of the juvenile endoneurial tissue mass could also counter a tendency for an increased EHP. Therefore, in the developing nerve a much more permeable endoneurial microvasculature (which is probably necessary for greater blood-nerve exchange of material to support the more active metabolism, and is also a consequence of tight junction dissolution and reformation during endothelial proliferation) does not lead to pathologic consequences because of the combination of a relatively permeable perineurium and epineurial lymphatics, and metabolic clearance of plasma-derived macromolecules.

FACTORS REGULATING BNI PERMEABILITY Microvessels in the endoneurium are the least permeable capillaries except for cerebral capillaries. The investing perineurium is even less permeable than the endoneurial capillaries. Hence, cellular components of the endoneurium secrete soluble factors responsible for maintaining the integrity of the tight junctional complexes and associated actin cytoskeletal components of endoneurial vascular endothelial cells and of cells composing the innermost layers of the perineurium. It should be recognized that endoneurial elements not only expend metabolic energy to maintain the junctional integrity of the endoneurium’s barrier tissues, but are at the same time obliged to devote metabolic energy to maintain the pumps and transporters for blood-nerve exchange of critical material needed by the endoneurial cellular elements. Thus the dynamically selective per-

meability properties of the BNI extract a considerable metabolic cost.

Neurovascular Unit in Nerve When considering the homeostatic mechanisms of the cerebral microenvironment, an important concept that is emerging with increasing validity and support from experimental data is that of the neurovascular unit.2,26 The structural components of the neurovascular unit are the cerebral endothelium, the neurons and astrocytes that make intimate contact with the endothelial cells, pericytes, and the extracellular matrix investing this functional unit. Functionally this unit subserves dynamic signaling among the vascular, glial, neuronal, and matrix elements of the cerebral microenvironment. There is direct synaptic and perisynaptic contact among these elements to effect changes extremely rapidly over very short distances. The modular organization of this unit and its microscopic size imply that it is repeated many times to limit control and regulation of the cerebral microvasculature to discrete small volumes of cerebral parenchyma. Signaling in the neurovascular unit is mostly directed at the cerebral endothelium to influence two important physiologic parameters of the cerebral microcirculation: localized cerebral blood flow and capillary permeability. Thus the neurovascular unit plays a crucial role in maintaining an adequate supply of oxygen and nutrients to neurons and glial cells. It plays an even more critical role in cerebral homeostasis by responding to changes in nutrient needs imposed by both physiologic and potentially pathologic stressors. The current concept of the neurovascular unit emphasizes the capillary endothelium, but it is expected that this concept will evolve to encompass both cerebral arterioles and venules because these components of the microcirculation play significant roles in regulating vascular resistance and blood-brain water exchange. What is the relevance of the neurovascular unit for the endoneurial microenvironment? Functionally, axons and Schwann cells can tolerate limited hypoxia and metabolic insults for brief periods of time (seconds to perhaps a few minutes) without suffering irreversible damage. Structurally, only pericytes are within a few microns of the endothelial cells and invested by a common basal lamina. Nevertheless, focal ischemic lesions in peripheral nerve are probably indicative of the presence of functional neurovascular units in the endoneurial microenvironment, with two important differences from their cerebral counterparts. In peripheral nerve, given the greater separation between axons and endothelial cells and between Schwann cells and endothelial cells, signaling molecules have to diffuse through greater distances to reach their targets. Hence the neurovascular unit in nerve, compared to that in cerebrum, represents a functional unit that operates in a larger volume and respond more slowly to perturbations.

Blood-Nerve Interface and Endoneurial Homeostasis

Cellular Control of BNI Permeability The perineurium and the endoneurial microvasculature form the BNI, which isolates the endoneurium by virtue of the “tight” intercellular junctions present in these two structures. Perineurial microvessels that are permeable to plasma macromolecules acquire tight intercellular junctions upon penetrating the innermost layers of the perineurium to enter the endoneurium.50 Thus one or more diffusible factors present in the endoneurium is responsible for the induction of tight intercellular junctions in barrier tissues circumscribing the endoneurium. Such an inducer is likely to be produced by axons, Schwann cells, or both, and regulate synthesis and elaboration of intercellular junctional proteins by perineurial and endothelial cells. In chronically transected nerves, the near-normal PS of the endoneurial endothelium to small nonelectrolytes,90 and the presence of “tight” intercellular junctions in the endothelium,25 strongly argue against axons as being the source of a regulatory factor for the maintenance of the tight junctions between vascular endothelial cells. However, the increased permeability of the perineurium accompanying axonal degeneration25,86,103 strongly suggests that axonally derived factors might be responsible for maintaining the relatively tight barrier properties of the perineurium. This working hypothesis on a dual source of regulatory proteins for modulating the permeability properties of the endoneurial vascular endothelium and perineurium needs to be rigorously tested by examining BNI permeability when either axons or Schwann cells are preferentially damaged by endoneurial injection of toxins. A corollary of this hypothesis is that Schwann cells have several phenotypes, and a major function of some of these subsets is regulation of endoneurial vascular permeability. This is accomplished through soluble factors released into the endoneurial extracellular space and diffusing less than 100 ␮ to reach the intended targets, the endothelial cells. This would be a form of paracrine signaling within a confined and specialized extracellular space. Conversely, based on reasons outlined earlier, perineurial junctional integrity appears to depend primarily on soluble factors secreted by axons and perhaps aided by Schwann cell–derived signaling molecules. A role for Schwann cells is postulated because of the expression of vascular endothelial cadherins by perineurial cells.73 Hence it is hypothesized that the baseline resting permeability of the two components of the BNI are regulated by axons and Schwann cells. An integral aspect of endoneurial homeostasis is the ability of the BNI to adaptively alter its permeability properties to meet the changing needs of endoneurial elements. Such changes may be dictated by programmed and gradual changes associated with growth and maturation, and microenvironmental disturbances precipitated by disease and trauma. How might these adaptive alterations in permeability of the BNI be initiated and effected?

661

Recent evidence strongly implicates the immune system as an active modulator of BNI permeability in a whole host of conditions ranging from trauma to metabolic neuropathies to vascular disorders.18,55,58,110 Thus it is postulated that hematogenous elements interacting directly with the endoneurial vascular elements effect increases in capillary permeability. The perineurium appears to be unusually resistant to inflammatory mediators.1 Hence immunomodulation of the BNI is expected to be limited to the nerve endothelial cells.

Molecular Basis of BNI Permeability Regulation Even though the molecular biology of junctional complexes of the BNI have not been investigated in detail, it is reasonable to assume that they are essentially similar to those of the cerebral microvasculature.5,17 Thus mediators of BNI permeability alterations have to utilize one or more of the following mechanisms: (1) modification of the actin cytoskeleton responsible for anchoring the junctional complexes to the intracellular scaffolding; (2) posttranslational modification of the constituent proteins (occludins, family of zonula occludens proteins, cadherins, claudins, catenins, etc.) of the junctional complexes; and (3) proteolytic degradation of the junctional components. Whereas the first two modes require signaling molecules to gain intracellular access, the last mode may be initiated and completed extracellularly at the junction itself. The classic cytokines and interleukins mediate inflammation-related changes in BBB. They are implicated to play a role in disease-related BNI changes as well. What are the events that initiate permeability changes and how are they sensed? It is not know whether hypoxia inducible factor plays an active role in endoneurial response to ischemia and hypoxia. Does the endoneurium contain angiotensin type 1 receptors that are activated by mechanical stress?111 They would be very attractive candidates for detecting changes in endoneurial pressure.

THERAPEUTIC IMPLICATIONS FOR PERIPHERAL NEUROPATHIES Some neuropathies are a direct outcome of microcirculatory failure in peripheral nerve, while other forms of neuropathy are not associated with microangiopathic failures, and a few exhibit a microcirculatory insufficiency which does not appear to have been the precipitating factor. What role can modification of the BNI play in the treatment of peripheral nerve diseases and trauma? There are at least three strategies. First, an increase in BNI permeability could deliver a higher concentration of therapeutics molecules to their targets. This could be

662

Functional Compartments and Nerve-Fiber Environment

achieved by modifying the drug molecules to increase their passive permeance across the BNI or utilize BNI transporters to gain better access to axons, Schwann cells, and other cells in the endoneurial compartment. Secondly, increased angiogenesis and endoneurial blood flow will directly address those clinical entities with microcirculatory failure. This would also indirectly facilitate the previous strategy. Finally, gene transfer has been successfully employed to treat a variety of nerve disorders in animal models. Therapeutic concentrations of several neurotrophic factors have been delivered with the following methods: intratissue injection of naked DNA, adenoviral gene transfer system. T-lymphocytes transduced with a recombinant retrovirus, and a herpes simplex-based vector. With accelerating advances in gene transfer technology, this last strategy is bound to play an increasing role in the treatment of peripheral neuropathies.

CONCLUSION It is advocated that the term blood-nerve interface is preferable to blood-nerve barrier. The former term encompasses the regulatory nature as well as the passive barrier properties of the perineurium and the endoneurial microvasculature. Historically, the term blood-nerve barrier, derived from the related term for the cerebral vasculature (blood-brain barrier), owes its origin to morphologic demonstrations of the inability of certain tracers to enter the immediate interstitial spaces of axons and neurons, and hence implied an all-or-none nature of the permeability of this barrier to tracers of various sizes. The original scientific term was, in German, Blut-Hirn Schrank, and a literal translation of that term would be “blood-brain cabinet,” with a strong emphasis on its passive barrier properties! As we learn more about the structure and functions of the perineurium and endoneurial microvasculature, especially the graded nature of its permeability and its dynamic adaptive responses to alterations in the endoneurial microenvironment, it seems more appropriate to refer to them as the blood-nerve interface. This conceptual thrust is likely to play a significant role in our understanding of peripheral nerve disorders and the evolution of therapeutic strategies and paradigms for the treatment of these disorders.

ACKNOWLEDGMENTS Experiments reported herein from the author’s laboratory were partially supported by grants RO1–NS30197 (National Institutes of Health), IBN-9420525 (National Science Foundation), and 23750 (MedCen Foundation). This chapter is dedicated to the memory of KNS (1929–1987) and KJN (1927–2001). Ms. Marianne Watkins

provided indispensable secretarial assistance during the preparation of this manuscript. My sincerest apologies are extended to colleagues whose publications are inadequately cited due to constraints of time and space.

REFERENCES 1. Abbott, N. J.: Inflammatory mediators and modulation of blood-brain barrier permeability. Cell. Mol. Neurobiol. 20:131, 2000. 2. Abbott, N. J.: Astrocyte-endothelial interactions and bloodbrain barrier permeability. J. Anat. 200:629, 2002. 3. Abbott, N. J.: Evidence for bulk flow of brain interstitial fluid: significance for physiology and pathology. Neurochem Int. 45:545, 2004. 4. Akert, K., Sandri, C., Weibel, E. R., et al.: The fine structure of the perineurial endothelium. Cell Tissue Res. 165:281, 1976. 5. Anderson, C. M., and Nedergaard, M.: Astrocyte-mediated control of cerebral microcirculation. Trends Neurosci. 26:340, 2003. 6. Bell, M. A., and Weddell, A. G. M.: A morphometric study of intrafascicular vessels in mammalian sciatic nerve. Muscle Nerve 7:524, 1984. 7. Boddingius, J.: Ultrastructural and histophysiological studies on the blood-nerve barrier and perineurial barrier in leprosy neuropathy. Acta Neuropathol. (Berl.) 64:282, 1984. 8. Bradbury, M. W. B., and Crowder, J.: Compartments and barriers in the sciatic nerve of the rabbit. Brain Res. 103:515, 1976. 9. Bundgaard, M., and Frokjaer-Jensen, J.: Functional aspects of the ultrastructure of terminal blood vessels: a quantitative study on consecutive segments of the frog mesenteric microvasculature. Microvasc. Res. 23:1, 1982. 10. Butt, A. M., Jones, H. C., and Abbott, N. J.: Electrical resistance across the blood-brain barrier in anaesthetized rats: a developmental study. J. Physiol. (Lond.) 429:47, 1990. 11. Crone, C.: The blood-brain barrier: a modified tight epithelium. In Suckling, A. J., Rumsby, M. G., and Bradbury, M. W. B. (eds.): The Blood-Brain Barrier in Health and Disease. Chichester, UK, Ellis Harwood Ltd, p. 17, 1986. 12. Crone, C., and Levitt, D. G.: Capillary permeability to small solutes. In Renkin, E. M., and Michel, C. C. (eds.): Handbook of Physiology, Section 2. The Cardiovascular System, Vol. IV. Microcirculation. Bethesda, MD, American Physiological Society, p. 411, 1984. 13. Crowley, N. R., Nelson, S. L., Weerasuriya, A., and Black, A. C.: Endoneurial hydrostatic pressure in rat sciatic nerve during development. Soc. Neurosci. Abstr. 22:772, 1996. 14. Enerback, L., Olsson, Y., and Sourander, P.: Mast cells in normal and sectioned peripheral nerve. Z. Zellforsch. 66:596, 1965. 15. Froehner, S. C., Davies, A., Baldwin, S. A., and Lienhard, G. E.: The blood-nerve barrier is rich in glucose transporter. J. Neurocytol. 17:173, 1988.

Blood-Nerve Interface and Endoneurial Homeostasis 16. Gerhart, D. Z., and Drewes, L. R.: Glucose transporters at the blood-nerve barrier are associated with perineurial cells and endoneurial microvessels. Brain Res. 508:46, 1990. 17. Gloor, S. M., Wachtel, M., Bolliger, M. F., et al.: Molecular and cellular permeability control at the blood-brain barrier. Brain Res. Rev. 36:258, 2001. 18. Griffin, J. W.: Vasculitic neuropathies. Rheum. Dis. Clin. North Am. 27:751, 2001. 19. Grotte, G.: Passage of dextran molecules across the bloodlymph barrier. Acta Chir. Scand. 211:1, 1956. 20. Haller, F. R., Haller, C., and Low, F. N.: The fine structure of cellular layers and connective tissue space at spinal nerve root attachment in the rat. Am. J. Anat. 133:109, 1972. 21. Haller, F. R., and Low, F. N.: The fine structure of the peripheral nerve root sheath in the subarachnoid space in the rat and other laboratory animals. Am. J. Anat. 131:1, 1971. 22. Karnovsky, M. J.: Morphology of capillaries with special reference to muscle capillaries. In Crone, C., and Lassen, N. A. (eds.): Capillary Permeability. Copenhagen, Munksgaard, p. 341, 1970. 23. Kristensson, K., and Olsson, Y.: The perineurium as a diffusion barrier to protein tracers: differences between mature and immature animals. Acta Neuropathol. (Berl.) 17:127, 1971. 24. Krnjevic, K.: Some observations on perfused frog sciatic nerves. J. Physiol. (Lond.) 123:338, 1954. 25. Latker, C. H., Wadhwani, K. C., Balbo, A., and Rapoport, S. I.: Blood-nerve barrier in the frog during wallerian degeneration: are axons necessary for maintenance of barrier function? J. Comp. Neurol. 309:650, 1991. 26. Lo, E. H., Broderick, J. P., and Moskowitz, M. A.: tPA and proteolysis in the neurovascular unit. Stroke 35:354, 2004. 27. Low, F. N.: The perineurium and connective tissue of peripheral nerve. In Landon, D. N. (ed.): The Peripheral Nerve. London, Chapman & Hall, p. 159, 1976. 28. Low, P. A.: Endoneurial potassium is increased and enhances spontaneous activity in regenerating mammalian nerve fibers: implications for neuropathic positive symptom. Muscle Nerve 8:27, 1985. 29. Low, P. A., and Dyck, P. J.: Increased endoneurial fluid pressure in experimental lead neuropathy. Nature 269:427, 1977. 30. Low, P. A., Marchand, G., Knox, F., and Dyck, P. J.: Measurement of endoneurial fluid pressure with polyethylene matrix capsule. Brain Res. 122:373, 1977. 31. Low, P. A., and Tuck, R. R.: Effects of changes in blood pressure, respiratory acidosis and hypoxia on blood flow in the sciatic nerve of the rat. J. Physiol. (Lond.) 347:513, 1984. 32. Mackenzie, M. L., Ghabriel, M. N., and Allt, G.: The bloodnerve barrier: an in vivo lanthanum tracer study. J. Anat. 154:27, 1987. 33. Magnani, P., Cherian, P. V., Gould, G. W., et al.: Glucose transporters in rat peripheral nerve: paranodal expression of GLUT1 and GLUT3. Metabolism 45:1466, 1996. 34. Malmgren, L. T., and Brink, J. J.: Permeability barriers to cytochrome-C in nerves of adult and immature rats. Anat. Rec. 181:755, 1975. 35. McCabe, J. S., and Low, F. N.: The subarachnoid angle: an area of transition in peripheral nerve. Anat. Rec. 164:15, 1969.

663

36. Mellick, R. S., and Cavanagh, J. B.: Longitudinal movement of radio-iodinated albumin within extravascular spaces of peripheral nerves following three systems of trauma. J. Neurol. Neurosurg. Psychiatry 30:458, 1967. 37. Michel, M. E., Shinowara, N. L., and Rapoport, S. I.: Presence of a blood-nerve barrier within blood vessels of frog sciatic nerve. Brain Res. 299:25, 1984. 38. Mizisin, A. P., and Kalichman, M. W.: Permeability and surface area of the blood-nerve barrier in galactose intoxication. Brain Res. 618:109, 1993. 39. Morris, J. H., Hudson, A. R., and Weddell, G.: A study of degeneration and regeneration in the divided rat sciatic nerve based on electron microscope observations. Parts I–III. Z. Zellforsch. 123:76, 1972. 40. Myers, R. R., Heckman, H. M., Galbraith, J. A., and Powell, H. C.: Subperineurial demyelination associated with reduced nerve blood flow and oxygen tension after epineurial vascular stripping. Lab. Invest. 65:41, 1991. 41. Myers, R. R., Powell, H. C., Costello, M. L., et al.: Endoneurial fluid pressure: direct measurement with micropipettes. Brain Res. 148:510, 1978. 42. Myers, R. R., Powell, H. C., Shapiro, H. M., et al.: Changes in endoneurial fluid pressure, permeability and peripheral nerve ultrastructure in experimental lead neuropathy. Ann. Neurol. 8:392, 1980. 43. Myers, R. R., Rydevik, B. L., Heckman, H. M., and Powell, H. C.: Proximodistal gradient in endoneurial fluid pressure. Exp. Neurol. 102:368, 1988. 44. Nesbitt, J. A., and Acland, R. D.: Histopathological changes following removal of the perineurium. J. Neurosurg. 53:233, 1980. 45. Nicely, M.: Measurement of the potential difference across the connective tissue sheath of frog sciatic nerve. Experientia 8:199, 1955. 46. Nukada, H., Powell, H. C., and Myers, R. R.: Perineurial window: demyelination in nonherniated endoneurium with reduced nerve blood flow. J. Neuropathol. Exp. Neurol. 51:523, 1992. 47. Ohi, T., Poduslo, J. F., Curran, G. L., and Dyck, P. J.: Quantitative methods for detection of blood-nerve barrier alterations in experimental animal models of neuropathy. Exp. Neurol. 90:365, 1985. 48. Ohno, K., Pettigrew, K., and Rapoport, S. I.: Lower limits of cerebrovascular permeability to nonelectrolytes in the conscious rat. Am. J. Physiol. Heart Circ. Physiol. 235:H299, 1978. 49. Olsson, Y.: Studies on vascular permeability in peripheral nerve I. Acta Neuropathol. (Berl.) 7:1, 1966. 50. Olsson, Y.: Studies on vascular permeability in peripheral nerve II. Acta Physiol. Scand. 284:1, 1966. 51. Olsson, Y.: Vascular permeability in the peripheral nervous system. In Dyck, P. J., Thomas, P. K., Lambert, E. H., and Bunge, R. P. (eds.): Peripheral Neuropathy, 2nd ed. Philadelphia, W. B. Saunders, p. 579, 1984. 52. Olsson, Y., and Reese, T. S.: Permeability of vasa nervorum and perineurium in mouse sciatic nerve studied by fluorescence and electron microscopy. J. Neuropathol. Exp. Neurol. 30:105, 1971. 53. Pappenheimer, J. R.: Passage of molecules through capillary walls. Physiol. Rev. 33:387, 1953.

664

Functional Compartments and Nerve-Fiber Environment

54. Poduslo, J. F., Curran, G. L., and Dyck, P. J.: Increase in albumin, IgG and IgM blood-nerve barrier indices in human diabetic neuropathy. Proc. Natl. Acad. Sci. U. S. A. 88:4879, 1988. 55. Polydefkis, M., Griffin, J. W., and McArthur, J.: New insights into diabetic polyneuropathy. JAMA 290;1371, 2003. 56. Powell, D. W.: Barrier function of epithelia. Am. J. Physiol. Gastrointest. Liver Physiol. 241:G275, 1981. 57. Powell, H. C., and Myers, R. R.: The blood-nerve barrier and the pathological significance of nerve edema. In Neuwelt, E. A. (ed.): Implications of the Blood-Brain Barrier and Its Manipulation, Vol. 1. New York, Plenum Press, p. 199, 1989. 58. Powell, H. C., and Myers, R. R.: Impact of inflammatory disease on the nerve microenvironment. J. Neurol. Sci. 220:131, 2004. 59. Powell, H. C., Myers, R. R., and Lampert, P. W.: Changes in Schwann cells and vessels in lead neuropathy. Am. J. Pathol. 109:193, 1982. 60. Rapoport, S. I.: Blood-Brain Barrier in Physiology and Medicine. New York, Raven Press, 1976, p. 45. 61. Reale, E., Luciano, L., and Spitznas, M.: Freeze-fracture faces of the perineurial sheath of the rabbit sciatic nerve. J. Neurocytol. 4:261, 1975. 62. Rechthand, E., Smith, Q. R., and Rapoport, S. I.: Facilitated transport of glucose from blood into peripheral nerve. J. Neurochem. 45:957, 1985. 63. Rechthand, E., Smith, Q. R., and Rapoport, S. I.: Transfer of nonelectrolytes from blood into peripheral nerve endoneurium. Am. J. Physiol. Heart Circ. Physiol. 252:H1175, 1987. 64. Rechthand, E., Smith, Q. R., and Rapoport, S. I.: A compartmental analysis of solute transfer and exchange across blood-nerve barrier. Am. J. Physiol. 255:317, 1988. 65. Reed, R. K., Johansen, S., and Noddeland, H.: Turnover rate of interstitial albumin in rat skin and skeletal muscle: effect of limb movements and motor activity. Acta Physiol. Scand. 125:711, 1985. 66. Renkin, E. M.: Multiple pathways of capillary permeability. Circ. Res. 41:735, 1977. 67. Rippe, B., and Haraldsson, B.: How are macromolecules transported across the capillary wall? News Physiol. Sci. 2:135, 1987. 68. Rundquist, I., Smith, Q. R., Michael, M. E., et al.: Sciatic nerve blood flow measured by laser Doppler flowmetry and [14C] iodoantipyrine. Am. J. Physiol. Heart Circ. Physiol. 248:H311, 1985. 69. Rydevik, B. L., Kwan, M. K., Myers, R. R., et al.: Effects of acute stretching on rabbit tibial nerve: an in vitro mechanical and histological study. J. Orthop. Res. 8:694, 1990. 70. Schmelzer, J. D., and Low, P. A.: The effects of hyperbaric oxygenation and hypoxia on the blood-nerve barrier. Brain Res. 473:321, 1988. 71. Shantha, T. S., and Bourne, G. H.: The “perineurial epithelium,” a metabolically active, continuous, protoplasmic cell barrier surrounding peripheral nerve fasciculi. J. Anat. 96:527, 1962. 72. Shinowara, N. L., Michel, M. E., and Rapoport, S. I.: Morphological correlates of permeability in the frog

73.

74. 75.

76. 77.

78.

79. 80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

perineurium: vesicles and transcellular channels. Cell Tissue Res. 227:11, 1982. Smith, M. E., Jones, T. A., and Hilton, D.: Vascular endothelial cadherin is expressed by perineurial cells of peripheral nerve. Histopathology 32:411, 1998. Smith, Q. R., and Allen, D. D.: In situ brain perfusion technique. Methods Mol. Med. 89:209, 2003. Spencer, P. S., Weinberg, H. J., and Raines, C. S.: The perineurial window: a new model of focal demyelination and remyelination. Brain Res. 96:323, 1975. Soderfeldt, B.: The perineurium as diffusion barrier to protein tracers. Acta Neuropathol. (Berl.) 27:55, 1974. Stoll, G., Jander, S., and Myers, R. R.: Degeneration and regeneration in the peripheral nervous system: from August Waller’s observation to neuroinflammation. J. Peripher. Nerv. Syst. 7:13, 2002. Stoll, G., Trapp, B. D., and Griffin, J. W.: Macrophage function during wallerian degeneration of rat optic nerve: clearance of degenerating myelin and Ia expression. J. Neurosci. 9:2327, 1989. Sunderland, S.: Nerves and Nerve Injury. Edinburgh, Livingstone, p. 55, 1978. Taylor, A. E., and Granger, D. N.: Exchange of macromolecules across the microcirculation. In Renkin, E. M., and Michel, C. C. (eds.): Handbook of Physiology: Section 2. The Cardiovascular System, Vol. IV. Microcirculation. Bethesda, MD, American Physiological Society, p. 467, 1984. Terho, P. M., Vuorinen, V. S., and Roytta, M.: The endoneurial response to microsurgically removed epi- and perineurium. J. Peripher. Nerv. Syst. 7:155, 2002. Thomas, P. K., and Bhagat, S.: The effect of extraction of the intrafascicular contents of peripheral nerve trunks on perineurial structure. Acta Neuropathol. (Berl.) 43:135, 1978. Thomas, P. K., Ochoa, J., Berthold, C.-H., et al.: Microscopic anatomy of the peripheral nervous system. In Dyck, P. J., and Thomas, P. K. (eds): Peripheral Neuropathy, 3rd ed. Philadelphia, W. B. Saunders, p. 28, 1993. Tsao, J. W., George, E. B., and Griffin, J. W.: Temperature modulation reveals three distinct stages of wallerian degeneration. J. Neurosci. 19:4718, 1999. Tserentsoodol, N., Shin, B. C., Koyama, H., et al.: Immunolocalization of tight junction proteins, occludin and ZO-1, and GLUT1 in the cells of the blood-nerve barrier. Arch. Histol. Cytol. 62:459, 1999. Wadhwani, K. C., Latker, C. H., Balbo, A., and Rapoport, S. I.: Perineurial permeability and endoneurial edema during wallerian degeneration of the frog peripheral nerve. Brain Res. 493:231, 1989. Wadhwani, K. C., Smith, Q. R., and Rapoport, S. I.: Facilitated transport of L-phenylalanine across the bloodnerve barrier of rat sciatic nerve. Am. J. Physiol. Regul. Integr. Comp. Physiol. 258:R1436, 1990. Waksman, B. H.: Experimental study of diphtheritic polyneuritis in the rabbit and guinea pig. III. The bloodnerve barrier in the rabbit. J. Neuropathol. Exp. Neurol. 20:35, 1961. Weerasuriya, A.: Permeability of endoneurial capillaries to K, Na, and Cl and its relation to peripheral nerve excitability. Brain Res. 419:188, 1987.

Blood-Nerve Interface and Endoneurial Homeostasis 90. Weerasuriya, A.: Patterns of change in endoneurial capillary permeability and vascular space during wallerian degeneration. Brain Res. 445:181, 1988. 91. Weerasuriya, A.: Permeabilities of endoneurial capillaries and perineurium of rat sciatic nerve to 22Na during development. Physiologist 33:A100, 1990. 92. Weerasuriya, A.: Changes in the permeability coefficientsurface area product (PS) of the blood-nerve interface (BNI) to sodium in degenerating and regenerating rat sciatic nerve. Paper presented at the 11th Biennial Meeting of the Peripheral Nerve Study Group, Aachen, Germany, July 1993. 93. Weerasuriya, A., Coath, G., and Crowley, N.: Endoneurial Na concentration and hydrostatic pressure in galactose neuropathy. FASEB J. 10:A764, 1996. 94. Weerasuriya, A., Curran, G. L., and Poduslo J. F.: Blood-nerve transfer of albumin and its implications for the endoneurial microenvironment. Brain Res. 494:114, 1989. 95. Weerasuriya, A., Curran, G. L., and Poduslo, J. F.: Developmental changes in blood-nerve transfer of albumin and endoneurial albumin content in rat sciatic nerve. Brain Res. 521:40, 1990. 96. Weerasuriya, A., Curran, G. L., and Poduslo, J. F.: Physiological changes in the sciatic nerve endoneurium of lead intoxicated rats: a model of endoneurial homeostasis. Brain Res. 517:1, 1990. 97. Weerasuriya, A., and Hockman, C. H.: Perineurial permeability to sodium during wallerian degeneration in rat sciatic nerve. Brain Res. 587:327, 1992. 98. Weerasuriya, A., Nelson, S. L., and Crowley, N. R.: Evidence for endoneurial fluid flow from endoneurial hydrostatic pressure measurements in transected and crushed rat sciatic nerves. Soc. Neurosci. Abstr. 24:267, 1998. 99. Weerasuriya, A., Nelson, S. L., and Crowley, N. R.: Development of an entrapment neuropathy is a two-stage process. Paper presented at the Peripheral Nerve Society Meeting, San Diego, July 1999.

665

100. Weerasuriya, A., and Rapoport, S. I.: Endoneurial capillary permeability to [14C] sucrose in frog sciatic nerve. Brain Res. 375:150, 1986. 101. Weerasuriya, A., Rapoport, S. I., and Taylor, R. E.: Modification of permeability of frog perineurium to [14C]sucrose by stretch and hypertonicity. Brain Res. 173:503, 1979. 102. Weerasuriya, A., Rapoport, S. I., and Taylor, R. E.: Ionic permeabilities of the frog perineurium. Brain Res. 191:405, 1980. 103. Weerasuriya, A., Rapoport, S. I., and Taylor, R. E.: Perineurial permeability increases during wallerian degeneration. Brain Res. 192:581, 1980. 104. Weerasuriya, A., Spangler, R. A., Rapoport, S. I., and Taylor, R. E.: AC impedance of the perineurium of the frog sciatic nerve. Biophys. J. 46:167, 1984. 105. Weiss, P., Wang, H., Taylor, A. C., and Edds, M. V.: Proximodistal fluid convection in the endoneurial spaces of peripheral nerves, demonstrated by colored and radioactive (isotope) tracers. Am. J. Physiol. 143:521, 1945. 106. Welch, K., and Davson, H.: The permeability of capillaries of the sciatic nerve of the rabbit to several materials. J. Neurosurg. 36:21, 1972. 107. Williams, P. L., and Hall, S. M.: Chronic wallerian degeneration: an in vivo and ultrastructural study. J. Anat. 109:487, 1971. 108. Windebank, A. J.: Metal neuropathy. In Dyck, P. J., and Thomas, P. K. (eds.): Peripheral Neuropathy. 3rd ed. Philadelphia, W. B. Saunders, p. 1549, 1993. 109. Windebank, A. J., McCall, J. T., Hunder, H. G., and Dyck, P. J.: The endoneurial content of lead related to the onset and severity of segmental demyelination. J. Neuropathol. Exp. Neurol. 39:692, 1980. 110. Zhou, L., and Griffin, J. W.: Demyelinating polyneuropathies. Curr. Opin. Neurol. 16:307, 2003. 111. Zou, Y., Akazawa, H., Qin, Y., et al.: Mechanical stress activates angiotensin II type 1 receptor without the involvement of angiotensin II. Nature Cell Biol. 6:499, 2004.

30 Nerve Blood Flow and Microenvironment PHILIP G. MCMANIS, PHILLIP A. LOW, AND TERRENCE D. LAGERLUND

Overview Unique Aspects of Special Physiology of Nerve Microvasculature Poor Autoregulation of Peripheral Nerve Peripheral Nerve as a NutritiveCapacitance Microvascular System

Peripheral Nerve Adaptation to Hypoxic Stress Regulation of Nerve Blood Flow Intrinsic Mechanisms Extrinsic Mechanisms Experimental Nerve Ischemia Models of Nerve Ischemia

OVERVIEW Peripheral nerve is unique in a number of respects. It is dependent on its communications with its parent cell body situated a great distance away. Indeed, the length:diameter ratio exceeds 109, so that export of nutrients and structural proteins is very inefficient. The function and indeed the survival of peripheral nerve depend also on its interactions with its microenvironment. The focus of this chapter is on experimental nerve ischemia as a particular component of this microenvironment. Because of space limitations, only cursory coverage is provided to other aspects of nerve microenvironment.

UNIQUE ASPECTS OF SPECIAL PHYSIOLOGY OF NERVE MICROVASCULATURE Peripheral nerve axon, especially its distal portion, is extremely long relative to its diameter and is a great distance from its parent cell body. The regulation and maintenance of nerve depend on the bidirectional axonal transport of structural proteins and growth factors from its neuron. It is exquisitely dependent on nerve microenvironment for its blood supply, oxygenation, and nutrition, and for the removal of toxic metabolic products. The phys-

Factors that Modulate the Effects of Nerve Ischemia Influence of Diabetic State on Nerve Ischemia Reperfusion Injury Pathogenesis of Nerve Ischemia

iology of nerve microenvironment has a number of unique characteristics that are discussed in this section.

Poor Autoregulation of Peripheral Nerve Autoregulation refers to the maintenance of constant blood flow, within a range of blood pressures (BPs) (the autoregulated range), when blood pressure is changed. Autoregulation is achieved by varying arteriolar tone and is achieved by myogenic rather than neurogenic mechanisms in most tissues. We demonstrated that peripheral nerve did not autoregulate.56 We found a curvilinear relationship of nerve blood flow (NBF) to mean BP, a characteristic of a passive system. Thus mammalian peripheral nerve autoregulates poorly, if at all,56 a finding that has subsequently been confirmed in anesthetized90 and nonanesthetized rats.99 When the rate of exsanguination is altered, the shape of the NBF:BP curve also changes. With steady-state recordings of blood flow using hydrogen (H2) polarography, a curvilinear relationship occurs.56 When exsanguination is rapid and blood flow is monitored in real time using laser Doppler velocimetry, a linear relationship occurs.100 The difference may relate to the time-dependent response of arterioles to intravascular pressure alterations.11 The concept of autoregulating and nonautoregulating systems is artificial because different organs autoregulate to 667

668

Functional Compartments and Nerve-Fiber Environment

different degrees. Furthermore, for intensively studied tissues such as brain, there is regional, temporal, and segmental heterogeneity of autoregulation. This range of autoregulatory capacities means that there is a momentto-moment distribution and redistribution of blood in such a way that tissues that have the greatest metabolic needs will have relatively the most constant blood flow, whereas tissues with the least needs will have blood flow that varies the most. The answer to the question as to whether a tissue is autoregulated or not should be: relative to what tissue? There is indeed heterogeneity of innervation of microvessels. For instance, dense noradrenergic and peptidergic innervation of vasa nervorum has been demonstrated,3 and dense innervation of small arteries is well known.15,16 Adrenergic innervation is reportedly increased in experimental diabetic neuropathy.25 These apparently disparate observations may relate to the regional and segmental heterogeneity of blood flow regulation described in other tissues.5,9 Regulation of blood flow in these tissues is largely neurogenic in vessels between 50 to 200 ␮m and largely myogenic in vessels less than 50 ␮m.29 In brain, nutritive vessel regulation appears to be metabolic and has been suggested to be due to adenosine.9 There is also considerable variation from one vessel bed to another. For instance, reactivity of carotid and mesenteric vessels is quite different. There is a characteristic NBF:BP curve, flat within the autoregulated range and sloping at either end.

Peripheral Nerve as a Nutritive-Capacitance Microvascular System Nerve microvasculature is physiologically unique in that the capillaries are much larger in caliber than those of the muscle bed so that red cells can pass unimpeded.7 These large capillaries are in continuity with venules and comprise a nutritive capacitance system.100 Such a system is disadvantaged because a small change in blood volume results in a disproportionate change in NBF. The morphologic basis of this system appears to be the large-diameter capillaries, with a median diameter of 8 to 9 ␮m, considerably larger than that of muscular capillaries7 and poorly developed arteriolar smooth muscle.6

Peripheral Nerve Adaptation to Hypoxic Stress Peripheral nerve is metabolically unique, being able to function relatively well on anaerobic metabolism and having powerful adaptive mechanisms. Compared to rat brain, nerve has about 10% of brain’s oxygen requirements but similar energy stores.57,98 When maximally active, nerve increases its energy demands by less than 100%, whereas brain increases severalfold.57 The relatively large energy stores and the low resting and maximal energy expenditure

enable nerve to function quite well on anaerobically generated high-energy phosphate. We have demonstrated that nerve will conduct impulses for many additional minutes when energy substrates are increased, as in diabetes.57 Indeed, Fink and Cairns28 demonstrated that mammalian peripheral nerve will conduct impulses for hours when provided with a limitless supply of glucose. Another strategy of resistance to ischemic conduction failure is a further downregulation of energy-requiring enzymes, a situation that occurs in aging54 and in chronic hypoxia.55 There is also a suggestion that acute hypovolemic stress also results in adaptive mechanisms; peripheral nerve responds by reducing its oxygen consumption acutely.100

REGULATION OF NERVE BLOOD FLOW The regulation of nerve blood flow is relatively well defined by predictions of the Poiseuille equation, which defines flow as a function of its diameter, pressure gradient, capillary length, and viscosity. However, consideration needs to be additionally focused on the status of the perineurium. Of great interest is the role of agents that dynamically change the caliber of the ends of the microvessel (arteriole and venule).

Intrinsic Mechanisms Microvascular flow is predictable from the Poiseuille equation: NBF ⫽ (P1 ⫺ P2) ⫻ (r4/8) ⫻ (␲/L), where NBF varies directly with the fourth power of its radius (r), the viscosity of blood (␲), capillary length (L), and the pressure gradient (P1 ⫺ P2) between the venular end (P2) and the arteriolar end (P1). Strictly speaking, the Poiseuille equation applies only to Newtonian fluids, and blood flow is non-Newtonian, so that the formulation is semiquantitative. The ultimate purpose of microvasculature is to deliver oxygen and remove metabolic waste products. That the Poiseuille equation has practical application is supported by detailed simulation studies on oxygen delivery to nerve.51 In a mathematical simulation of the release, diffusion, and consumption of oxygen in the capillaries and surrounding tissue of peripheral nerve under steady-state conditions, the Krogh-Erlang equation was used to calculate oxygen tension in tissue, and numerical solution of the differential equation governing oxygen release from hemoglobin and diffusion was used to calculate oxygen tension in the capillary. Using average measured values for the parameters of oxygen solubility, diffusion coefficient, capillary diameter, capillary density, NBF, oxygen consumption rate, and arterial oxygen tension in rat peripheral nerve, we calculated the endoneurial oxygen tension as a function of distance from the nearest capillary and distance along the capillary from the arterial end to the venous end (Fig. 30–1). Oxygen tension is highest at the

669

Nerve Blood Flow and Microenvironment

Capillary Length The length of the capillary influences the drop in hydrostatic pressure along the vessel, although the largest dissipation of pressure is along small arteries and terminal arterioles.38 In nearly all tissues studied, 40% to 60% of the reduction in pressure occurs above the first-order arterioles. The second major drop is across terminal arterioles, with little evidence for precapillary sphincters in most tissues.

FIGURE 30–1 Oxygen delivery based on capillary model using Krogh-Erlang equation. (From Lagerlund, T. D., and Low, P. A.: A mathematical simulation of oxygen delivery in rat peripheral nerve. Microvasc. Res. 34:211, 1987, with permission.)

arterial end and lower at the venous end. It also falls off with distance from the capillary. There is a point (lethal corner) comprising the furthest distance from the arteriolar end of the capillary where oxygenation is poorest (see Fig. 30–1). The range of calculated values agreed with experimental measurements obtained from the sciatic nerves of rats. Conditions that adversely affected oxygen delivery include reduced capillary diameter, increased intercapillary distance, reduced blood flow, and reduced arterial oxygen tension. The lower experimentally obtained oxygen tensions in sciatic nerves of diabetic rats could be accounted for reasonably by this model on the basis of a 33% reduction in NBF (consistent with previously measured flow reduction). Viscosity of Blood Hemorheologic mechanisms also have major influences on blood flow and include the concentration of hemoglobin, serum proteins, the presence of abnormal hemoglobins, and factors affecting the coagulability of blood. Blood hematocrit is the major determinant of whole blood viscosity, but erythrocyte aggregability and deformability are also important contributors to blood viscosity.26,81 Capillary Radius A small change in radius results in a major change in NBF because NBF varies with the fourth power of the radius. Capillary diameter is altered in certain peripheral neuropathies. For instance, in diabetic neuropathy, there is endothelial cell hypertrophy and hyperplasia, and intimal and smooth muscle cell proliferation leading to a reduction in lumen diameter,37 as well as capillary thrombosis101 and closure.27

Pressure Gradient A fall in systemic BP will reduce P1 and an increase in venous pressure will increase P2. Particularly in a tissue that autoregulates poorly, P1 is an important regulator of NBF. P1 is altered dynamically by arteriolar tone and passively by blood volume. Less focus has been given to P2, the venular end, which is also vasoregulated, and the status of both P1 and P2 are important. NBF has been found by all recent studies to regress against systemic BP56,90,99,100 in anesthetized and decerebrate nonanesthetized rats.99 Because nerve microvasculature is a nutritive capacitance system, a small reduction in blood volume, as might occur in hypovolemia or exsanguination, results in a disproportionate reduction in NBF100 and precedes the reduction in systemic BP.

Extrinsic Mechanisms Perineurial Pinch and Endoneurial Edema Nerve is supplied by an intrinsic interconnecting system of microvessels. Feeding into the intrinsic system is an epineurial system, the extrinsic system. The two systems are interconnected via arterioles, which traverse nerve perineurium. What is not known is whether the interconnections are so diffuse and uniform that endoneurial ischemia results only when many perforating arterioles are occluded, or whether each segment of nerve has a major controlling influence on its underlying intrinsic nerve blood flow. The former model would be one in which nerve ischemia occurs by a hemodynamic mechanism. The latter would be at least in part a local and regional supply and ischemic mechanism. Myers et al.71 suggested that perineurial distortion, as might occur in nerve edema, may result in distortion of perineurium creating a perineurial pinch mechanism, and provided mathematical modeling data to demonstrate its feasibility. In earlier experimental studies in which peripheral nerve was undercut (to devascularize nerve) for long stretches or the regional supply was ligated, nerve fiber degeneration did not occur.1,10,24,60 However, the focus of these studies had been on the effect of arterial rather than arteriolar occlusion. Additionally, the end point of fiber degeneration may be too crude because hypoxic/ischemic effects, including conduction failure, occur well before fiber degeneration. Epineurial vasoconstriction induced by norepinephrine over a short

670

Functional Compartments and Nerve-Fiber Environment

segment of nerve (2 cm) on the underlying endoneurial blood flow, measured simultaneously in the subperineurial and centrifascicular sites using microelectrode H2 polarography, resulted in a dramatic reduction in NBF in both sites at concentrations as low as 10⫺7 M.40 We excluded a systemic effect of norepinephrine by demonstrating that the local ischemia was unassociated with NBF reduction in the contralateral sciatic nerve or with an increase in plasma norepinephrine above baseline (M. Kihara and P. A. Low, unpublished observations). We interpret our data as indicating that there is local vasoregulation of NBF by epineurial arterioles, but that the small residual blood flow is sufficient to prevent fiber degeneration. Endoneurial edema, by perineurial distortion, may contribute to the NBF reduction demonstrated in experimental galactose neuropathy. Vasoregulation of Peripheral Nerve: Vasoconstrictors Regional NBF is regulated by systemic BP56 and the balance between neural vasoconstrictors and vasodilators (Table 30–1). The major neurotransmitter of mammalian peripheral nerve is norepinephrine,3,85,104 which is confined to the microvasculature of epineurium and is essentially absent in endoneurium.86 Other vasoconstrictors are endothelin and vasopressin. In a study of the role of ␣-adrenergic innervation in nerve vasoregulation,111 NBF was monitored using a number of techniques. We used laser Doppler velocimetry, endoneurial measurement by microelectrode H2 polarography, and epineurial vasoreactivity by computerized videoangiology. Both ␣ agonists and antagonists were applied locally and systemically. Vasoconstriction mediated by ␣ agonists was regularly produced and was blocked by prior treatment with ␣ antagonists. The videoangiologic recordings showed markedly heterogeneous vasoreactivity to norepinephrine along segments of arterioles, suggesting a segregation of ␣ receptors. Additionally, there is also a local noradrenergic regulation of underlying endoneurial NBF by epineurial arterioles. Sympathetic stimulation results in a reduction

in NBF and chemical sympathectomy results in an increase in NBF. Vasoconstriction is known to be mediated by epineurial ␣-adrenergic receptors40,111 including vasopressin, 5-hydroxytryptamine, and endothelin receptors. 110 5-Hydroxytryptamine reduces NBF with a predominant effect on arteriovenous flow.21 In a previous study, we reported an EC50 (estimated concentration causing 50% of maximal constriction) of 10⫺4.9 M based on a doseresponse study of the epineurial application of norepinephrine, a method that generates EC50 values about two orders of magnitude lower than intra-arterial methods.40 The EC50 for endothelin, using identical methodology, was about three orders of magnitude lower at 10⫺8 M.110 An identical value of EC50 for endothelin was reported by Zochodne et al.110 We undertook a dose-effect study of vasopressin on NBF, its interactions with ␣ adrenoreceptors, and its effect on ischemic conduction failure. Topical epineurial application of vasopressin caused a concentrationdependent reduction of NBF (EC50 ⫽ 6.6 ⫻ 10⫺9 M; asymptote ⫽ 73.9% NBF reduction). The topical application of subthreshold concentrations of vasopressin and norepinephrine alone resulted in no change in NBF, but combined application resulted in a dramatic reduction (72.3%), suggesting synergistic action. We also demonstrated that vasoconstriction caused by combined vasopressin and norepinephrine can produce partial conduction block of sciatic-tibial nerve.91 Vasoregulation of Peripheral Nerve: Vasodilators These potent vasoconstrictor actions are balanced by vasodilation and mediated by calcitonin gene–related peptide (CGRP),108 substance P (SP),108 nitric oxide,18,41 and the prostaglandins.42 CGRP fibers are more dense in nerve than SP, and CGRP is a more potent vasodilator.108 In addition to adrenergic innervation, there is also prominent peptidergic innervation of vasa nervorum.3 Prostacyclin (largely confined to endothelial cells) is the major vasodilator and inhibitor of platelet aggregation of microvessels, and thromboxane A2 (largely confined to platelets) has the

Table 30–1. Vasoregulation of Peripheral Nerve Site of Action

Reference(s)

Vasoconstrictors Norepinephrine Endothelin Vasopressin

Epineurial ␣ adrenoreceptors Endothelial cell Epineurial; ? endothelial

Zochodne and Low,111 Kihara and Low40 Zochodne et al.,110 Cameron et al.18 Sasaki and Low91

Vasodilators CGRP SP Nitric oxide Prostaglandins

Endothelium Endothelium Endothelium Endothelium

Zochodne and Ho108 Zochodne and Ho108 Kihara and Low,41 Cameron et al.18 Kihara and Low42

CGRP = calcitonin gene–related peptide; SP = subtance P.

Nerve Blood Flow and Microenvironment

opposite effects. The ratio of prostacyclin to thromboxane A2 is considered to be important in the maintenance of vascular tone.70 We demonstrated that nerve biosynthesis of 6-keto prostaglandin F1␣ (6KPGF1␣), the stable metabolite of prostacyclin, was largely confined to nerve sheath,104 suggesting another mechanism of epineurial regulation of endoneurial nerve blood flow. Observations similar to ours were made in diabetic rat aorta89 and heart, in which altered regulation of phospholipase activity was suggested as the mechanism of reduced endogenous 6KPGF1␣.88 One possible mechanism of reduced prostacyclin biosynthesis is the reduction of nerve norepinephrine. Norepinephrine release results in the increased synthesis and release of prostaglandin I2 (PGI2) metabolites30 by ␣ receptor–mediated and calcium/calmodulin-dependent mechanisms. Calmodulin is known to activate phospholipase A2,105 resulting in a breakdown of membrane phospholipids to generate arachidonic acid,17 whose availability is the rate-limiting step in prostaglandin synthesis.34 Norepinephrine may also be reduced by the generation of reduced oxygen species, which is thought to be increased in chronic experimental diabetes39 and may be generated from norepinephrine.12 Effect of Maturation and Aging on Nerve Blood Flow We have studied the effects of development (ages 2 to 12 weeks)46 and aging (to 30 months)44 in the sciatic nerve of rats using microelectrode H2 polarography. NBF was highest in 2-week-old rats and progressively declined during development. There was an inverse relationship between NBF and mean arterial pressure, which was lowest at 2 weeks of age, gradually increasing through the next 4 weeks and remaining relatively constant thereafter. The progressive decrease in NBF, associated with an increase in BP, occurs because of a progressive increase in endoneurial vascular resistance. This progressive increase in nerve vascular resistance is due to an increase in plasma viscosity and hematocrit, an increase in innervation of arterioles and small arteries, and a decrease in capillary density.46 The higher NBF in immature rats is likely to be a developmentally adaptive mechanism subserving the special metabolic and functional needs of immature nerves. It permits greater blood-nerve exchange of materials to accommodate the greater metabolic needs of rapidly elongating and myelinating axons and proliferating Schwann cells. Rat peripheral nerve is progressively more resistant to ischemic/anoxic conduction failure with increasing age, a phenomenon that is lowest in immature nerves.54 This resistance was paralleled by a progressive age-related decline in endoneurial oxygen consumption. Endoneurial ATP and creatine phosphate concentrations and expenditure were also progressively reduced with age. Anoxia resulted in progressively smaller reductions in these nucleotide phosphates with increasing age to 8 months,

671

after which time little further change occurred. These findings indicate that the major mechanism of resistance to ischemic conduction failure is a progressive decline in energy requirements. At no stage in development was there any compelling evidence of autoregulation of NBF. Adult peripheral nerve, with increasing age, undergoes a progressive increase in resistance to ischemia and reduction in energy utilization. Nerve conduction velocity and amplitude of nerve action potential increased progressively with age to 8 months, after which time no further increases were demonstrated in the rat.54 This increase was associated with greater resistance to ischemic/anoxic conduction failure and a progressive age-related decline in endoneurial consumption of oxygen, ATP, and creatine phosphate. NBF regressed negatively with increasing age, and this decline was associated with an increase in nerve vascular resistance. Twenty minutes of nerve stimulation resulted in an increase in blood flow by about 50% in adult animals, and this response did not decline with increasing age. As indices of oxygen free radical activity, we measured conjugated dienes, hydroperoxides, and norepinephrine from 2 to 30 months. There was a gradual decline of all indices with increasing age. We concluded that NBF declines with aging as a result of reduced microvascular caliber. These vessels retain their hyperemic response, and oxygen free radical activity is less with increasing age.44 These findings, taken together, suggest that increasing age is associated with reduced utilization of high-energy phosphate.

EXPERIMENTAL NERVE ISCHEMIA Models of Nerve Ischemia Unlike richly perfused tissues such as brain, kidney, and heart, in which ischemic injury is readily produced, peripheral nerve is remarkably resistant to both ischemic conduction failure and ischemic fiber degeneration because of its low energy needs and extensive anastomoses.61 Early attempts to produce experimental ischemia by ligating supplying arteries or undercutting nerve were largely unsuccessful. Models of ischemia have, however, subsequently been developed. These include ligation of supplying arteries to the limb nerves, resulting in greater degree of ischemic fiber degeneration at the level of the tibial nerve and less severe alterations of sciatic nerve.32,49,77,83 The severity of ischemic stress can be increased by the use of agents that induce vasoconstriction, such as arachidonic acid,82 norepinephrine,111 and endothelin.110 A highly successful model of nerve ischemia can be induced by the intra-arterial injection of microemboli of sufficient size to occlude supplying capillaires.77 The severity of reduction of NBF and of ischemic fiber degeneration is linearly related to the dose of microemboli.47 For physiologic studies of nerve ischemia and the effects of reperfusion injury, it was

672

Functional Compartments and Nerve-Fiber Environment

necessary to develop a model of near-total yet reversible ischemia. Our model of severe nerve ischemia, produced by ligating collateral arteries followed by the temporary occlusion of the abdominal aorta and both iliac arteries, consistently occluded NBF.94

Factors that Modulate the Effects of Nerve Ischemia The severity of ischemic fiber degeneration is dependent on the severity and duration of the ischemic stress. Nukada and Dyck77 used polystyrene microspheres of a size (14 ␮m) sufficient to plug capillaries and precapillaries. These were injected into the arterial supply of rat sciatic nerves, producing widespread segmental occlusion of capillaries in lower limb nerves. The clinical and pathologic effect was dose related. One million microspheres produced selective capillary occlusion but no nerve fiber degeneration; approximately 6 million microspheres also produced selective capillary occlusion and associated foot and leg weakness, sensory loss, and fiber degeneration beginning in a central core of the distal sciatic nerve; and 30 million microspheres caused both capillary and arterial occlusion and a greater neuropathologic deficit. From these observations, it is inferred that (1) occlusion of isolated precapillaries and capillaries does not produce ischemic fiber degeneration; (2) occlusion of many microvessels results in central fascicular fiber degeneration, indicating that these cores are watershed regions of poor perfusion; and (3) stereotyped pathologic alterations of nerve fibers and Schwann cells are related to dose, anatomic site, and time elapsed since injection. We evaluated the relationship of microsphere dose and blood flow. The dose of microspheres regressed with the degree of hind limb paresis, reduction in NBF, degree of fiber pathology, and ischemic conduction failure of the tibial nerve.47 That ischemic fiber degeneration is due to ischemia is supported by the demonstration that the severity of fiber degeneration regresses inversely with NBF (Fig. 30–2).47 All nerves with severe fiber degeneration had flows of less than 3 mL/100 g/min. Intercapillary distance is a determinant of NBF, and the increased intercapillary distance associated with endoneurial edema results in a corresponding reduction in NBF, as predicted by the Poiseuille equation.64 In experimental nerve ischemia, centrifascicular fiber degeneration is typical, with subperineural sparing.77 The difference in intercapillary distance in centrifascicular regions versus the subperineurial area has been proposed as being responsible,53,63 although diffusion of oxygen from surrounding areas is likely more important.63 A major determinant of the severity of ischemic fiber degeneration is the temperature of nerve during ischemia. Because peripheral nerve has a large ischemic safety factor, hypothermia (by reducing metabolic demands) greatly

FIGURE 30–2 Relationship between severity of ischemic fiber degeneration and nerve blood flow. Severity of fiber degeneration is inversely related to measured perfusion.

reduces the severity of ischemic fiber degeneration. We evaluated the influence of temperature on the effect of ischemia to the sciatic nerve in the rat. Ischemia was produced by embolization of microspheres (diameter 14 ␮m) into the supplying arteries at three temperatures (37°, 32°, and 28° C) and maintaining the nerve at the defined temperature for an additional 4 hours. End points (evaluated 7 days after embolization) of function, electrophysiology, NBF (in mL/100 g/min), and severity of ischemic fiber degeneration demonstrated a close relationship between temperature and severity of nerve fiber degeneration.45 Hypothermia at 28° C dramatically reduced the severity of ischemic fiber degeneration. The timing of hypothermia is important. Its neuroprotective value is much greater during the ischemic insult than during reperfusion, although the latter was partially effective and ameliorated the development of endoneurial edema, presumably by reducing the severity of blood-nerve barrier breakdown.68

Influence of Diabetic State on Nerve Ischemia A consistent finding in experimental diabetic neuropathy is a reduction in NBF. Decreased endoneurial perfusion has been found repetitively (about 100 reports) by 12 programs, although it has been questioned by two investigators.20,109 Table 30–2 summarizes the current status of NBF experimental results. Cameron et al.19 reported that there is a tight relationship between NBF and nerve conduction in experimental diabetic neuropathy (Fig. 30–3). The reduction in NBF is associated with a reduction in endoneurial oxygen tension.102 Chronic hyperglycemia results in a large deficit in NBF. Both auto-oxidative and ischemia-induced lipid peroxidation occurs with resultant peripheral sensory

673

Nerve Blood Flow and Microenvironment

Table 30–2. Results of Studies on Nerve Blood Flow in Experimental Diabetes Investigator/Program

Method

Low, Mayo (Rochester, USA) Powell/Myers (San Diego, USA) Greene (Michigan, USA) Cameron/Cotter (Scotland, UK) Gispen (Netherlands, UK) Hotta (Nagoya, Japan) Stevens (Michigan, USA) Tomlinson (London, UK) Ueno (Kanagawa, Japan) Yasuda (Ohtsu, Japan) Wright/Nukada (New Zealand) Yorek (Wisconsin, USA) Kihara (Nagoya, Japan) Nakamura (Nagoya, Japan) Schratzberger (Boston, USA) Ueno (Japan) Yamamoto (Osaka, Japan) Van Dam (Utrect, Netherlands) Van Buren (Utrect, Netherlands) Zochodne (Calgary, Canada) Williamson (USA)

H2; LDF; iodo-AP LDF H2 H2; LDF LDF LDF; H2 H2 LDF LDF LDF LDF H2 H2 H2 LDF, imaging LDF LDF LDF LDF LDF; H2 Microemboli

Results b b b b b b b b b b b b b b b b b b b ? c

H2 ⫽ hydrogen polarography; iodo-AP ⫽ iodoantipyrine radiography; LDF ⫽ laser Doppler flow velocimetry.

neuropathy in streptozotocin-induced diabetes in the rat. Free radical defenses, especially involving antioxidant enzymes, have been suggested to be reduced, but scant information is available on these defenses in chronic hyperglycemia. We evaluated the gene expression of glu-

FIGURE 30–3 Relationship between nerve blood flow and conduction velocity in experimental diabetic neuropathy.

tathione peroxidase, catalase, and superoxide dismutase (cuprozinc and manganese separately) in L4/5 dorsal root ganglion (DRG) and superior cervical ganglion, as well as enzyme activity of glutathione peroxidase in DRG and sciatic nerve in experimental diabetic neuropathy of 3-month and 12-month durations. We also evaluated nerve electrophysiology of caudal, sciatic-tibial, and digital nerves. A nerve conduction deficit was seen in all nerves in experimental diabetic neuropathy at both 3 and 12 months. Gene expression of glutathione peroxidase, catalase, cuprozinc superoxide dismutase, and manganese superoxide dismutase was not reduced in experimental diabetic neuropathy at either 3 or 12 months. Catalase messenger RNA (mRNA) was significantly increased in experimental diabetic neuropathy at 12 months. Glutathione peroxidase enzyme activity was normal in sciatic nerve. We conclude that gene expression is not reduced in peripheral nerve tissues in very chronic experimental diabetic neuropathy. Changes in enzyme activity may be related to duration of diabetes or due to posttranslational modifications.48 Excessive Susceptibility Nukada first demonstrated excessive vulnerability of diabetic nerves to ischemia in rats that had been diabetic for 4 months.74 He ligated the supplying arteries to diabetic (duration 4 months) rat sciatic nerve and demonstrated

674

Functional Compartments and Nerve-Fiber Environment

FIGURE 30–4 Ischemia-reperfusion injury (3 hours of ischemia, 7 days of reperfusion) to an experimental diabetic neuropathy model showing excessive fiber degeneration and immunolabeling of activated caspase-3, 8-hydroxydeoxyguanosine, and 4-hydroxynonenal in diabetic tibial nerve compared with age-matched controls.

excessive fiber degeneration. Of interest is that, whereas resistance to ischemic fiber degeneration develops early, the excessive vulnerability to ischemic fiber degeneration is time dependent, being seen at 4 months and not apparently present at 2 weeks.75 In Figure 30–4, nerves in a model with experimental diabetic neuropathy (duration of diabetes of 4 months) underwent extensive fiber degeneration in response to 3 hours of ischemia, whereas corresponding control nerve was relatively unaffected. Schwann cells demonstrated prominent immunolabeling to 8-hydroxydeoxyguanosine and hydroxynonenal, indicating oxidative injury (Fig. 30–4). Caspase-3 immunoreactivity indicates commitment to the efferent limb of the apoptotic pathway. Resistance to ischemic conduction failure is due largely to the ability of diabetic nerves to continue conducting longer than normal nerves because of the large increase in energy stores,58 and, because diabetic nerve is partially depolarized,14 there is reduced energy cost involved in impulse generation.87 The pathophysiology of excessive susceptibility to ischemic fiber degeneration is not fully elucidated. Areas of focus have been on the role of vasoconstrictors (increased) and vasodilators (reduced) and the role of oxidative injury and the inflammatory response. Nitric oxide synthase activity is reduced in sciatic nerve of experimental diabetic neuropathy41,80 and is reversed by insulin treatment early on. Experimental diabetes is associated with denervation supersensitivity to vasoconstrictors.107 Application of topical endothelin-1 to sciatic nerve trunk resulted in axonal damage that exceeded that of nondiabetics by a factor greater than 5. The pattern of axonal degeneration was multifocal but not centrofascicular and did not vary with fascicular

area.107 There are focal electrophysiologic alterations in these rats. There is some uncertainty as to whether there is receptor supersensitivity to endothelin107,110 or a change in the balance of nerve vasodilators to vasoconstrictors,52 because some studies have not found, at least in in vitro experiments, supersensitivy to endothelin.41 Limited information is available on the mechanisms of this susceptibility. Using an in vitro model, Wachtler et al.103 studied the role of intracellular pH (pHi) and intracellular calcium concentrations ([Ca2⫹]i). Isolated rat spinal roots were preincubated for 3 to 6 hours in either 5- or 25-mM D-glucose before transient exposure to gaseous hypoxia or cyanide. Intracellular pH and Ca2⫹ concentrations were measured photometrically by means of the fluorescent dyes carboxy-SNARF-1 and a combination of Calcium Green-1 and Fura Red, respectively. They observed that 25-mM D-glucose resulted in stronger intracellular acidification and much slower posthypoxic recovery of pHi as compared to 5-mM D-glucose. From these and the calcium experiments, they concluded that hypoxia-induced nerve acidification rather than a rise in [Ca2⫹]i might contribute to ischemic lesions found in diabetic neuropathy.

Reperfusion Injury Although ischemia alone will cause ischemic fiber degeneration, reperfusion will aggravate ischemic injury to the axon, the Schwann cell, and the blood-nerve barrier by oxidative injury. In rat sciatic-tibial nerve subjected to ischemia sufficient to cause conduction failure (1 or 3 hours of complete ischemia), two interesting observations were found with reperfusion. First, with reperfusion NBF was restored to only 55% and 45% of resting values

Nerve Blood Flow and Microenvironment

following 1 and 3 hours, respectively, indicating reduced reperfusion94 as a result of swelling of endothelial cells.2 Immunolabeling for 8-hydroxydeoxyguanosine, a marker of oxidative DNA injury, is demonstrable for endothelial cells (see Fig. 30–4). Second, the blood-nerve barrier breakdown occurred during reperfusion as a result of oxidative injury.94 We subsequently evaluated the effect of ischemic reperfusion of rat sciatic-tibial nerve seeking biochemical and pathologic evidence of blood-nerve barrier disruption and lipid peroxidation. Ischemia caused by the ligation of the supplying arteries to sciatic-tibial nerve was maintained for 3 hours, followed by reperfusion. Reperfusion resulted in an increase in nerve lipid hydroperoxides, greatest at 3 hours, followed by a gradual decline over the next month. Nerve edema and ischemic fiber degeneration consistently became more severe with reperfusion, indicating that oxidative stress impairs the blood-nerve barrier (edema) and causes ischemic fiber degeneration. Reduced reperfusion was greatest over distal sciatic nerve and mid-tibial nerve at day 7. The most ischemic segment (mid-tibial) of nonreperfused ischemic nerves (duration 3 hours) underwent both edema and ischemic fiber degeneration that were as pronounced as those of other segments after reperfusion and underwent a smaller increase in NBF with reperfusion, suggesting that ischemia alone can also cause ischemic fiber degeneration and edema. The type of fiber degeneration was that of axonal degeneration.72 Free radical–mediated lipid and protein oxidation results in carbonyl formation in nerve. Its presence in tissue can be detected histochemically. Reperfusion following 7 hours of near-complete ischemia resulted in histochemically detectable carbonyl compounds in microvessels of reperfused nerves.2 Endothelial cells were positively stained and swollen. ␣-Tocopherol pretreatment reduced staining and size of endothelial cells.2 In a subsequent study, He et al.,31 using a sensitive enzyme-linked immunosorbent assay, measured protein carbonyl content in experimental ischemia to nerve followed by reperfusion. They found that carbonyl content was unaffected by ischemia alone but increased by 55% after 12 to 18 hours of reperfusion, correlating with the onset of nerve pathology. Pretreatment with the xanthine oxidase inhibitor allopurinol prevented these abnormalities, suggesting that xanthine oxidase activity precedes oxidative damage during reperfusion injury. Nukada et al.76 differentiated the pathologic changes caused by reperfusion from those of ischemia alone. Changes caused by less severe ischemia followed by reperfusion were characterized by vascular swelling, endoneurial and intramyelinic edema, demyelination, and a paucity of axonal degeneration. Breakdown of the blood-nerve barrier was characteristic. Reperfusion results in alterations to the blood-nerve barrier, endothelial cells, and Schwann cells. Part of the response is mediated by an inflammatory response. Nukada et al.79 investigated the time course of

675

acute inflammation and its role in the development of demyelination in reperfused rat sciatic, tibial, and peroneal nerves after a 5-hour period of near-complete ischemia. Polymorphonuclear neutrophil migration was seen early in the endoneurial lesion. After 18 hours of reperfusion, there was maximal intercellular adhesion molecule-1 (ICAM-1) expression on endoneurial vessels, and polymorphonuclear neutrophil accumulation was then prominent, reaching a peak 24 hours after reperfusion. Endoneurial mononuclear macrophages increased nearly fourfold after 48 to 72 hours of reperfusion. Macrophages were observed invading Schwann cells and myelinated lamellae, with associated demyelination. Thus this study provides evidence of macrophage-associated demyelination after reperfusion similar to that seen in inflammatory neuropathies. We evaluated the mRNA expression of the pro-inflammatory cytokines tumor necrosis factor-␣ (TNF-␣) and interleukin-1␤ (IL-1␤) in rat sciatic and tibial nerves following ischemia-reperfusion injury, using competitive reverse transcriptase–polymerase chain reaction, to explore the role of cytokines in ischemia-reperfusion injury.66 The expressions of both TNF-␣ and IL-1␤ mRNA were related to severity of ischemia and occurred with reperfusion rather than ischemia alone. TNF-␣ gene expression peaked at 24 hours of reperfusion, whereas that of IL-1␤ peaked at 12 hours. These data support the notion that the proinflammatory cytokines TNF-␣ and IL-1␤ are involved in the inflammatory response of ischemia-reperfusion injury to the peripheral nervous system and may be involved in the pathophysiology of ischemic fiber degeneration. Nitric oxide synthase alterations also occur during ischemia-reperfusion. In a study evaluating mRNA and protein expressions of neuronal (nNOS), inducible (iNOS), and endothelial (eNOS) nitric oxide synthases in peripheral nerve after ischemia-reperfusion, the investigators reported an interesting pattern of isoform expression.84 They evaluated the effects of ischemia (2 hours) followed by reperfusion (0 or 3 hours). Ischemia alone resulted in a reduction in both nNOS and eNOS and their protein expressions. After ischemia-reperfusion (2 hours of ischemia followed by 3 hours of reperfusion), expression of both nNOS and eNOS mRNA and protein decreased further. In contrast, iNOS mRNA significantly increased after ischemia and was further upregulated (14-fold) after reperfusion, whereas iNOS protein was not detected. The results highlight the importance of the inflammatory response associated with reperfusion following ischemia. Recently, we explored the recovery process by extending the period of observation following reperfusion to 42 days.33 Pathologically, three phases were identifiable: phase 1 (0 to 3 hours): minimal pathologic changes and minimal edema; phase 2 (7 days, 14 days): prominent fiber degeneration and endoneurial edema; and phase 3 (28 days, 42 days): abundant small regenerating fiber clusters and minimal edema. Compound muscle action potential was the most sensitive

676

Functional Compartments and Nerve-Fiber Environment

index of neural deficits and recovery, showing progressive recovery beyond 14 days. Severe functional deficits developed immediately and persisted, with a trend to recovery at the 42-day time point. It was concluded that reperfusion by oxidative injury worsened nerve function and aggravated fiber degeneration, but in the longer time frame permitted fiber regeneration to occur. Neuroprotection Knowledge of the pathogenesis of ischemia-reperfusion injury has naturally led to studies designed to minimize the effects of this injury. Most of the studies to date have focused on hypothermia, antioxidants, and reducing the inflammatory response. Because peripheral nerve has a large ischemic safety factor, hypothermia (by reducing metabolic demands) is potentially an efficacious technique to rescue nerve from ischemic fiber degeneration. We therefore evaluated the influence of temperature on the severity of ischemic fiber degeneration in sciatic nerve resulting from embolization with microspheres (14 ␮m) into its supplying arteries. The limb was embolized at three temperatures (37°, 32°, and 28° C) and was maintained at each temperature for an additional 4 hours. End points, evaluated 7 days after embolization, for the embolized limb were (1) behavioral scores (0 to 11, in increasing limb function); (2) compound nerve action potential of sciatic-tibial nerve; (3) sciatic NBF (in mL/100 g/min); and (4) histologic grade, expressed as percentage of fibers undergoing ischemic fiber degeneration. NBF was reduced in all groups varying with temperature, and all indices of nerve structure and function were significantly improved with hypothermia. We concluded that hypothermia, easily achievable in a limb nerve, is highly efficacious in the rescue of nerve from ischemic fiber degeneration.45 We subsequently evaluated the effectiveness of hypothermia in protecting peripheral nerve from ischemia followed by reperfusion injury to rat peripheral nerve. Ischemia was caused by ligation of supplying arteries and collateral vessels, and reperfusion resulted when the ligatures were released after a predetermined period of ischemia (3 or 5 hours). The effects of temperature (28°, 32°, and 37° C) were evaluated. End points were behavioral score, nerve electrophysiology, and nerve histology. The groups treated at 36° to 37° C underwent marked fiber degeneration associated with a reduction in action potential and impairment in behavioral score. The groups treated at 28° C (for both 3 and 5 hours) showed significantly less (P ⬍ .01; analysis of variance, Bonferoni post hoc test) reperfusion injury for all indices (behavioral score, electrophysiology, and neuropathology), and the groups treated at 32° C had scores intermediate between the groups treated at 36° to 37° C and at 28° C. Our results showed that cooling the limbs dramatically protects the peripheral nerve from ischemia-reperfusion injury.67 Once hypothermia was applied in the ischemic period, the

resultant neuroprotection continued into the reperfusion period, even if nerve temperature was then raised during the reperfusion period. These results indicate that hypothermic neuroprotection is dramatically more efficacious during the intra-ischemic period than during reperfusion, when a lesser degree of neuroprotection ensued.68 As part of our focus on improving peripheral nerve salvage from ischemic fiber degeneration, we evaluated whether hyperbaric oxygen (HBO), by increasing delivery of oxygen to ischemic nerve, will rescue peripheral nerve rendered ischemic by microembolization. HBO appears to be less effective than hypothermia but is effective in improving behavioral, electrophysiologic, and morphologic indices of nerve function.43 We concluded that HBO will effectively rescue fibers from ischemic fiber degeneration, providing the ischemia is not extreme. The potent antioxidant ␣-lipoic acid was demonstrated to provide neuroprotection from ischemic fiber degeneration resulting from ischemia-reperfusion injury when the ischemic insult was not overwhelming.66 Distal sensory conduction was significantly improved in the 3-hour ischemia group treated with ␣-lipoic acid (P ⬍ .05). The antioxidant failed to show favorable effects if the duration of ischemia was longer (5 hours of ischemia). These results suggest that ␣-lipoic acid is efficacious for moderate ischemia-reperfusion, especially on distal sensory nerves. In a study of mild ischemic rat facial nerve injury induced by embolization of thrombin-coated microspheres,69 rapid and sustained labeling of facial nuclei (peaking by day 1 and lasting 8 days) of mRNA of c-Jun and growth-associated protein-43 was seen, followed by a delayed peak (day 3) for CGRP. Superoxide dismutase alleviated the behavioral impairment and decreased the CGRP mRNA expression at the third day after injury. The protocol for facial nerve paresis results in ischemia of both axons and facial neurons.69

PATHOGENESIS OF NERVE ISCHEMIA Data are gradually accumulating on the molecular pathogenesis of nerve ischemic fiber degeneration. There is a complex interplay of ischemia, eicosanoids, oxidative injury, the inflammatory response, and norepinephrine. Ischemia would result in phospholipase activation and phospholipid breakdown, liberating arachidonic acid with its cascade to produce the prostaglandins and leukotrienes. Reduced oxygen species are generated during ischemia and especially during reperfusion by mechanisms described earlier. The formation of lipid hydroperoxides would inhibit prostacyclin synthetase. The increased biosynthetic rate of thromboxane A2 coupled with the ischemia-related inhibition of prostacyclin production by endothelial cells would result in vasoconstriction, aggravating the ischemic insult. In nerve, there is a breakdown of the blood-nerve barrier after 3 hours of ischemia but not after 1 hour.

677

Nerve Blood Flow and Microenvironment

Reperfusion after 1 hour of ischemia will, however, regularly cause a breakdown of the barrier with a reduction in NBF94 as a result of endothelial cell swelling, consistently found in areas associated with demyelination.78 We used the blood-nerve barrier as a physiologic index of the oxidative injury, which is known to increase microvascular permeability.23,50 The permeability–suface area product may be increased as a result of an increase in permeability or an increase in the surface area of endoneurial capillaries. In this study, NBF remained persistently reduced with reperfusion. Therefore, the progressive increase in the product must be due to an increase in permeability. Our results suggest that reperfusion injury of nerve requires near-total ischemia of several hours’ duration, because it occurs following 3 hours but not 1 hour of occlusion. Using this paradigm, we demonstrated that reperfusion resulted in an increase in nerve lipid hydroperoxides, greatest at 3 hours, followed by a gradual decline over the next month.72 Endoneurial edema, indicative of bloodnerve barrier breakdown, consistently became more severe with reperfusion, indicating that oxidative stress impairs the barrier. Protein carbonyl content was unaffected by ischemia alone but increased by 55% after 12 to 18 hours’ reperfusion, correlated with the onset of nerve pathology, suggesting a possible role of peroxynitrite.2,31 Hypoxia/ischemia results in an increase in tissue reducing equivalents57,73 and an increase in tissue xanthine.36 Ischemia or hypoxia has been suggested to activate a calcium-dependent protease (possibly calpain), which converts cytosolic xanthine dehydrogenase to xanthine oxidase.4 This important reaction occurs in capillary endothelium.35 The above three conditions create an oxygen free radical–generating system. Furthermore, reperfusion accelerates reduced oxygen species formation by introducing oxygen into a system primed and generating reduced oxygen species.62 In nerve, pretreatment with the xanthine oxidase inhibitor allopurinol prevented these abnormalities, suggesting that xanthine oxidase activity precedes oxidative damage during reperfusion injury.31 Nukada et al.79 further explored the role of the adhesion molecule ICAM-1 and the inflammatory response. Polymorphonuclear neutrophil migration was seen early in the endoneurial lesion. After 18 hours of reperfusion, there was maximal ICAM-1 expression on endoneurial vessels, and polymorphonuclear neutrophil accumulation was then prominent, reaching a peak 24 hours after reperfusion. Endoneurial mononuclear macrophages increased nearly fourfold after 48 to 72 hours of reperfusion. Macrophages were observed invading Schwann cells and myelinated lamellae, with associated demyelination. Ischemia induces TNF-␣ mRNA in nerve,66 which in turn activates the nuclear factor (NF)-␬ pathway, causing oxidative stress.13 The expressions of both TNF-␣ and IL-1␤ mRNA were related to severity of ischemia and

occurred with reperfusion rather than ischemia alone.65 TNF-␣ gene expression peaked at 24 hours of reperfusion, whereas that of IL-1␤ peaked at 12 hours. These data support the notion that the pro-inflammatory cytokines TNF-␣ and IL-1␤ are involved in the inflammatory response of ischemia-reperfusion injury to the peripheral nervous system and may be involved in the pathophysiology of ischemic fiber degeneration. There is some indirect evidence for intracellular calcium accumulation with ischemia in peripheral nerve. Calcium ionophore will cause vesicular disruption of myelin93,96 and axonal degeneration,92 and nerve reconnection in a calciumfree medium resulted in improved functional recovery.22 The vasoconstriction and microvascular ischemia/hypoxia in disorders such as diabetes have been in part ascribed to perturbations of prostaglandins and reduced oxygen species generation. Lipid hydroperoxides are increased and inhibit prostacyclin synthetase activity, resulting in a reduced prostacyclin:thromboxane ratio and vasoconstriction and platelet aggregation.106 Three abnormalities in PGI2 have been described in diabetes. Apart from a reduced synthesis of PGI2, serum from patients with diabetes or rats with streptozotocin-induced diabetes contains less PGI2 releasestimulating factor or may have more of an inhibitory factor.59 The presence of a stimulating factor appears to be well established in normal serum.95,97 Third, platelets of diabetic patients have reduced responsiveness to PGI2,8 and dipyridamole will restore reponsiveness.39 The current pathogenetic schema of the pathogenesis of nerve ischemia-reperfusion injury may be summarized as follows. There is energy run-down,112 poly(ADP-ribose) polymerase (PARP) activation, NF-␬B activation, and activation of adhesion molecules including ICAM-1. At this time, it is known that there is immunolabeling of endothelial cells with ICAM-1, PARP, and 8-hydroxydeoxyguanosine (indicating oxidative DNA injury). A prominent inflammatory infiltration with polymorphs is seen. Next there is a switch to mononuclear cells, upregulation of iNOS, and increased expression of certain cytokines, including TNF-␣.65 These changes occur between 48 hours and 1 week. There is immunolabeling of Schwann cells with 8-hydroxydeoxyguanosine and caspase-3 and terminal deoxynucleotidyl transferase–mediated dUTP nick end-labeling (TUNEL) positivity, indicating commitment to the apoptotic pathway (see Fig. 30–4).

REFERENCES 1. Adams, W. E.: The blood supply of nerves. II. The effects of exclusion of its regional sources of supply on the sciatic nerve of the rabbit. J. Anat. 77:243, 1943. 2. Anderson, G. M., Nukada, H., and McMorran, P. D.: Carbonyl histochemistry in rat reperfusion nerve injury. Brain Res. 772:156, 1997.

678

Functional Compartments and Nerve-Fiber Environment

3. Appenzeller, O., Dhital, K. K., Cowen, T., and Burnstock, G.: The nerves to blood vessels supplying blood to nerves: the innervation of vasa nervorum. Brain Res. 304:383, 1984. 4. Battelli, M. G., Corte, E. D., and Stirpe, F.: Xanthine oxidase type d (dehydrogenase) in the intestine and other organs of the rat. Biochem. J. 126:747, 1972. 5. Baumbach, G. L., and Heistad, D. D.: Regional, segmental, and temporal heterogeneity of cerebral vascular autoregulation [review]. Ann. Biomed. Eng. 13:303, 1985. 6. Bell, M. A., and Weddell, A. G.: A morphometric study of intrafascicular vessels of mammalian sciatic nerve. Muscle Nerve 7:524, 1984. 7. Bell, M. A., and Weddel, A. G. M.: A descriptive study of the blood vessels of the sciatic nerve in the rat, man and other mammals. Brain 107:871, 1984. 8. Betteridge, D. J., El Tahir, K. E., Reckless, J. P., and Williams, K. I.: Platelets from diabetic subjects show diminished sensitivity to prostacyclin. Eur. J. Clin. Invest. 12:395, 1982. 9. Bevan, J. A.: Autonomic pharmacologist’s guide to the cerebral circulation. Trends Pharmacol. Sci. 5:234, 1984. 10. Blunt, M. J., and Stratton, K.: The immediate effects of ligature of vasa nervorum. J. Anat. 90:204, 1956. 11. Borgstrom, P., Grande, P. O., and Lindbom, L.: Responses of single arterioles in vivo in cat skeletal muscle to change in arterial pressure applied at different rates. Acta Physiol. Scand. 113:207, 1981. 12. Boveris, A., and Chance, B.: The mitochondrial generation of hydrogen peroxide: general properties and effect of hyperbaric oxygen. Biochem. J. 134:707, 1973. 13. Bowie, A., and O’Neill, L. A.: Oxidative stress and nuclear factor-kappaB activation: a reassessment of the evidence in the light of recent discoveries. Biochem. Pharmacol. 59:13, 2000. 14. Brismar, T., Sima, A. A., and Greene, D. A.: Reversible and irreversible nodal dysfunction in diabetic neuropathy. Ann. Neurol. 21:504, 1987. 15. Burnstock, G.: Innervation of vascular smooth muscle: histochemistry and electron microscopy. Clin. Exp. Pharmacol. Physiol. Suppl. 2:7, 1975. 16. Burnstock, G., Griffith, S. G., and Sneddon, P.: Autonomic nerves in the precapillary vessel wall. J. Cardiovasc. Pharmacol. 6(Suppl.):344, 1984. 17. Burton, K. P., Buja, L. M., Sen, A., et al.: Accumulation of arachidonate in triacylglycerols and unesterified fatty acids during ischemia and reflow in the isolated rat heart: correlation with the loss of contractile function and the development of calcium overload. Am. J. Pathol. 124:238, 1986. 18. Cameron, N. E., Cotter, M. A., and Hohman, T. C.: Interactions between essential fatty acid, prostanoid, polyol pathway and nitric oxide mechanisms in the neurovascular deficit of diabetic rats. Diabetologia 39:172, 1996. 19. Cameron, N. E., Eaton, S. E., Cotter, M. A., and Tesfaye, S.: Vascular factors and metabolic interactions in the pathogenesis of diabetic neuropathy. Diabetologia 44:1973, 2001. 20. Chang, K., Ido, Y., LeJeune, W., et al.: Increased sciatic nerve blood flow in diabetic rats: assessment by “molecular” vs. particulate microspheres. Am. J. Physiol. 273:E164, 1997. 21. Day, T. J., Lagerlund, T. D., and Low, P. A.: Analysis of H2 clearance curves used to measure blood flow in rat sciatic nerve. J. Physiol. (Lond.) 414:35, 1989.

22. de Medinaceli, L., Wyatt, R. J., and Freed, W. J.: Peripheral nerve reconnection: mechanical, thermal, and ionic conditions that promote the return of function. Exp. Neurol. 81:469, 1983. 23. Del Maestro, R. F., Bjork, J., and Arfors, K. E.: Increase in microvascular permeability induced by enzymatically generated free radicals. II. Role of superoxide anion radical, hydrogen peroxide, and hydroxyl radical. Microvasc. Res. 22:255, 1981. 24. Denny-Brown, D., and Brenner, C.: Paralysis of nerve induced by direct pressure and by tourniquet. Arch. Neurol. Psychiatry 51:1, 1944. 25. Dhital, K., Lincoln, J., Appenzeller, O., and Burnstock, G.: Adrenergic innervation of vasa and nervi nervorum of optic, sciatic, vagus and sympathetic nerve trunks in normal and streptozotocin-diabetic rats. Brain Res. 367:39, 1986. 26. Dintenfass, L.: Haemorheology of diabetes mellitus. J. Pharmacol. Exp. Ther. 278:1262, 1979. 27. Dyck, P. J., Hansen, S., Karnes, J., et al.: Capillary number and percentage closed in human diabetic sural nerve. Proc. Natl. Acad. Sci. U. S. A. 82:2513, 1985. 28. Fink, B. R., and Cairns, A. M.: Differential tolerance of mammalian myelinated and unmyelinated nerve fibers to oxygen lack. Reg. Anesth. 7:2, 1982. 29. Folkow, B.: Autoregulation in muscle. Circ. Res. 15:19, 1964. 30. Gilmore, N., Vane, J. R., and Wyllie, J. H.: Prostaglandins released by the spleen. Nature 218:1135, 1968. 31. He, K., Nukada, H., McMorran, P. D., and Murphy, M. P.: Protein carbonyl formation and tyrosine nitration as markers of oxidative damage during ischaemia-reperfusion injury to rat sciatic nerve. Neuroscience 94:909, 1999. 32. Hess, K., Eames, R. A., Darveniza, P., and Gilliatt, R. W.: Acute ischaemic neuropathy in the rabbit. J. Neurol. Sci. 44:19, 1979. 33. Iida, H., Schmelzer, J. D., Schmeichel, A. M., et al.: Peripheral nerve ischemia: reperfusion injury and fiber regeneration. Exp. Neurol. 184:997, 2003. 34. Irvine, R. F.: How is the level of free arachidonic acid controlled in mammalian cells? Biochem. J. 204:3, 1982. 35. Jarasch, E. D., Grund, C., Bruder, G., et al.: Localization of xanthine oxidase in mammary-gland epithelium and capillary endothelium. Cell 25:67, 1981. 36. Jennings, R. B. and Reimer, K. A.: Lethal myocardial ischemic injury. Am. J. Pathol. 102:241, 1981. 37. Johnson, P. C., Doll, S. C., and Cromey, D. W.: Pathogenesis of diabetic neuropathy. Ann. Neurol. 19:450, 1986. 38. Joyner, W. L., and Davis, M. J.: Pressure profile along the microvascular network and its control. Fed. Proc. 46:266, 1987. 39. Karpen, C. W., Pritchard, K. A. Jr., Merola, A. J., and Panganamala, R. V.: Alterations of the prostacyclinthromboxane ratio in streptozotocin induced diabetic rats. Prostaglandins Leukot. Med. 8:93, 1982. 40. Kihara, M., and Low, P. A.: Regulation of rat nerve blood flow: role of epineurial alpha- receptors. J. Physiol. (Lond.) 422:145, 1990. 41. Kihara, M., and Low, P. A.: Impaired vasoreactivity to nitric oxide in experimental diabetic neuropathy. Exp. Neurol. 132:180, 1995.

Nerve Blood Flow and Microenvironment 42. Kihara, M., and Low, P. A.: Vasoreactivity to prostaglandins in rat peripheral nerve. J. Physiol. (Lond.) 484:463, 1995. 43. Kihara, M., McManis, P. G., Schmelzer, J. D., et al.: Experimental ischemic neuropathy: salvage with hyperbaric oxygenation. Ann. Neurol. 37:89, 1995. 44. Kihara, M., Nickander, K. K., and Low, P. A.: The effect of aging on endoneurial blood flow, hyperemic response and oxygen-free radicals in rat sciatic nerve. Brain Res. 562:1, 1991. 45. Kihara, M., Schmelzer, J. D., Kihara, Y., et al.: Efficacy of limb cooling on the salvage of peripheral nerve from ischemic fiber degeneration. Muscle Nerve 19:203, 1996. 46. Kihara, M., Weerasuriya, A., and Low, P. A.: Endoneurial blood flow in rat sciatic nerve during development. J. Physiol. (Lond.) 439:351, 1991. 47. Kihara, M., Zollman, P. J., Schmelzer, J. D., and Low, P. A.: The influence of dose of microspheres on nerve blood flow, electrophysiology, and fiber degeneration of rat peripheral nerve. Muscle Nerve 16:1383, 1993. 48. Kishi, Y., Nickander, K. K., Schmelzer, J. D., and Low, P. A.: Gene expression of antioxidant enzymes in experimental diabetic neuropathy. J. Peripher. Nerv. Syst. 5:3, 2000. 49. Korthals, J. K., and Wisniewski, H. M.: Peripheral nerve ischemia. Part 1. Experimental model. J. Neurol. Sci. 24:65, 1975. 50. Korthuis, R. J., Granger, D. N., Townsley, M. I., and Taylor, A. E.: The role of oxygen-derived free radicals in ischemiainduced increases in canine skeletal muscle vascular permeability. Circ. Res. 57:599, 1985. 51. Lagerlund, T. D., and Low, P. A.: A mathematical simulation of oxygen delivery in rat peripheral nerve. Microvasc. Res. 34:211, 1987. 52. Low, P. A., Lagerlund, T. D., and McManis, P. G.: Nerve blood flow and oxygen delivery in normal, diabetic, and ischemic neuropathy. Int. Rev. Neurobiol. 31:355, 1989. 53. Low, P. A., Nukada, H., Schmelzer, J. D., et al.: Endoneurial oxygen tension and radial topography in nerve edema. Brain Res. 341:147, 1985. 54. Low, P. A., Schmelzer, J. D., and Ward, K. K.: The effect of age on energy metabolism and resistance to ischaemic conduction failure in rat peripheral nerve. J. Physiol. (Lond.) 374:263, 1986. 55. Low, P. A., Schmelzer, J. D., Ward, K. K., and Yao, J. K.: Experimental chronic hypoxic neuropathy: relevance to diabetic neuropathy. Am. J. Physiol. 250:E94, 1986. 56. Low, P. A., and Tuck, R. R.: Effects of changes of blood pressure, respiratory acidosis and hypoxia on blood flow in the sciatic nerve of the rat. J. Physiol. (Lond.) 347:513, 1984. 57. Low, P. A., Ward, K., Schmelzer, J. D., and Brimijoin, S.: Ischemic conduction failure and energy metabolism in experimental diabetic neuropathy. Am. J. Physiol. 248:E457, 1985. 58. Low, P. A., Yao, J. K., Kishi, Y., et al.: Peripheral nerve energy metabolism in experimental diabetic neuropathy. Neurosci. Res. Commun. 21:49, 1997. 59. Lubawy, W. C., and Valentovic, M.: Streptozocin-induced diabetes decreases formation of prostacyclin from arachidonic acid in intact rat lungs. Biochem. Med. 28:290, 1982. 60. Lundborg, G.: Ischemic nerve injury: experimental studies on intraneural microvascular pathophysiology and nerve

61.

62. 63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73. 74. 75.

76.

77.

78.

79.

679

function in a limb subjected to temporary circulatory arrest. Scand. J. Plast. Reconstr. Surg. 6:3, 1970. Lundborg, G.: Structure and function of the intraneural microvessels as related to trauma, edema formation, and nerve function. J. Bone Joint Surg. Am. 57:938, 1975. McCord, J. M.: Oxygen-derived free radicals in postischemic tissue injury. N. Engl. J. Med. 312:159, 1985. McManis, P. G., and Low, P. A.: Factors affecting the relative viability of centrifascicular and subperineurial axons in acute peripheral nerve ischemia. Exp. Neurol. 99:84, 1988. McManis, P. G., Low, P. A., and Yao, J. K.: Relationship between nerve blood flow and intercapillary distance in peripheral nerve edema. Am. J. Physiol. 251:E92, 1986. Mitsui, Y., Okamoto, K., Martin, D. P., et al.: The expression of proinflammatory cytokine mRNA in the sciatic-tibial nerve of ischemia-reperfusion injury. Brain Res. 844:192, 1999. Mitsui, Y., Schmelzer, J. D., Zollman, P. J., et al.: Alphalipoic acid provides neuroprotection from ischemiareperfusion injury of peripheral nerve. J. Neurol. Sci. 163:11, 1999. Mitsui, Y., Schmelzer, J. D., Zollman, P. J., et al.: Hypothermic neuroprotection of peripheral nerve of rats from ischaemiareperfusion injury. Brain 122:161, 1999. Mitsui, Y., Schmelzer, J. D., Zollman, P. J., et al.: Hypothermic neuroprotection of peripheral nerve of rats from ischemia-reperfusion injury: intraischemic vs. reperfusion hypothermia. Brain Res. 827:63, 1999. Mohri, D., Satomi, F., Kondo, E., et al.: Change in gene expression in facial nerve nuclei and the effect of superoxide dismutase in a rat model of ischemic facial paralysis. Brain Res. 893:227, 2001. Moncada, S., and Vane, J. R.: Arachidonic acid metabolites and the interactions between platelets and blood-vessel walls. N. Engl. J. Med. 300:1142, 1979. Myers, R. R., Murakami, H., and Powell, H. C.: Reduced nerve blood flow in edematous neuropathies: a biomechanical mechanism. Microvasc. Res. 32:145, 1986. Nagamatsu, M., Schmelzer, J. D., Zollman, P. J., et al.: Ischemic reperfusion causes lipid peroxidation and fiber degeneration. Muscle Nerve 19:37, 1996. Neely, J. R., and Feuvray, D.: Metabolic products and myocardial ischemia. Am. J. Pathol. 102:282, 1981. Nukada, H.: Increased susceptibility to ischemic damage in streptozocin-diabetic nerve. Diabetes 35:1058, 1986. Nukada, H.: The susceptibility of rat diabetic nerve to ischemia: increased or decreased? J. Neurol. Sci. 119:162, 1993. Nukada, H., Anderson, G. M., and McMorran, P. D.: Reperfusion nerve injury: pathology due to reflow and prolonged ischaemia. J. Peripher. Nerv. Syst. 2:60, 1997. Nukada, H., and Dyck, P. J.: Microsphere embolization of nerve capillaries and fiber degeneration. Am. J. Pathol. 115:275, 1984. Nukada, H., and McMorran, P. D.: Perivascular demyelination and intramyelinic oedema in reperfusion nerve injury. J. Anat. 185:259, 1994. Nukada, H., McMorran, P. D., and Shimizu, J.: Acute inflammatory demyelination in reperfusion nerve injury. Ann. Neurol. 47:71, 2000.

680

Functional Compartments and Nerve-Fiber Environment

80. Omawari, N., Dewhurst, M., Vo, P., et al.: Deficient nitric oxide responsible for reduced nerve blood flow in diabetic rats: effects of L-NAME, L-arginine, sodium nitroprusside and evening primrose oil. Br. J. Pharmacol. 118:186, 1996. 81. Palinski, W., Torsellini, A., and Doni, L.: Influence of platelet activation of erythrocyte deformability. Thromb. Haemost. 49:84, 1983. 82. Parry, G. J., and Brown, M. J.: Arachidonate-induced experimental nerve infarction. J. Neurol. Sci. 50:123, 1981. 83. Parry, G. J., and Brown, M. J.: Selective fiber vulnerability in acute ischemic neuropathy. Ann. Neurol. 11:147, 1982. 84. Qi, W. N., Yan, Z. Q., Whang, P. G., et al.: Gene and protein expressions of nitric oxide synthases in ischemia-reperfused peripheral nerve of the rat. Am. J. Physiol. Cell Physiol. 281:C849, 2001. 85. Rechthand, E., Hervonen, A., Sato, S., and Rapoport, S. I.: Distribution of adrenergic innervation of blood vessels in peripheral nerve. Brain Res. 374:185, 1986. 86. Rechthand, E., Smith, Q. R., Latker, C. H., and Rapoport, S. I.: Altered blood-nerve barrier permeability to small molecules in experimental diabetes mellitus. J. Neuropathol. Exp. Neurol. 46:302, 1987. 87. Ritchie, J. M.: A note on the mechanism of resistance to anoxia and ischaemia in pathophysiological mammalian myelinated nerve. J. Neurol. Neurosurg. Psychiatry 48:274, 1985. 88. Rosen, P., Senger, W., Feuerstein, J., et al.: Influence of streptozotocin diabetes on myocardial lipids and prostaglandin release by the rat heart. Biochem. Med. 30:19, 1983. 89. Roth, D. M., Reibel, D. K., and Lefer, A. M.: Vascular responsiveness and eicosanoid production in diabetic rats. Diabetologia 24:372, 1983. 90. Rundquist, I., Smith, Q. R., Michel, M. E., et al.: Sciatic nerve blood flow measured by laser doppler flowmetry and [14C]iodoantipyrine. Am. J. Physiol. 248:H311, 1985. 91. Sasaki, H., and Low, P. A.: Extreme vasoreactivity of rat epineurial arterioles to vasopressin. Am. J. Physiol. 271:H1307, 1996. 92. Schlaepfer, W. W.: Calcium-induced degeneration of axoplasm in isolated segments of rat peripheral nerve. Brain Res. 69:203, 1974. 93. Schlaepfer, W. W.: Vesicular disruption of myelin simulated by exposure of nerve to calcium ionophore. Nature 265:734, 1977. 94. Schmelzer, J. D., Zochodne, D. W., and Low, P. A.: Ischemic and reperfusion injury of rat peripheral nerve. Proc. Natl. Acad. Sci. U. S. A. 86:1639, 1989. 95. Seid, J. M., Jones, P. B., and Russell, R. G.: The presence in normal plasma, serum and platelets of factors that stimulate the production of prostacyclin (PGI2) by cultured endothelial cells. Clin. Sci. 64:387, 1983. 96. Smith, K. J., Hall, S. M., and Schauf, C. L.: Vesicular demyelination induced by raised intracellular calcium. J. Neurol. Sci. 71:19, 1985.

97. Snopko, R., Guffy, T., Rafelson, M., and Hall, E.: Serum stimulation of prostacyclin synthesis in aortically, venously and microvascularly derived endothelial cells. Clin. Physiol. Biochem. 5:70, 1987. 98. Stewart, M. A., Passonneau, J. V., and Lowry, O. H.: Substrate changes in peripheral nerve during ischaemia and wallerian degeneration. J. Neurochem. 12:719, 1965. 99. Sundqvist, T., Oberg, P. A., and Rapoport, S. I.: Blood flow in rat sciatic nerve during hypotension. Exp. Neurol. 90:139, 1985. 100. Takeuchi, M., and Low, P. A.: Dynamic peripheral nerve metabolic and vascular responses to exsanguination. Am. J. Physiol. 253:E349, 1987. 101. Timperley, W. R., Boulton, A. J., Davies-Jones, G. A., et al.: Small vessel disease in progressive diabetic neuropathy associated with good metabolic control. J. Clin. Pathol. 38:1030, 1985. 102. Tuck, R. R., Schmelzer, J. D., and Low, P. A.: Endoneurial blood flow and oxygen tension in the sciatic nerves of rats with experimental diabetic neuropathy. Brain 107:935, 1984. 103. Wachtler, J., Mayer, C., Rucker, F., and Grafe, P.: Glucose availability alters ischaemia-induced changes in intracellular pH and calcium of isolated rat spinal roots. Brain Res. 725:30, 1996. 104. Ward, K. K., Low, P. A., Schmelzer, J. D., and Zochodne, D. W.: Prostacyclin and noradrenaline in peripheral nerve of chronic experimental diabetes in rats. Brain 112:197, 1989. 105. Wong, P. Y., and Cheung, W. Y.: Calmodulin stimulates human platelet phospholipase A2. Biochem. Biophys. Res. Commun. 90:473, 1979. 106. Ziboh, V. A., Maruta, H., Lord, J., et al.: Increased biosynthesis of thromboxane A2 by diabetic platelets. Eur. J. Clin. Invest. 9:223, 1979. 107. Zochodne, D. W., Cheng, C., and Sun, H.: Diabetes increases sciatic nerve susceptibility to endothelin-induced ischemia. Diabetes 45:627, 1996. 108. Zochodne, D. W., and Ho, L. T.: Influence of perivascular peptides on endoneurial blood flow and microvascular resistance in the sciatic nerve of the rat. J. Physiol. (Lond.) 444:615, 1991. 109. Zochodne, D. W., and Ho, L. T.: Normal blood flow but lower oxygen tension in diabetes of young rats: microenvironment and the influence of sympathectomy. Can. J. Physiol. Pharmacol. 70:651, 1992. 110. Zochodne, D. W., Ho, L. T., and Gross, P. M.: Acute endoneurial ischemia induced by epineurial endothelin in the rat sciatic nerve. Am. J. Physiol. 263:H1806, 1992. 111. Zochodne, D. W., and Low, P. A.: Adrenergic control of nerve blood flow. Exp. Neurol. 109:300, 1990. 112. Zollman, P. J., Awad, O., Schmelzer, J. D., and Low, P. A.: Effect of ischemia and reperfusion in vivo on energy metabolism of rat sciatic-tibial and caudal nerves. Exp. Neurol. 114:315, 1991.

31 Pathology of Peripheral Neuron Cell Bodies MICHAEL J. GROVES AND FRANCESCO SCARAVILLI

Pathologic Alterations in Peripheral Neuron Cell Bodies following Axotomy Chromatolysis Histologic and Ultrastructural Features Factors Affecting the Extent and Duration of Chromatolysis after Axotomy Chromatolysis Produced by Insults Other Than Axotomy Summary Intracytoplasmic Vacuolation Histologic and Ultrastructural Features of Axotomy-Induced Vacuolation Vacuolation Produced by Insults Other Than Axotomy Summary Biochemical Alterations Early Changes Cytoskeletal Proteins Neuropeptides Transcription Factors Enzymes Glycoconjugates Neurotrophic Factors and Receptors Ion Channels Summary Neuron Death Mechanisms of Neuron Death Identified Biochemical Events in Neuron Death Factors Affecting the Magnitude of Neuron Death Summary

Pathologic Alterations in the Cell Bodies of Peripheral Neurons in Age and Disease Aging General Observations Lipofuscin

Melanin Neurofibrillary Tangles Hereditary Neuropathies Hereditary Motor and Sensory Neuropathy (Charcot-Marie-Tooth Disease) Hereditary Sensory and Autonomic Neuropathy Types I through V Spinal Motor Atrophy Hirschsprung’s Disease Giant Axonal Neuropathy Neuroaxonal Dystrophy Defects in DNA Repair (Ataxia-Telangiectasia, Xeroderma Pigmentosum, Cockayne’s Syndrome) Disorders Caused by Trinucleotide Repeat Instability Friedreich’s Ataxia Spinocerebellar Ataxias X-Linked Spinal and Bulbar Muscular Atrophy (Kennedy’s Disease) Neuronal Intranuclear Hyaline Inclusion Disease Lysosomal Storage Disorders Gaucher’s Disease GM1 Gangliosidosis GM2 Gangliosidosis Neuronal Ceroid Lipofuscinosis (Batten Disease) Sphingomyelin Lipidosis (Niemann-Pick Disease) Mucopolysaccharidosis Mucolipidosis Mannosidosis Fabry’s Disease Type II Glycogenosis (Pompe’s Disease) and Other Glycogen Storage Diseases Chédiak-Higashi Syndrome

Leukodystrophies Refsum’s Disease Tangier Disease (Hereditary High-Density Lipoprotein Deficiency) Neurologic Mutants Dystonia Musculorum The Sprawling Mouse and the Mutilated Foot Rat The Wobbler Mouse Mouse Motor Neuron Degeneration Inherited Dog Neuropathies Neuropathies Produced by Systemic Metabolic Disorders Porphyria Diabetes Mellitus Amyloidosis Infective and Inflammatory Neuropathies Varicella Zoster and Herpes Zoster Infection Poliovirus Cytomegalovirus Infection Human Immunodeficiency Virus Other Viruses Producing Peripheral Neuropathy Diphtheria Leprosy Sensory Ganglionitis in Sjögren’s Syndrome Lymphoma Paraneoplastic Neuropathy Toxins Hexacarbons Metals Vincristine Pyridoxine Amiodarone Acrylamide Organophosphates

683

684

Pathology of the Peripheral Nervous System

The study of the pathologic alterations to neuronal perikarya following axotomy has been an area of research since Nissl reported alterations in the staining properties and appearance of motor neurons.263 Since the last edition of Peripheral Neuropathy, many advances in technique have enabled the molecular events underlying pathologic alterations, as well as those involved in the death of neurons, to be studied. We summarize these new data from animals and humans, as well as some of the historical background, by dividing this chapter into two parts. The first largely concerns the effects of axotomy upon peripheral neurons in experimental animals, with some data from human studies. The second describes pathologic alterations seen in defined human neuropathies, with some data from animal models.

Pathologic Alterations in Peripheral Neuron Cell Bodies following Axotomy CHROMATOLYSIS Histologic and Ultrastructural Features Chromatolysis, also known as the axon reaction, encompasses a sequence of morphologic changes in neuronal cell bodies following an insult, most often after axotomy. Most reports have involved the use of histologic DNA/ RNA stains such as toluidine blue and cresyl fast violet to visualize these changes because they show the Nissl substance very clearly4,50,154,168,263,362,363,394,494 (Fig. 31–1). The morphologic features of axotomy-induced chromatolysis in adult animals appear to be relatively consistent in motor, sensory, and sympathetic138,139,263,371 neuron populations, as well as in different species.44,50,291,357,363,371 The term chromatolysis refers to the dispersion of the clumps of Nissl substance (rough endoplasmic reticulum) from the center to the periphery of the perikaryon, resulting in loss of RNA staining in the central region (Fig. 31–1). In the case of axotomy-induced chromatolysis, this can progress to, and result in, an almost complete absence of staining in the center of the neuron. The first indication of chromatolysis is often the relocation of the nucleus from the center to the periphery of the perikaryon (nuclear eccentricity); this relocation may persist far longer than the loss of central basophilia.232 Indentations on the inward side of the eccentric nucleus340 may also be a feature (Fig. 31–1D). Ultrastructurally, the perikaryal changes involve a disruption of the cytoskeletal network of filaments and microtubules, which also become shorter, leading to a dispersion of rough endoplasmic reticulum to the periphery

and alteration of neuron shape.291 The process of chromatolysis does not itself appear to affect the speed of fast axonal transport,248 which is carried out by microtubules. The disappearance of the paraphyte processes of satellite cells, which normally interdigitate with the cytoplasm of the associated DRG neuron, has also been observed265 (Fig. 31–1D). The normal function of these processes and cells is not known; “glial” cells around large-diameter dorsal root ganglion (DRG) neurons proliferate after axotomy or nerve damage in the adult rat and human, often forming “perineuronal onion bulbs,”392 the function or significance of which is not clear. Another alteration often seen in conjunction with chromatolysis is swelling or shrinkage of the perikaryon, although care should be taken when evaluating volume changes in peripheral neurons because even very small variations in fixation technique will have quite profound consequences in relation to perikaryal volume. Furthermore, if proportions of large and small neurons are being measured, a preferential loss of small or large neurons could be misinterpreted as an overall decrease or increase in mean perikaryal volume, respectively. The occurrence of swelling or shrinkage seems to depend upon the type and duration of injury as well as the type of neuron and age of the animal. For example, following permanent axotomy of the sciatic nerve, chromatolytic adult rat DRG neurons undergo a decrease in perikaryal size initially affecting all sizes of neuron; however, after 6 weeks, only the larger neurons were reduced in size.367 This was maintained for up to 26 weeks. In contrast, transient axotomy of the sciatic nerve (produced by a crush injury) produced an initial decline followed by an increase in mean perikaryal area to levels above control values.366,494 There are many other reports of volume changes in peripheral neuron perikarya following peripheral axotomy,263 but in general it seems that permanent axotomy leads to a permanent decrease in volume in some or all of the injured neurons, whereas transient axotomy followed by axonal regeneration leads to an increased perikaryal volume that is linked to certain biochemical alterations (see below). Nuclear and nucleolar enlargement also seems to be a feature of axotomy-induced chromatolysis. Adult hamster facial neurons showed an increase in size that peaked at 2 days after transection of the facial nerve, preceding the peak chromatolytic response at 4 days.221 Nuclei and nucleoli of adult rat DRGs were also reported to show phasic increases in volume after crushing of the sciatic nerve, peaking at 3 to 4 and 8 to 11 days later, respectively.494 In the latter study significant volume changes were found in neuronal perikarya, nuclei, and nucleoli in the contralateral, uninjured DRGs: Such bilateral effects of a unilateral nerve injury have been reported to occur in rat trigeminal ganglia324 and elsewhere.242

FIGURE 31–1 A, Photomicrograph of normal adult rat lumbar DRG neurons stained with cresyl fast violet (Nissl stain). The large, circular nuclei are positioned in the center of the perikaryon, and show very little chromatin. They have either a single large, centrally positioned nucleolus in large neurons (arrow), or two to four small, peripherally located nucleoli in “small dark” neurons (arrowheads). The Nissl substance (rough endoplasmic reticulum) is evenly distributed in smaller or larger clumps throughout the cytoplasm. Bar: 10 ␮m. B, Chromatolytic features in an adult rat large DRG neuron stained with cresyl fast violet, 2 weeks after transection of the sciatic nerve. The nucleus and Nissl substance are displaced to the periphery, resulting in a lack of staining in the center of the neuron. Bar: 10 ␮m. C, Chromatolytic human spinal motor neuron from an autopsy of a patient with Guillain-Barré syndrome. Displacement of the nucleus and Nissl substance to the periphery and lack of central staining are evident. (Luxol fast blue/cresyl fast violet stain.) Bar: 10 ␮m. D, Electron micrograph of a chromatolytic adult rat DRG neuron 2 weeks after sciatic transection. The nucleus (N) is indented on the side facing the center of the neuron, a feature often seen in chromatolytic neurons. The rough endoplasmic reticulum (Nissl substance) is concentrated around the periphery with the mitochondria, Golgi apparatus, and other organelles concentrated in the center. This can clearly be seen at a higher magnification (E). Bar: 1 ␮m.

686

Pathology of the Peripheral Nervous System

Factors Affecting the Extent and Duration of Chromatolysis after Axotomy The age of the animal when axotomy occurs appears to be a crucial determinant in the appearance, speed of onset, duration, and outcome of chromatolysis, and whether it is seen at all. Peripheral nerve transection in newborn (⬍24 hours old) rats, hamsters, or guinea pigs does not produce any chromatolytic DRG neurons: affected neurons become smaller and more intensely stained,34 and many die.363 However, sciatic transection in 7- to 8-day-old kittens did lead to chromatolysis of some DRG neurons.5,368 Nuclear and nucleolar swelling220 was also absent in axotomized facial motor neurons of 15-day-old hamsters, but did occur at 25 days. The authors suggested that this difference may be due to the younger neurons being at the peak growth phase, and not capable of further enlargement. This sharp delineation with regard to age in the morphologic response to axotomy indicates a critical difference in neuronal metabolism between newborn and slightly older animals: Many rat DRG neurons do not attain their mature phenotype until the week after birth.22 Kerezoudi et al.232 reported a greater proportion of DRG neurons with eccentric nuclei at 6 months after peripheral axotomy performed on young rats than at 6 months after axotomy performed on aged rats. Chromatolysis appears to be more intense and resolves less rapidly in old animals than in young ones.233 The location of the injury or insult and the type of peripheral neuron are also major determinants in the time of onset and extent of chromatolytic changes. Axotomy of the centrally directed axons of DRG neurons (rhizotomy) does not produce any observable morphologic change in the perikarya of the affected neurons, regardless of the distance of the injury from the DRG.264 Likewise, proximal section of the central axons of the trigeminal and geniculate nerves and the inferior ganglia of the vagus and glossopharyngeal nerves of the monkey did not produce any significant chromatolysis in these ganglia, in contrast to transection of their peripheral axons.44 A lesion to the peripheral axons of DRG neurons induces a chromatolytic response in perikarya 24 to 48 hours later, depending upon the distance of the injury from the DRG.152,264 Small DRG neurons seem to respond to axotomy earlier than large neurons,363 a difference that is reflected in the increased expression of growth-associated protein-43 (GAP-43) that occurs in small DRG neurons before large ones.428 Adult rat thoracic spinal cord hemisection leads to chromatolytic changes in some lumbosacral motoneurons after 21 days,315 which is an example of transsynaptic effects because the motor neurons are unlikely to have been injured directly by the lesion. Following peripheral axotomy of motor axons, any changes start to be seen within 7 days.363,371 Unlike adult rat DRG neurons, few

(if any) chromatolytic adult rat spinal motor neurons are seen 1 month after sciatic transection (M. J. Groves and F. Scaravilli, personal observations). The duration of the chromatolytic response following axotomy depends largely on whether axons are permitted to regenerate or not. Crush injuries (axonotmesis) produce a transient axotomy that, as long as the perineurium is not breached, regenerates at the same rate in adult rats regardless of the width or duration of the crush.37 Resolution of the chromatolytic appearance occurs once contact with the peripheral targets been reestablished.367,444 Permanent axotomy, in which axonal regeneration is prevented, results in a longer duration of the chromatolytic response, which gradually diminishes.163,366

Chromatolysis Produced by Insults Other Than Axotomy Diseases and toxins that disrupt axonal transport or stimulate a regenerative response may produce the alterations characteristic of chromatolysis in peripheral neurons by producing a functional axotomy, or by inducing the same pattern of gene expression that follows axotomy. Chromatolytic motor neurons have been found in autopsies of patients with a wide range of acquired and inherited diseases, such as chronic inflammatory demyelinating polyneuropathy,316 acute motor axonal neuropathy/acute motor and sensory axonal neuropathy,161 WerdnigHoffmann disease,61 poliovirus,30 amyotrophic lateral sclerosis,281 macroglobulinemia,373 and radiation injury in dogs.25,354 Chromatolysis of enteric72 and autonomic neurons174 has been reported in cases of equine grass sickness. Chromatolytic neurons have been reported in sensory ganglia in certain inherited progressive sensory neuropathies in dogs,78 mice,94 and humans (see below). Toxins will also produce chromatolysis of sensory neurons, particularly those that inhibit axonal transport. Examples include intrathecal injection of local anesthetics329 and intraneural injection of cytotoxic lectins,500,516 systemic 2,5-hexanedione,432 3-acetylpyridine,20 and capsaicin.192 Sympathectomy using an immunotoxin will lead to a large proportion of sympathetic neuron cell bodies becoming chromatolytic in the adult rat.348 Drugs that block axonal transport provide interesting clues regarding the initiation of chromatolysis. Following transection of the hypoglossal nerve of adult cats and rats, a subepineurial injection of colchicine will delay the onset of chromatolysis and the increased uptake of 2-deoxyglucose (an early indicator of injury) in the adult rat hypoglossal nucleus.419 Furthermore, subepineurial injections of extracts of transected nerve into uninjured hypoglossal nerves stimulated 2-deoxyglucose uptake in the hypoglossal nucleus.420 Local application of vinblastine

Pathology of Peripheral Neuron Cell Bodies

(which inhibits microtubule polymerization and axonal transport) to the rat sciatic nerve produces some, but not all, of the changes associated with chromatolysis.121 Systemic administration to adult rats causes swelling of large DRG neurons that appears to be due to the accumulation of neurofilaments following the inhibition of anterograde transport.464

Summary When all of the disparate insults that can produce the appearance of chromatolysis in peripheral neurons are considered, the overall impression is that a stimulus produced in nerve at the site of injury is involved in triggering the chromatolytic response. The appearance of chromatolysis may therefore be governed by mechanisms that are switched on by the specific activation of genes involved in axonal regeneration that are quite distinct from those expressed during axonal elongation in development.266 This may also explain why neonatal responses to axotomy are qualitatively (and quantitatively) different from those in the adult. When the concomitant biochemical changes that occur after axotomy are considered, chromatolysis does appear to be a process associated with a regenerative, and not a degenerative, response to an insult by neurons.

INTRACYTOPLASMIC VACUOLATION Histologic and Ultrastructural Features of Axotomy-Induced Vacuolation There are occasional reports in the literature of “atypical” neurons in ipsilateral DRGs following axotomy or peripheral nerve damage that have single or multiple large intracytoplasmic vacuoles,3,44,50,164,232,290,324 up to 50 ␮m in diameter (Fig. 31–2). They usually have a low incidence but persist for long periods of time if axonal regeneration is prevented.50,164,232,290 This form of pathology is not confined to DRG neurons: Occasional grossly distorted, vacuolated sympathetic neurons have been observed in the adult rat 1 month after sciatic transection (M. J. Groves and F. Scaravilli, personal observations). Large intracytoplasmic vacuoles have been reported in adult mouse motor neurons 2 weeks after ventral root avulsion, an injury that leads to the death of a significant number of them.261 The morphology of vacuolated adult rat DRG neurons after peripheral axotomy is relatively constant: Neurons develop large spherical spaces that increase the perikaryal volume, often reducing the neuronal cytoplasm to a thin rim around the circumference and compressing the nucleus (see Fig. 31–2B). These vacuoles usually occur singly, leading Cavanaugh50 to call the affected neurons

687

“signet ring” cells, but neurons containing multiple vacuoles are not uncommon. Authors have assumed that they are not large invaginations (or “insudation”220) of the plasma membrane. Their intracellular location is confirmed by the absence of systemically administered horseradish peroxidase.164 Ultrastructurally, the vacuoles seen after axotomy appear to be identical to those seen after acrylamide intoxication in rats.220 They are membrane bound and have occasional villus-type protrusions of neuronal cytoplasm extending into the vacuole (see Fig. 31–3G below). They are filled with an aqueous phase containing a sparse, flocculent material (seen in electron micrographs) that is probably proteinaceous in nature and are probably derived from swollen Golgi apparatus or endoplasmic reticulum. Vacuolated rat DRG neurons induced by peripheral axotomy belong to the large DRG neuron population, based on the expression of large DRG neuron markers such as stage-specific embryonic antigen-4 (SSEA-4) and phosphorylated neurofilament (see Fig. 31–2F). Vacuolated neurons actively transport a retrograde tracer (the B subunit of cholera toxin) (see Fig. 31–2D); they express the TrkC neurotrophin-3 (NT-3) receptor (M. J. Groves and F. Scaravilli, personal observations), and their incidence can be reduced by administration of NT-3 (but not nerve growth factor [NGF] or brain-derived neurotrophic factor [BDNF]) to the proximal stump.164 Axotomy-induced vacuolation did not appear to lead to the immediate death of affected neurons, because the proportion of axotomized neurons expressing large neuron markers actually increased overall. We have found only one vacuolated neuron that could be stained using the in situ end-labeling technique for single-stranded DNA,498 and other authors using the terminal deoxynucleotidyl transferase–mediated dUTP nick end-labeling (TUNEL) method for DNA strand breaks137 similarly found only occasional ones that could be stained.290 None of these stained neurons showed condensed bodies of DNA characteristic of apoptosis, indicating that other mechanisms (such as excitotoxicity) may be involved in this DNA damage that could lead to the death of some vacuolated neurons. Vacuolated neurons can be seen in uninjured DRGs from rats at an incidence of around 0.005% to 0.01%, and their numbers seem to increase with age.164 Kerezoudi et al.232 also found that they increase in incidence with age in rabbit DRGs. Following successful regeneration of axons after crushing the sciatic nerve, the incidence of vacuolation returned to control levels by 2 to 3 months; after permanent transection their incidence remained relatively high (1% to 2%), gradually declining over 6 months to levels that were still higher than in controls.164 It would appear that, like chromatolytic neurons, the rate of disappearance or resolution of large

688

Pathology of the Peripheral Nervous System

FIGURE 31–2 A, Cresyl fast violet–stained adult rat L4 DRG neuron containing a single large vacuole, 1 month after sciatic nerve transection. Bar: 10 ␮m. B, Cresyl fast violet–stained adult rat L4 DRG neuron containing three vacuoles (same animal as A). Bar: 10 ␮m. C, Hematoxylin and eosin–stained adult rat sympathetic neuron from a sacral sympathetic ganglion, 1 month after sciatic transection, containing a single large vacuole. Bar: 10 ␮m. D, Vacuolated adult rat lumbar DRG neurons stained for peroxidase, 40 hours after injection of a cholera toxin–horseradish peroxidase conjugate into the sciatic nerve proximal stump transected 2 months previously. This showed that vacuolated neurons are axotomized neurons, possess axons projecting to the lesion site 2 months after axotomy, and are capable of active transport. Bar: 10 ␮m. Figure continued on opposite page

vacuoles in DRG neurons is governed by the rate of axonal regeneration.

Vacuolation Produced by Insults Other Than Axotomy Toxins and ionizing radiation also lead to vacuolated, or “cavitated,” sensory neurons. Acrylamide220 produces vacuolation of some rat DRG neurons at the same dose as that required to trigger the appearance of chromatolysis in DRG neurons within 1 week of administration (30 mg/kg391). Intraneural pronase injection,254 neonatal exposure to capsaicin,192 3-acetylpyridine administration,20 and acute radiation injury354 also produced DRG neuron vacuolation, indicating that vacuolation may be a

general reaction to insult. Most of these insults are thought to interfere with axonal transport, and (as in chromatolysis) this may be the key underlying event that triggers vacuolation. Vacuolation has also been reported in sympathetic neurons of young rats following NT-3 antiserum treatment272 and exposure to large doses of ethanol.209,210 The cytostatic drug doxorubicin (Adriamycin), when injected into the tongue, produces intracytoplasmic vacuolation in mouse hypoglossal motor neurons 14 days later,29 as does trimethyl tin and methylmercury intoxication in rat and gerbil spinal motor neurons.325,434 Vacuolated spinal motor neurons have also been seen in pigs with an inherited lower motor neuron disease,336 and in cases of porphyria49 and tetanus60 in humans. Vacuolated human DRG and sympathetic neurons

Pathology of Peripheral Neuron Cell Bodies

689

FIGURE 31–2 Continued E, Cryosections of adult rat lumbar DRG 1 month after sciatic transection, immunostained for growth-associated protein-43 (GAP-43). Vacuolated neurons seen after peripheral axotomy always demonstrate intense GAP-43 immunoreactivity. Bar: 10 ␮m. F, Cryosections of adult rat lumbar DRG 1 month after sciatic transection immunostained for the oligosaccharide conjugate stage-specific embryonic antigen-4 (SSEA-4), which is expressed by a subset of large sensory neurons. Vacuolated neurons express large sensory neuron markers such as SSEA-4 and phosphorylated neurofilament. Bar: 10 ␮m. G, Electron micrograph of an intracytoplasmic vacuole (v) within an adult rat L4 DRG neuron 2 months after transection of the ipsilateral sciatic nerve. Vacuoles are filled with an aqueous phase, with villus-like cytoplasmic processes occasionally projecting into it (arrowhead). Abundant lipofuscin (arrows) is usually present in vacuolated neurons. (Uranyl acetate and lead citrate stain.) Bar: 1 ␮m.

have also been described in diabetic neuropathy,92,156 although one of these reports is probably describing postmortem artifact in human DRGs rather than true vacuolation.156 Vacuolated human DRG neurons are also seen in herpes zoster infection.90 There are reports in the literature of “vacuolization” involving the formation of small fluid-filled vacuoles. This pathology seems to be associated with an excitotoxic insult to neurons, such as laser ablation of spiral ganglion neurons423 or in motor neurons following ventral root avulsion, and is further discussed in the section on Neuron Death.64 “Vacuolization” has also been reported in a transgenic mouse model of motor neuron degeneration (Mnd) involving the overexpression of superoxide dismutase 1,21 and vacuolation of motor neuron perikarya was found in one patient (of eight examined) with Creutzfeldt-Jakob disease.508

Summary Although grossly distorting intracytoplasmic vacuolation is seen as an incidental finding in DRGs of normal rats (particularly aged rats), its increased incidence is a significant pathologic alteration that has been described in sensory,

sympathetic, and motor neuron perikarya after trauma or exposure to certain neurotoxins or radiation, and in some neuropathies such as diabetes. It is not known how these large vacuoles relate to the formation of smaller cytoplasmic vacuoles/vesicles (“vacuolization”), or to the chromatolysis that is normally also present in other neurons. The incidence of large vacuoles in DRG neurons appears to be dependent upon neuron type, and may be a specific reaction to injury by certain subsets of peripheral neurons. The fact that it has been largely ignored is probably due to its relatively low incidence; however, knowledge of the mechanisms involved in vacuolation may be useful in understanding wider disease processes affecting neurons in which it is seen.

BIOCHEMICAL ALTERATIONS Numerous alterations to the biochemical content and processes of peripheral neurons have been described after an insult, particularly in sensory and motor neurons following peripheral axotomy. Many of these changes seem to reflect a switch to a regenerative state; broadly, neurotransmitters, neuropeptides, and enzymes associated with

690

Pathology of the Peripheral Nervous System

nerve conduction are downregulated while proteins associated with axonal growth and neuroprotection are upregulated. It is becoming apparent that the biochemical changes described so far in the literature are part of a complex interaction between neurons, axons, and satellite and Schwann cells, as well as peripheral and central targets, seemingly mediated at least in part by cytokines and neurotrophic factors.409 The restoration of the normal biochemical content of peripheral neurons appears to be governed by considerations similar to those involved in resolution of chromatolysis, but there are exceptions to this generalization.

Early Changes The first biochemical changes described after peripheral axotomy appear to consist of upregulation of proteins that probably have a neuroprotective role, and the relationship between ionic and transcriptional changes (e.g., activating transcription factor; see below) in this early phase is not yet established. One of the earliest changes described in peripheral neurons is the increased uptake of 2-deoxyglucose in rat hypoglossal motor neurons,420 and a massive influx of calcium ions into the axoplasm of the transected axon that may not reach the perikaryon.531,532 Twenty-four hours after axotomy of the rat hypoglossal nerve, injured motor neurons upregulate the calcium-buffering protein calbindin, which may protect the neuron from the deleterious effects of free calcium ions.247 Calbindin expression after axotomy has also been reported in rat sympathetic ganglion neurons.380 Another potentially neuroprotective molecule, heat shock protein 27 (HSP27), a chaperone protein thought to protect cellular macromolecules from damage, is upregulated in DRG neurons within 48 hours after sciatic transection at mid-thigh level.71,259 The stressactivated protein kinase p38 pathway is activated in spinal motor neurons after axotomy, which leads to the induction of HSP25,313 and HSP72 expression was observed in rat superior cervical ganglion neurons 24 hours after axotomy.198 Increased activity of ornithine decarboxylase, an enzyme involved in the synthesis of polyamines, which may protect DNA from damage, was detected in DRG neurons 13 hours after sciatic injury,493 and also in rat superior cervical ganglion after axotomy.139 The urokinase plasminogen activator receptor starts to be upregulated in adult mouse DRG neurons 8 hours after a crush injury to the sciatic nerve,414 and probably facilitates axonal regeneration by dissolving the extracellular matrix.

Cytoskeletal Proteins Changes in the expression of cytoskeletal components following axotomy reflect, as with many other macromol-

ecules, a switch of synthetic activity from that required for neurotransmission to that for axonal regeneration. Neurofilament messenger RNA (mRNA) is downregulated, while tubulin and actin mRNAs are upregulated, in rat DRG neurons beginning at around 5 days after sciatic axotomy.327,504 This probably underlies decreases in perikaryal volume seen following insult, because neurofilament content is a major determinant of the size of the axon and perikaryon.127,193 Dorsal rhizotomy also produces alterations in microtubule, actin, and neurofilament synthesis, although of a smaller magnitude than those produced by peripheral axotomy.157 Peripherin, an intermediate filament protein normally expressed in small DRG neurons, is upregulated in large DRG neurons (Fig. 31–3C and D) of adult rats505 by 7 days after sciatic crush and declines to control levels as axons regenerate. GAP43, a phosphoprotein involved in axonal elongation, is upregulated within 2 days in small-diameter DRG neurons after sciatic nerve transection, whereas L-type DRG neurons upregulate GAP-43 between 4 and 14 days later428 (Fig. 31–3B). Adult rat motor neurons demonstrate alterations similar to those in cytoskeletal components after axotomy that are dependent upon the distance from the site of peripheral axotomy to the cell body.115 ␣I- and ␤II-tubulin, actin, GAP-43,148,458,459 and peripherin470 are upregulated after axotomy (see Fig. 31–3), whereas neurofilament312,470 and microtubule-associated proteins443 are downregulated. Interestingly, peripheral axotomy in adult rats and cats will induce the dye coupling of spinal motor neurons normally only seen during embryonic development, without any alterations to connexin (a gap junction protein) mRNA expression.54,55 Avulsion of the sciatic nerve leads to the death of a proportion of the affected spinal motor neurons; this is preceded by an aberrant accumulation of phosphorylated neurofilaments in perikarya.283 Abnormal phosphorylation and glycosylation of neurofilaments has also been implicated in the loss of motor neurons in Werdnig-Hoffmann disease.61

Neuropeptides Neuropeptides are peptides or polypeptides synthesized by neurons that have neuromodulatory or neurotransmitter functions. With one or two exceptions, their synthesis decreases after axotomy as neuronal synthetic activity switches to proteins required for axonal regeneration. The neuropeptide galanin, thought to have an analgesic role, is not normally expressed at all but is upregulated in DRG neurons within 24 hours of sciatic transection in adult rats,485 and within 14 days in monkeys.525 The expression of substance P starts to decline in DRG neurons around 3 days after sciatic transection.217,485 Similarly, calcitonin gene–related peptide (CGRP) is mainly expressed by smalldiameter DRG neurons but shows a more delayed decrease

Pathology of Peripheral Neuron Cell Bodies

691

FIGURE 31–3 A, Cryosection of normal adult rat lumbar DRG immunostained for growth-associated protein-43 (GAP-43). Immunoreactivity is confined to small- to medium-diameter neurons. Bar: 10 ␮m. B, Cryosection of adult rat ipsilateral lumbar DRG, 2 weeks after transection of the sciatic nerve. Following peripheral axotomy, neuronal GAP-43 immunoreactivity increases in intensity and in the proportion of immunoreactive neurons, with many larger diameter neurons becoming immunoreactive (arrow). Bar: 10 ␮m. C, Paraffin section of normal adult rat lumbar DRG immunostained for the intermediate filament protein peripherin. As with GAP-43, immunoreactivity is confined to the smaller DRG neurons. Bar: 10 ␮m. D, A 4-␮m wax section of adult rat lumbar DRG immunostained for peripherin 1 month after sciatic transection. Larger DRG neurons are starting to become immunoreactive (arrows). Bar: 10 ␮m. Figure continued on following page

in expression than substance P following axotomy165 (see Fig. 31–3E to H). However, the number of DRG neurons that express CGRP seems to increase following dorsal rhizotomy, possibly as a result of its accumulation in neurons.206 CGRP and substance P downregulation in DRG neurons after axotomy is probably due to decreased target-derived NGF,481 as shown when endogenous NGF was removed with administration of a specific antiserum.409 However, guinea pigs appear not to show decreased CGRP immunoreactivity after axotomy,376 whereas monkeys do.525 Other DRG neuron neuropeptides show altered patterns of expression beginning around 1 week after peripheral nerve injury: somatostatin and cholecystokinin levels are decreased while vasoactive peptide and neuropeptide Y are

upregulated, mainly in small DRG neurons.225,530 These alterations are largely reversed following successful axon regeneration. Unlike in DRG neurons, galanin, substance P, vasoactive peptide, somatostatin, and ␣-CGRP (see Fig. 31–3) are upregulated, and cholecystokinin downregulated, in rat spinal motor neurons following sciatic transection.15,349,525 Noradrenergic neurons of the adult rat superior cervical ganglion increase their expression of substance P, neuropeptide Y, galanin, and vasoactive intestinal peptide mRNA after axotomy.530 This upregulation is induced by leukemia inhibitory factor (LIF) produced in the lesioned nerve435 that, with NGF upregulation in macrophages, is partly triggered by interleukin-1.43 The differences

692

Pathology of the Peripheral Nervous System

FIGURE 31–3 Continued E, Cryosection of L5 DRG immunostained for calcitonin gene–related peptide (CGRP), a neuropeptide expressed in 50% to 60% of normal adult rat lumbar DRG neurons, mainly of small diameter. Bar: 10 ␮m. F, Cryosection of ipsilateral adult rat lumbar DRG immunostained for CGRP, 2 weeks after sciatic transection. CGRP expression is downregulated in DRG neurons after peripheral axotomy. Bar: 10 ␮m. G, Cryosection of normal adult rat lumbar spinal cord immunostained for CGRP, showing the faint immunoreactivity normally present in spinal motor neurons. Bar: 10 ␮m. H, Cryosections of adult rat lumbar spinal cord, immunostained for CGRP 1 month after sciatic transection. CGRP immunoreactivity in spinal motor neurons is transiently increased in intensity following peripheral axotomy. Bar: 10 ␮m.

between neuronal populations in the axotomy-induced alterations to neuropeptide expression are probably due to differences in trophic factor dependence as well as differences in the function of the neuropeptides concerned.

Transcription Factors Most, if not all, pathologic alterations to neurons are ultimately due to altered gene expression.184 These alterations are starting to be understood using molecular biologic techniques such as DNA microarrays and reverse transcription–polymerase chain reaction, but it is already evident that different neuronal populations show differing patterns of transcription factor activation and inhibition that probably underlie differences in reaction to insult.

One of the first transcription factors known to be upregulated following axotomy is activating transcription factor 3, which binds to DNA as a homodimer or as a heterodimer with c-Jun; the time course of its induction depends upon the distance of the nerve cell body from the site of injury.471 N-Shc mRNA is predominantly expressed in hypoglossal motor neurons compared with Shc mRNA, but axotomy reverses this pattern of expression while SCK and Grb2 mRNA expression is unaffected.450 The LIM-type homeobox gene islet-1 plays an important role in differentiation and axonal elongation of sensory and motor neurons during development; however, following peripheral axotomy in adult rats, islet-1 mRNA is downregulated, with maximal depletion occurring at 3 to 7 days, but returns to control levels when regeneration is complete.194 Six days after sciatic

693

Pathology of Peripheral Neuron Cell Bodies

nerve transection, when neuronal death is occurring, increased immunoreactivity for the transcription factor c-Jun and the proapoptotic protein Bax takes place in spinal cord motor neurons and small-diameter DRG neurons of young rats.144 c-Jun expression and activation is elevated in neonatal and adult rat sensory and motor neurons after sciatic axotomy, as well as in some uninjured sensory neurons, and remains elevated until axonal regeneration is complete.183 c-Jun expression does not predict neuronal death or survival according to one study,422 although another study found that the axotomy-induced upregulation of c-Jun starts within 3 hours in neonatal motor neurons, becoming maximal between 1 and 10 days, and expression of a certain form of c-Jun was associated with apoptotic motor neurons.46 c-Fos expression is elevated in neonatal and adult motor neurons after axotomy, but not altered in sensory neurons.183

Enzymes Alterations to enzyme activity following axotomy seem to reflect the change to a regenerating state: enzymes involved with neurotransmission are downregulated while those thought to have roles in axonal elongation and neuroprotection are upregulated. Motor neurons synthesize acetylcholinesterase and choline acetyltransferase, the enzymes responsible for the degradation and synthesis of acetylcholine, respectively. These enzymes are downregulated in lumbar spinal motor neurons following peripheral axotomy65,116 and thoracic spinal cord hemisection315 in the adult rat. Fluoride-resistant acid phosphatase (FRAP) is probably the same enzyme as thiamine monophosphatase, and is exclusive to the nonpeptidergic small-diameter DRG neuron population.256 This neuron subpopulation also expresses an oligosaccharide glycoconjugate recognized by the LA4 antibody and Griffonia simplicifolia isolectin B4.416 FRAP activity in rat DRGs starts to decline around 3 days after sciatic transection but recovers to control levels by 21 days,455 and does not decline at all after rhizotomy. Conversely, carbonic anhydrase activity is found in large-diameter DRG neurons and—like expression of another large-diameter DRG neuron marker, SSEA-4165—is unaffected by sciatic transection.346 Some enzyme levels are increased. Nitric oxide synthase (NOS) produces nitric oxide, an important molecule in many pathologic processes because it reacts with superoxide radicals to produce highly damaging hydroxyl and nitronium radicals.85 NOS is upregulated in adult rat,480,526 but not guinea pig or monkey,376,526 DRG neurons within 48 hours of sciatic axotomy, indicating a potentially important species difference. The levels of urokinase and tissue plasminogen activator mRNA increase in adult mouse DRG neurons 3 days after crushing the sciatic nerve; the proteins are transported to the regenerating axon tip, where they are thought to assist regeneration by dissolving the extracellular matrix.414

NOS upregulation is also induced in adult rat motor neurons following ventral root avulsion, and this can be blocked by administration of BDNF.326 The induction of NOS activity in rat facial motor neurons after peripheral axotomy seems to be age dependent, since neonatal motor neurons show little evidence of any increase in activity, and upregulation is only seen in rats over 1 week old.277 A subpopulation of adult rat sympathetic neurons (about 2%) upregulates NOS by 1 week following axotomy,239 while no increase in the number of NOS-expressing enteric neurons could be detected following a crush injury to gut or treatment with colchicine.104 Expression of superoxide dismutases, the copper-, zinc-, and manganese-containing enzymes that catalyze the breakdown of toxic superoxide radicals, is also increased in rat facial motor neurons by peripheral axotomy,523 which represents an increase in neuroprotection. Tyrosine kinase activity, as indicated by increased phosphotyrosine immunoreactivity, has been reported to increase in the hypoglossal nucleus and the dorsal motor nucleus of the vagus nerve 48 hours after axotomy of the hypoglossal and vagus nerves, respectively514; this is in keeping with the upregulation in phosphatase/kinase activity reported after axotomy.266

Glycoconjugates The function of the (known) oligosaccharide glycoconjugates in DRG neurons is unknown, although they are likely to be involved in cell-cell recognition in the dorsal horn of the spinal cord. As with carbonic anhydrase, glycoconjugates associated with large DRG neurons such as SSEA-4165 appear to be unaffected by sciatic transection. In contrast, the expression of oligosaccharide glycoconjugates by the small-diameter DRG neuron population decreases after peripheral axotomy,347 particularly the glycoconjugate recognized by the LA4 antibody and G. simplicifolia isolectin B4,165,496 which may be a sensitive indicator of peripheral axon injury.

Neurotrophic Factors and Receptors Neurotrophic factors are proteins secreted by tissues, glia, and sometimes neurons themselves. Among other functions, they maintain the phenotype of the target neuron population and may even be required for their survival in adulthood. Neurotrophic factor expression in peripheral neurons before and after insult shows wide differences between, and also within, motor, sensory, and sympathetic neuron populations. The most researched of these molecules are the neurotrophins, basic polypeptides with a characteristic and highly conserved structure that are ligands for the p75 lowaffinity neurotrophin receptor (p75NTR) and one of the three membrane-bound tyrosine kinase (Trk) receptors.456 NGF, BDNF, and NT-3 appear to be the most widely expressed

694

Pathology of the Peripheral Nervous System

and important neurotrophins for the peripheral nervous system in development, maintenance of neuronal phenotype, and survival.1,461,481 Other proteins that can stimulate or protect peripheral neurons are usually referred to as neurotrophic factors and include cytokines and growth factors.456 BDNF mRNA is normally expressed in a small number of adult rat DRG neurons that also express the high-affinity nerve growth factor receptor TrkA.12 It is upregulated in response to NGF administration,295 inflammation,59 and sciatic nerve crush and transection406,463 in rat DRG neurons that express TrkB (the high-affinity BDNF receptor) and TrkC (the high-affinity NT-3 receptor), and is transported to the spinal cord, where it may modulate somatosensory pathways. This altered pattern of expression in DRGs was apparent 2 days after sciatic transection,295 and would appear at present to be specific to sensory neurons. NT-3 and NGF mRNA and protein are normally expressed in some adult rat DRG neurons, but transection of the spinal nerve seems to produce a slight decrease in this neuronal expression 1 week later, while satellite cells surrounding large-diameter DRG neurons upregulate NT-3 and NGF mRNA and protein.527 This upregulation in non-neuronal cells is probably responsible for the reported overall increase of NGF mRNA in rat DRGs following axotomy,406 and has also been reported for glial cell line–derived neurotrophic factor (GDNF).175 Most adult rat lumbar motor neurons express mRNA for BDNF and the related neurotrophin-4, and peripheral nerve transection causes an upregulation of BDNF mRNA between 2 and 7 days later.176,240 However, adult rat nodose ganglion neurons contained no mRNA for any of the main neurotrophins (NGF, BDNF, NT-3) before or after transection of the vagus nerve.258 The neuronal expression of neurotrophin receptors in neurons also shows differing responses to peripheral axotomy. Immunoreactivity for p75NTR declines in ipsilateral DRG neurons,299 and is upregulated in the contralateral neurons, during the first week after sciatic transection. This becomes maximal at around 3 weeks and remains depressed for the duration of observation (8 weeks528); a similar effect occurs in sympathetic neurons.360 The expression of p75NTR in adult rat motor neurons is induced by axotomy, but only in those allowed to regenerate.41 Immunoreactivity and mRNA for the high-affinity NGF receptor TrkA in DRG neurons starts to be downregulated 3 days after spinal nerve ligation411 or sciatic transection.24,299 TrkA, TrkC, and p75NTR expression also declines in nodose ganglion neurons after vagotomy.257 Caution must be exercised when assessing data on TrkB and TrkC expression because of the existence of alternatively spliced isoforms that lack the catalytic kinase domain. TrkB and TrkC mRNA and protein decrease in DRG neurons 1 week after sciatic transection in young and old rats.24 However, when mRNA was extracted from the whole DRG, increased levels of mRNA for TrkB and TrkC were found 1 week after axotomy, which was largely due to an increase

in the nonkinase forms110 and which also shows that non-neuronal expression of nonkinase Trk receptors is upregulated in DRGs after axotomy. Adult motor neurons normally express the neurotrophin receptors TrkB and TrkC, but, following peripheral axotomy, mRNA for TrkC is downregulated while that of TrkB is upregulated.116,176 The more severe injury produced by ventral root avulsion leads to the downregulation of both TrkB and TrkC.176 The role of other neurotrophic factors and cytokines in neuronal responses to injury is starting to be clarified: GDNF, LIF, glial growth factor, transforming growth factor (TGF), and ciliary neurotrophic factor (CNTF) all have trophic effects on various peripheral neuron populations.456 Basic fibroblast growth factor (bFGF) is mitogenic for a wide variety of cell types of mesodermal and neuroectodermal origin; bFGF protein and mRNA begin to be upregulated in rat DRG neurons 15 hours after axotomy, before declining after 1 week.218 This period corresponds to the peak incidence of mitotic activity in glial cells of axotomized rat ganglia.126,270 LIF and GDNF receptor mRNAs are upregulated in adult rat spinal motor neurons following a crush injury to their target muscle, sciatic nerve axotomy, or ventral root avulsion.176 Intriguingly, mRNA for a critical component of signal transduction for the LIF/GDNF receptor complex, known as gp130, was downregulated after ventral root avulsion.176 CNTF receptor mRNA was downregulated and the LIF receptor upregulated in rat facial motor neurons within 24 hours of axotomy of the facial nerve, returning to control levels after successful regeneration.170 The mRNA for the GDNF receptor is also upregulated in rat DRG neurons after sciatic nerve transection, but another component of the receptor complex, c-Ret, is not.224 The mRNA for epidermal growth factor receptor, which is also the receptor for TGF-␣, is upregulated in rat DRG neurons within 24 hours of peripheral nerve injury and remains so for at least 14 days, while TGF-␣ is upregulated in surrounding satellite cells.511

Ion Channels Pathologic alterations to voltage- and ligand-gated ion channels in sensory neurons could contribute to the ectopic spontaneous discharge associated with neuropathic pain. In other neuronal populations their expression could also significantly alter the normal function of affected neurons, or make them more or less vulnerable to excitotoxic neuron death. The expression of mRNAs for subunits of voltage-gated calcium channels are downregulated in adult rat DRG neurons 7 days after axotomy or chronic constriction injury of the sciatic nerve.236 Immunoreactivity for the ␣ and ␤1 and ␤2 subunits of voltage-gated sodium channels in human DRG neurons is also decreased after DRG avulsion.77 Immunoreactivity for SK1 and IK1 (calcium-activated potassium ion channels) was decreased in avulsed human DRGs

Pathology of Peripheral Neuron Cell Bodies

from 8 hours to 12 months after surgery, and was influenced by NGF levels in cultures of neonatal rat DRG neurons.33 The P2X3 receptor is a ligand-gated cation channel expressed by the nonpeptidergic subset of small-diameter DRG neurons169; it is activated by the binding of extracellular adenosine 5⬘-triphosphate and may therefore be involved in inflammatory pain. After transection of the tibial and common peroneal nerves or the infraorbital nerve, P2X3 mRNA is downregulated in associated rat DRG or trigeminal ganglion neurons, but upregulated in adjacent uninjured neurons, 3 days after injury.472 The number of neurons expressing P2X3 immunoreactivity in human DRGs has been reported to decrease following DRG avulsion.520 The expression of P2X3 channel and acid-sensing ion channel-3 (ASIC-3) immunoreactivity seems to be upregulated in enteric neurons in inflammatory bowel disease,519,521 and immunoreactivity for ASIC-1, -2 and -3 decreased in human DRG neurons following nerve avulsion.521 Motor neurons also downregulate ion channel expression after axotomy. mRNAs encoding sodium channel types I, II, and III are all decreased,208 as are glutamate-activated N-methyl-D-aspartate (NMDA) receptors.350 This last observation could conceivably alter the degree of excitotoxic neuron death in motor neurons, for example, after ventral root avulsion (see below).

Summary In conclusion, the biochemical changes described following peripheral axotomy largely mirror chromatolysis in the time course and in the factors that govern its duration and severity. Generally, axotomy-induced biochemical changes reflect the increased synthesis of macromolecules involved in regeneration and neuroprotection, and downregulation of molecules associated with neurotransmission. To the investigator, biochemical changes may be more sensitive indicators of injury than morphologic ones, and could give clues to the nature of the injury when the underlying controlling mechanisms are known.

NEURON DEATH One possible outcome of injury to peripheral neurons is the death of all or some of the affected neurons. Neuron death is now thought to encompass several distinct morphologic appearances that reflect different biochemical processes (e.g., apoptotic, autophagic, excitotoxic, cytoplasmic, necrotic) that could be regarded as representing a spectrum of neuron death mechanisms. Necrosis would be at one end, representing death as a direct consequence of very acute and severe insults, and apoptosis at the other, representing the most controlled and deliberate form of neuron destruction that is largely a secondary effect of an insult.64,345,353 The factors that determine which mechanism

695

will operate in a particular neuronal population after insult are not well understood; readers interested in a more detailed analysis of neuronal death are referred to some excellent monographs and reviews.64,241,345

Mechanisms of Neuron Death Axotomy-Induced Neuron Loss Death of neurons after axotomy has often been inferred from evidence of neuronophagia, or from neuronal counts that demonstrate a loss when compared to control groups of neurons.4,50,107,363,440 Neuronophagia—clumps of presumed cells with phagocytic activity in ganglia (nodules of Nageotte) or the anterior horn of the spinal cord—results from the death of a neuron at that site and the subsequent infiltration of microglial cells or macrophages (Fig. 31–4). Until recently, the methods used for estimating neuron number involved potentially biased profile-based counting methods.66 Without observations of degenerating neurons, neuronal loss inferred from neuron counting can produce misleading results, particularly if obtained from single dorsal root or sympathetic ganglia, because of the large variation within (between left and right) and between animals. Sensory neuron loss following unilateral axotomy has been reported in a variety of species, including humans,3,14,16,44,50,163,290,363,366,440 with most authors finding a selective loss of small-diameter DRG neurons.3,50,165,363,367,452 Motor neurons are vulnerable to axotomy-induced death in immature animals226 but appear to be more resistant in adults.261,357 Sympathetic neurons also seem to be lost in adult rats after axotomy.181 Apoptosis in the Peripheral Nervous System The discovery of apoptosis as a particular type of death with a defined physiologic role,234 and its rediscovery during the hunt for proto-oncogenes, stimulated interest in the field of “thanatology.” Apoptosis was thought to require the de novo synthesis of macromolecules278,510 as part of a controlled destruction mechanism that does not trigger an inflammatory response. It is now recognized that other forms of cell death involve macromolecular synthesis, and that apoptosis can occur in the absence of a cell nucleus,64 so many of the original assumptions are being challenged, particularly in the context of the nervous system. Apoptosis is recognized as a vital process in the regulation of cell number during the development of the peripheral nervous system.333 Indeed, descriptions of naturally occurring cell death during development frequently appear in the historical literature, starting with Vogt in 1842.63 The term often used to describe the morphology of this “natural” death was pyknosis, which describes the condensation and/or fragmentation of the neuron and its nucleus into darkly staining, irregularly shaped bodies. Pyknotic, degenerating, or necrotic neurons are mentioned in some studies on DRG neuron death following axotomy3,50,63,363 and in mouse

696

Pathology of the Peripheral Nervous System

FIGURE 31–4 A, Neuronophagia of motor neurons (arrows) from an autopsy of a patient with poliomyelitis. The spaces formerly occupied by motor neurons are filled with phagocytic cells removing cellular debris. (Hematoxylin and eosin stain.) Bar: 10 ␮m. B, Hematoxylin and eosin–stained human lumbar DRG from an autopsy of a patient with a paraneoplastic sensory neuropathy. Neuronophagia is often seen in aged or neuropathic human DRGs, where the clusters of phagocytic cells are known as nodules of Nageotte (arrows). Bar: 20 ␮m. C, Nodules of Nageotte are seldom seen in rat DRGs unless an insult is very acute and severe, such as exposure to neurotoxic agents. The nodules of Nageotte (arrows) indicated here were present in a lumbar DRG of an adult rat exposed to large doses of triethyl tin. (Toluidine blue–stained 1-␮m resin section.) Bar: 10 ␮m. (C courtesy of J. M. Jacobs.)

models of hereditary neuropathy386 from the time period before apoptosis was characterized in depth. Apoptosis was originally, and in many respects still is, a term that described a specific morphology. Nuclear DNA condenses into “caps” or nucleosomes (spherical bodies of highly condensed DNA that stains intensely with DNA stains; Fig. 31–5), the cell and nuclear volumes decrease, and the plasma membrane frequently shows “blebbing” (small protrusions of cytoplasm510). Ultrastructurally (Fig. 31–6), the cell cytoplasm becomes electron dense and filled with numerous osmiophilic bodies and fluid-filled vesicles that are probably the remains of organelles and membranes, although mitochondria are initially preserved. The nuclear membrane also becomes involuted and indistinct while the nuclear material condenses into spherical bodies. The ultimate fate of cells undergoing apoptosis is thought to involve them being phagocytosed, and the time taken for the whole process is very short, around 1 to 3 hours,64,509 leaving no obvious trace. The exposure of extracellular phosphatidylserine residues on lipids in the plasma membrane stimulates macrophages to engulf the cell,113 which in DRGs could lead to the formation of a nodule of Nageotte. The time between cultured sympathetic neurons having a normal morphology

and being completely unrecognizable as neurons following apoptosis induced by withdrawal of NGF was found to be 2 to 3 hours.102 This short time span means that a small observed incidence of neuronal apoptosis, maintained for a period of time, would lead to a significant loss of neurons. One of the first detailed descriptions of the morphology and ultrastructure of degenerating adult peripheral neurons (which today would be called apoptotic death) was of dying trigeminal ganglion neurons after transection of the infraorbital nerve.3 These authors also reported that degenerating neurons were not seen after 30 days posttrauma and that none were seen in the control, contralateral ganglia. They found that the number of degenerating neurons was small compared to the estimated neuronal loss, and speculated that this may be due to a rapid removal process. The features of the degenerating neurons described in this work are very similar to the features of apoptotic adult rat162,163,290 or mouse105 DRG neurons after peripheral axotomy. At least some of these apoptotic DRG neurons undergoing axotomy-induced apoptosis can be stained with the in situ end-labeling (ISEL)163 or TUNEL105,290 techniques (see Fig. 31–5) (see below). The numbers of apoptotic neurons seen in vivo in these reports were low, starting to

Pathology of Peripheral Neuron Cell Bodies

697

FIGURE 31–5 A, This photomicrograph shows an adult rat lumbar DRG neuron that is probably in the early stages of apoptosis 2 weeks after transection of the sciatic nerve. The nucleus and neuron have an irregular shape, and the nucleus (arrow) contains numerous small and darkly staining bodies. (Cresyl fast violet–stained paraffin section.) Bar: 10 ␮m. B and C, Cresyl fast violet–stained paraffin sections of ipsilateral lumbar DRG, 1 month after sciatic transection. The small clumps of DNA probably coalesce into larger bodies while the perikaryon starts to be broken down. At this stage the neurons undergoing apoptosis have an irregular shape, darkly stained cytoplasm, indistinct nuclear outline, and several spherical bodies of extremely condensed DNA. Bar: 10 ␮m. D, The end result of the apoptotic process is a body with no discernible structure (a “ghost cell”), apart from remains of condensed DNA (arrow). The size of the structure and the presence of an associated satellite cell (arrowhead) indicate that it is neuronal in origin. (Cresyl fast violet–stained paraffin section.) Bar: 10 ␮m.

appear from 1 day after adult rat sciatic nerve transection and ligation at mid-thigh level.290 The peak incidence occurred at 2 weeks; interestingly, apoptotic DRG neurons were still observed 6 months later.163,290 This axotomyinduced DRG neuron loss (relative to contralateral, uninjured DRGs) can be reduced by the administration of NT-3 and NGF,162,268,337,366 and NT-3 administration to the proximal stump also reduced the incidence of apoptotic DRG neurons.162 Following axotomy of a peripheral nerve, the initial indication that a DRG neuron is going to undergo apoptosis following chromatolytic changes seems to be the formation

of small “droplets” of condensed DNA in the nucleus, which also starts to develop an indistinct nuclear membrane (see Fig. 31–5). These DNA droplets seem to become bigger and the nucleus more irregularly shaped and indistinct, while the progressive loss of basophilia in the center of the perikaryon becomes more pronounced. The Nissl substance around the periphery of the perikaryon stains more intensely, and the shape of the neuron becomes irregular and shrunken. The neuron seems to become physically harder in paraffin-embedded material, which is probably due to the cross-linking of proteins and other constituents by transglutaminase. Finally, the nucleus becomes

698

Pathology of the Peripheral Nervous System

FIGURE 31–6 A, Apoptotic DRG neurons can be identified in resin-embedded material using similar criteria as for paraffin sections, although the remains of the nucleus and DNA are more difficult to identify. The condensed DNA (arrows) and darker staining cytoplasm distinguish them from nonapoptotic neurons. (Toluidine blue–stained 1-␮m resin section.) Bar: 10 ␮m. B, Ultrastructurally, the cytoplasm of apoptotic neurons is more electron dense than that of other nonapoptotic cells and neurons, and contains numerous vesicles, lipid bodies, and lamellar structures. The distorted nucleus (N) is located at the extreme periphery and contains condensed masses of DNA (arrows). The entire structure is surrounded by an associated satellite cell (S), but no paraphyte processes interdigitate with the neuron. (Uranyl acetate and lead citrate staining.) Bar: 1 ␮m. C, At a higher magnification it is possible to see that numerous intact mitochondria are present (arrows), thought to be a hallmark of apoptosis. The nuclear membrane is visible in places (arrowheads). Bar: 1 ␮m. D, Electron micrograph of a degenerating mouse spinal motor neuron, 8 days after injection of serum from a patient with an amyotrophic lateral sclerosis–like syndrome. The key features are dilatation of endoplasmic reticulum and Golgi apparatus (arrows), “blown” mitochondria (arrowheads), and a nucleus that appears near normal. This appearance suggests that apoptosis is not occurring and that some disruption to membrane integrity/ionic homeostasis has happened. Bar: 10 ␮m. (D courtesy of A. Pullen.)

unidentifiable, all basophilia has disappeared, and what remains is a seemingly empty sack containing the condensed remnants of neuronal DNA, a so-called ghost cell.102 Adult rat or mouse spinal motor neurons do not appear to die following transection of a peripheral nerve, whereas neonatal motor neurons seem to undergo a type of death that has similarities to apoptosis and necrosis. This involves DNA fragmentation261,277,372 that is mediated by NMDA receptor–linked ion channels.158,255,372 However, spinal

motor neuron death following ventral root avulsion in neonatal or adult mice does not appear to involve DNA fragmentation, and the process is more reminiscent of necrosis.261 Sciatic nerve avulsion causes adult rat spinal motor neurons to die via a process that has the morphologic features of apoptosis283 and that involves DNA strand breaks.267 These results suggest that there is a hierarchy of nerve injuries that determines the mechanism of motor neuron death in neonates and in adults, with ventral root

Pathology of Peripheral Neuron Cell Bodies

avulsion being the most severe and acute form of injury and peripheral nerve transection the least severe. This is probably due to significant trophic support being supplied by surrounding glia and neurons in adult spinal cord332 that is disrupted by root avulsion. Reperfusion following spinal cord ischemia has also been reported to produce apoptosis of rabbit spinal motor neurons, as determined by TUNEL staining and upregulation of caspase enzymes.51,179 Apoptosis in sympathetic neurons has been studied in vitro, where it is triggered by removal of NGF from the medium containing the cultured neurons within 24 to 48 hours.101,102,123,278 Its occurrence in vivo has not been reported. Necrosis True necrosis is thought to occur when cells are overwhelmed by an insult without having time to initiate a controlled destruction, leading to cessation of synthetic functions, loss of cell volume regulation, influx of water, and cell lysis.275 It is often reported as occurring in the central nervous system (CNS) following acute ischemic injury, with cell necrosis in the center of infarcts, which are often surrounded by a penumbra of damaged cells that have had time to initiate apoptosis. There are few, if any, images of neurons undergoing necrosis in vivo, presumably because it happens too rapidly. Reports of unequivocal necrosis (and not any other form of neuron death) in peripheral neurons are rare and should be treated with caution; acute injuries that produce this pattern of injury are not commonly encountered or studied in peripheral neurons. Frank necrosis of rabbit cranial and DRG sensory neurons has been described following acute intoxication with methylmercury,213 a toxin that is thought to inactivate membrane ion transport. In this report, neurons undergoing morphologic changes typical of necrosis can be seen: an electron-lucent cytoplasm, dilatation of Golgi apparatus and endoplasmic reticulum, and mitochondrial swelling and bursting while nuclear morphology remains almost normal.64 Intraneural injection of the cytotoxic lectins ricin, abrin, modeccin, or volkensin has been reported to produce necrotic rat sensory neurons and spinal motor neurons,499,500 although the presence of irregular, shrunken, dark nuclei in DRGs suggests that the process may not be true necrosis (if indeed true necrosis exists as a distinct entity). Herpes zoster infection can produce neuron death via a process that appears similar to necrosis, because a massive inflammatory cell infiltrate was involved, no structures resembling apoptotic (or pyknotic) bodies were reported despite a massive loss of neurons in the affected DRGs, and some degenerating neurons showed evidence of vacuolation.90 It is likely that the form of neuron death most closely resembling “true” necrosis would be produced by complement activation; this results in the formation of the membrane attack complex that forms a pore in the membrane to which the immunoglobulin/complement has bound.

699

The resultant influx of water and ions would, if enough pores were present in the membrane, cause neuron lysis. Excitotoxic Neuron Death Excitotoxic, or cytoplasmic, neuron death64,100,261 incorporates features of both apoptosis (nuclear DNA condensation, electron-dense cytoplasm) and “necrosis” (dilatation of endoplasmic reticulum, Golgi cisternae, and mitochondria). Its role has been investigated in research on some types of neurodegenerative disease, and experimentally in neurons using NMDA, kainic acid, and other analogues of glutamate. These agonists activate glutamate-activated ion channels, some types of which (e.g., the AMPA glutamate receptor subunit GluR2) are expressed in motor neurons. This may make them vulnerable to excitotoxic insults through the overstimulation of these receptors/ion channels and the subsequent disruption of intracellular ionic homeostasis.47 The activation and opening of ion channels would be a key event in excitotoxic mechanisms, allowing an influx of calcium and other ions that overwhelm the normal calcium buffering mechanisms, leading to the activation of calcium-activated degradative enzymes in the cytoplasm.57 It is therefore likely that this would be associated with a particular progression of morphologic changes, recognition of which could assist in the elucidation of pathologic mechanisms. Such a mechanism has been suggested to underlie some forms of motor neuron disease,478 although the process by which motor neurons die in motor neuron disease remains unclear to date.285,410 It is likely that the more severe the injury to motor neurons (ventral root avulsion ⬎ sciatic avulsion ⬎ ventral root transection ⬎ nerve transection), the more rapid and “cytoplasmic” the pattern of neuron death becomes. In sensory neurons, lidocaine has been reported to be neurotoxic via an excitotoxic mechanism involving membrane depolarization and calcium influx and release from intracellular stores151; it is interesting to compare this finding with reports of motor neuron pathology following intrathecal administration of local anesthetics.329 In a process very similar to excitotoxic cell death, laser ablation of spiral ganglion neurons423 produced features of both apoptosis (nuclear condensation) and necrosis (cell edema and “vacuolization”). An excitotoxic mechanism is suspected to underlie the sensory neurotoxicity of capsaicin and resiniferatoxin177,180,191: These molecules bind to and activate vanilloid receptor type 1, a ligand-gated divalent ion channel. Other Types of Neuron Death Reports of other forms of motor neuron death are beginning to appear in the literature. Motor neurons from transgenic mice overexpressing superoxide dismutase appeared to die via a mechanism that is not apoptosis, and instead involved mitochondrial swelling.21 Some axotomized facial motor neurons have recently been found to accumulate ␣-synuclein, and then to die slowly via a nonapoptotic process.305

700

Pathology of the Peripheral Nervous System

Identified Biochemical Events in Neuron Death One characteristic of apoptosis is activation of a calciumactivated endonuclease that digests DNA into 200–base pair fragments; these fragments form a characteristic “ladder” pattern in DNA gels.509 This degraded DNA can be visualized in situ using histochemical techniques that detect DNA strand breaks (TUNEL137) or single-stranded DNA produced by calcium-activated endonuclease (ISEL).498 However, both techniques stain both necrotic and apoptotic (i.e., dying) cells, and the point at which DNA degradation becomes detectable using ISEL or TUNEL is not known, nor is the precise point at which the “point of no return”

occurs. More recent developments have utilized the detection of specific components of the apoptotic pathway, such as activated caspases, transglutaminase, or poly(ADPribose) polymerase (Fig. 31–7). Caution should be exercised in using these techniques because it is still not clear how many other apoptotic pathways exist that do not involve these molecules. In our view, the actual type of cell/neuron death can only be characterized with any confidence by using morphologic criteria (particularly electron microscopy) on well-fixed (i.e., perfused) tissues. It is now recognized that apoptosis can be triggered by a variety of factors, among them DNA damage,284,506 loss of contact with targets and target-derived trophic factors,282 and defects in the ubiquitin degradative pathway.302 This

FIGURE 31–7 A and B, Photomicrographs of dying adult rat DRG neurons (arrows) stained with the ISEL technique, 1 month after sciatic transection. Histochemical staining methods for damaged DNA, such as TUNEL and ISEL, are not specific for apoptosis but can assist in identifying apoptotic neurons or cells in tissue sections by revealing degraded DNA. Bar: 10 ␮m. C and D, Immunohistochemical methods for apoptosis-specific proteins may be a more reliable way of identifying apoptotic cells and neurons. C, Photomicrograph of an adult rat DRG neuron, immunostained for activated caspase 3: the remains of the nucleus are just visible (arrows). Paraffin section of ipsilateral DRG, 2 weeks after sciatic nerve transection. Bar: 10 ␮m. D, The normally occurring apoptotic neuron death (arrows) in embryonic DRG can also be identified using activated caspase 3 immunohistochemistry. Paraffin section of embryonic day 14/15 rat embryo DRG. Bar: 10 ␮m.

Pathology of Peripheral Neuron Cell Bodies

initial “initiator” stage249 is thought to be followed by the activation of at least two signaling cascades (the effector stage), one involving the activation of the caspase family of proteases, which are thought to be specific for apoptosis, and at least one other caspase-independent pathway.302 Caspases 3, 6, and 7 are thought to be common to most apoptotic pathways, and immunohistochemistry for activated caspases has been used to detect apoptosis in situ in mouse models of motor neuron disease,166 in injuryinduced motor neuron changes,87,284,479 and in sensory neurons after NGF withdrawal.314 The caspase pathway can be activated by proteins released from damaged or permeabilized mitochondria.344 This might explain the combined features of apoptosis and necrosis seen in excitotoxic or “cytoplasmic” neuron death, in which mitochondria may be damaged by loss of ionic homeostasis. The final phase of apoptosis is the degradation phase, which involves the destruction of cellular organelles, cross-linking of cytoplasmic constituents, and DNA destruction. One enzyme involved at this stage is apoptosis-related transglutaminase type 2, thought to be involved in the cross-linking of cytoplasmic constituents; this has been detected in degenerating motor neurons in the Mnd transgenic mouse model of motor neuron disease.196 Apoptosis appears to be regulated by a balance between the expression of pro- and antiapoptotic signals. Differences in the expression of pro- (such as Bax) and antiapoptotic (such as Bcl-2 and Bcl-XL) proteins have been proposed as a basis of the selective vulnerability of sensory and motor neuron populations to axotomy-induced apoptosis.144 The role of p75NTR in this equilibrium is intriguing: It appears to be both pro- and antiapoptotic depending upon the type of neuron and the stage of development of the animal.17,36,45,56,124 Moreover, Bcl-2 (normally antiapoptotic) seems to promote, and be required for, the p75-mediated apoptotic death of cultured sensory neurons, while Bcl-XL (normally proapoptotic) seems to inhibit this.73 In axotomized facial motor neurons, the downregulation of Bcl-2 was greater the closer the site of axotomy was to the facial nucleus, which may be the explanation for proximal stump length–dependent survival in axotomized neurons.490,518 Excitotoxic neuron death mediated by NMDA receptors has been reported to involve the production of nitric oxide and oxygen radicals167,267 that can kill cells or neurons by reacting with, and inactivating, cellular macromolecules. Nitrotyrosine residues in proteins are indicative of the presence of nitric oxide and superoxide radicals and can be detected using immunohistochemistry. Such a mechanism has been proposed to underlie certain transgenic models of motor neuron disease,257 and nitrotyrosine residues have been detected in cases of motor neuron disease.382,474 A more important action of nitric oxide in chronic disease might be the inhibition of neuronal energy production.38 Disruption of calcium homeostasis certainly appears to produce many of the features of excitotoxic death in motor

701

neurons.2 Overexpression of the calcium-buffering protein parvalbumin in transgenic mice475 can protect motor neurons from this form of neuron death. Moreover, administration of certain calcium channel blockers can reduce facial motor neuron apoptosis and loss following transection of the facial nerve in neonatal rats.462 Theoretically, staining for soluble Ca2⫹ ions in histologic sections would identify apoptotic and excitotoxic neuron death in tissue sections, because calcium ions appear to be crucial in both proposed mechanisms. This is technically difficult because of postmortem dispersion of soluble ions. Complement activation by the Fc portions of bound antibodies leads to the formation of the membrane attack complex. This is essentially a pore or channel composed of various activated complement proteins that is inserted into the cell membrane of the target cell or neuron, leading to the lysis (as a result of loss of ionic homeostasis) of the cell/neuron being attacked. Consequently, in tissue sections this will appear similar to necrotic or excitotoxic death in morphology but in theory could be distinguished by the presence of activated complement proteins such as C3d or C9neo. C3d bound to the axolemma of motor axons has been reported in acute motor axonal neuropathy,171 which may result in the chromatolytic motor neurons reported in acute motor and sensory axonal neuropathy.161 Complement activation has also been described in the death of hypoglossal motor neurons produced by intraneural injection of ricin, in which complement C9 and the complement inhibitor clusterin immunoreactivities were found in neuronal perikarya, and C3d found in surrounding glial cells.465 Why complement should be activated by neuronal death is not clear. The stimulus for the macrophage and T-lymphocyte invasion of adult rat DRGs after transection of the sciatic nerve is also not clear202; the question of whether they are essentially protective or deleterious toward neurons is an interesting one. The mechanism of neuron death in paraneoplastic neuropathies might be expected to involve complement activation, but so far this has not been reported; neuron loss in anti-Hu–associated paraneoplastic neuropathy and encephalomyelitis did not apparently involve apoptosis or complement activation.26

Factors Affecting the Magnitude of Neuronal Death There seems little doubt that axotomy in the neonate produces a more rapid and much greater incidence of sensory and motor neuron death than in the adult.5,16,34,189,199,263,368,394,517 This may be due to the greater dependence of neonatal neurons upon target-derived trophic factors, or to a greater vulnerability of neurons with growing axons than of those without.495 There also appear to be qualitative differences in that condensed DNA caps, as well as the spherical nucleosomes seen in adults, were observed in neonatal apoptotic neurons16; these caps have not been reported or observed in adult rats.

702

Pathology of the Peripheral Nervous System

The distance between DRGs and the site of peripheral axotomy in adult rats directly affects the rate and magnitude of neuron death.413,518 Transection with ligation of the proximal and distal stumps of the adult rat spinal nerve will produce a loss of around 35% of L5 DRG neurons by 45 days later.484 An identical lesion at mid-thigh level produced a neuron loss of around 14% at 8 weeks and 37% at 32 weeks.452 The type of injury may also determine the extent of neuron loss: Transection, ligation, and “capping” of both distal and proximal stumps of the sciatic nerve at mid-thigh level appeared to produce an eventual neuron loss of around 35% in injured DRGs,290,452 whereas transection of the sciatic nerve and ligation of the proximal stump at mid-thigh level produced a loss of only 7% to 14%.163 This smaller loss could be due to neurotrophic factors produced in the distal stump being able to diffuse out of the cut end of the stump. The apparent differences in susceptibility to axotomyinduced neuron death between populations of adult peripheral neurons, particularly between motor and sensory/sympathetic neurons, may ultimately reflect biochemical differences between placode-derived and neural crest–derived neuron populations and their respective responses to cytokines and neurotrophic factors. However, it is also possible that susceptibility to death is purely related to neuron size. This question will not be answered until the mechanisms causing neuron death are established.

Summary This is a rapidly developing field, but it is already clear that the morphology and biochemistry of dying neurons can give valuable insights into the underlying mechanisms responsible. Mechanisms are likely to be different in adult and neonatal animals, but whether axotomy-induced neuron death is triggered by a loss of axoplasm, or something released from the injury site that is neurotoxic, or a lack of a survival factor transported from the periphery is still not known.

Pathologic Alterations in the Cell Bodies of Peripheral Neurons in Age and Disease AGING

Table 31–1. Incidence of Nodules of Nageotte with Age in Human Dorsal Root Ganglia Age

Sex

7 mo 13 yr 17 yr 17 yr 17 yr 35 yr 52 yr 53 yr 60 yr 69 yr 75 yr 80 yr 91 yr

F F F M M M F F F F M F F

Nageotte Nodules (%) 0 ⬍0.1 1.0 0.17 0.87 1.5 0.92 0.35 2.2 0.78 3.5 0.8 1.5

From Scaravilli, F.: Changes in neuronal structure and cell populations with ageing. In Thomas, P. K. (ed.): Peripheral Nerve Changes in the Elderly: New Issues in Neurosciences. Amsterdam, Wiley, p. 95, 1988, with permission.

of nodules have been observed in human DRGs after amputation of the upper arm.440 Age-related changes in autonomic ganglia are more obvious in humans than in other mammals, and in the superior cervical ganglion more than elsewhere.131 These agerelated changes mainly affect the cell processes. Enlarged argyrophilic neurites containing highly phosphorylated neurofilaments were observed in human paravertebral sympathetic neurons399; swollen axon terminals containing aggregates of filaments have also been reported.397,398,401 Sensory neurons in rats352 and humans108 have been reported to increase in number with age, whereas motor neurons appear to decrease in number but increase in size in rats.211 Enteric neurons have been reported to decline in number with age in experimental animals, based on neuron counts of gut neurons in young and old rats.76 Andrews9 observed that autonomic neurons increase in size postnatally, albeit not in a synchronized way: sympathetic neurons innervating the submandibular gland continue to grow into maturity from early postnatal periods, while those innervating the iris do not. Similarly, neuronal atrophy observed in aging is by no means a general phenomenon; rat DRG neurons appear to increase in size with age (M. J. Groves and F. Scaravilli, personal observations).

General Observations Nodules of Nageotte are frequently seen in DRGs from autopsies of aged patients without symptoms of neuropathy (Table 31–1). An age-related loss of sensory neurons may be behind the loss of unmyelinated axons often seen in sural nerves of aged patients, as well as the loss of deep tendon reflexes often observed in the elderly (M. J. Groves and F. Scaravilli, personal observations). Larger numbers

Lipofuscin Progressive accumulation of lipofuscin is the most obvious change in aging neurons of the peripheral, including the autonomic, nervous system. Lipofuscin is a brown pigment, fluorescent at all wavelengths, that is usually formed by the peroxidation and polymerization of unsaturated membrane lipids following attack by free radicals, and as such is often

703

Pathology of Peripheral Neuron Cell Bodies

thought of as a “wear and tear” pigment. In mice, the incidence of (all types of) neurons containing lipofuscin varies from none at birth to about 90% in 30-month-old animals. Mouse DRG and motor neurons containing lipofuscin increased in number in both types of cells from birth,320 and the increased amount of lipofuscin in motor neurons of aged rats has been suggested to be the reason for their increased size.211 Whereas in young mice this pigment is widespread in the cytoplasm of neurons, in older animals it clusters around the nucleus at one end of the neuron as well as near the axon hillock.320 In humans, the incidence of DRG and motor neurons containing lipofuscin (Table 31–2) had reached the values existing in the elderly by the end of the third decade.384 Increased lipofuscin deposits and lamellar cytoplasmic bodies have been reported in autonomic ganglia of 26- to 28-month-old Wistar rats.429 Together with a change in amount, age also brings about a modification of the physical and histochemical characteristics of lipofuscin.320 The relationship between lipofuscin and lysosomes has been studied in motor neurons of the mouse: Primary lysosome-like, autophagic vacuole–like, and mature pigment granules represent early, intermediate, and late stages, respectively.408 Aging is not the only factor involved in the process of accumulation of lipofuscin. Ubiquinated lipofuscin granules have been described in surviving motor neurons in cases of autosomal recessive spastic paraplegia.487 An increased amount of lipofuscin in DRG cells has been described in Tangier disease.396 Stress also seems to produce an increase in neuronal lipofuscin,231 whereas vitamin E298 and the drug centrophenoxine319 seem to have a preventive effect on its accumulation in CNS neurons. The role of vitamin E in protecting membrane lipids against attack by free radicals, and hence against lipofuscin accumulation, is illustrated by a new Table 31–2. Percentage of Human Dorsal Root Ganglion and Motor Neurons Containing Lipofuscin at Various Ages

Age

Sex

DRG Neurons with Lipofuscin (%)

7 mo 13 yr 17 yr 17 yr 27 yr 52 yr 69 yr 80 yr 91 yr

F F M M M F F F F

0 20 14 13 52 76 64 66 61

Motor Neurons with Lipofuscin (%) 0 ⬍0.1 ⬍3 24 64 81 65 50 63

From Scaravilli, F.: Changes in neuronal structure and cell populations with ageing. In Thomas, P. K. (ed.): Peripheral Nerve Changes in the Elderly: New Issues in Neurosciences. Amsterdam, Wiley, p. 95, 1988, with permission.

syndrome522 of vitamin E deficiency. This is produced by a missense mutation of the ␣-tocopherol transfer protein gene that, in addition to retinal atrophy and severe degeneration of the posterior columns of the dying-back type, also produces a massive accumulation of lipofuscin in DRG neurons.

Melanin A melanin-like pigment has been described in some neurons, particularly those of the human trigeminal and spinal ganglia.88 This appears early in life, and seems to precede lipofuscin and to increase in incidence with age. Scaravilli observed that staining methods for melanin gave positive results in the regions of the cytoplasm of sensory and autonomic ganglion (but not motor) cells occupied by lipofuscin.384 Ultrastructural features of melanin also are similar to those of lipofuscin19; moreover, the finding that neuromelanin in the pigmented nuclei of the brainstem resembles mature lipofuscin190 indicates that, in old sensory ganglia, lipofuscin bodies may also become melanized.265 The observation that the pigment is histochemically similar to melanin, and is present in sensory and autonomic nerve cells but not in motor neurons, may be explained by melanocytes, autonomic neurons, and sensory neurons having a common origin during development (the neural crest).

Neurofibrillary Tangles Neurofibrillary tangles (NFTs) are most often associated with the neuropathology of CNS neurons affected by Alzheimer’s disease and related neurodegenerative disorders, but have also been observed in motor neurons of patients with Alzheimer’s disease,378,404 with Parkinson dementia complex and amyotrophic lateral sclerosis from Guam,338,397,473 and with progressive supranuclear palsy (PSP) (Fig. 31–8A).227,235,486 NFTs were also observed in the nucleus of Onufrowicz of patients with PSP.390 Transgenic mice expressing the G272V mutant form of tau develop NFTs in their motor neurons.153 NFTs have been observed in dorsal root ganglion neurons in old Wistar rats,476 in the upper cervical ganglia of a 76-year-old man,229 and in cervical DRG neurons of patients with PSP.323

HEREDITARY NEUROPATHIES Hereditary Motor and Sensory Neuropathy (Charcot-Marie-Tooth Disease) The mutations that cause hereditary motor and sensory neuropathy (HMSN) types I and III (Déjérine-Sottas disease) are located within the genes encoding the myelin compaction proteins PMP22 and P0 expressed by Schwann cells,98 and

704

Pathology of the Peripheral Nervous System

FIGURE 31–8 A, Paraffin section of DRG from an autopsy of a patient with SCA7, immunostained for polyglutamine with the 1C2 antibody. Despite postmortem artifact (severe shrinkage of neuronal perikarya), the immunoreactive nuclear inclusion is clearly visible (arrow). A neighboring nodule of Nageotte (arrowhead) indicates previous loss of a neuron. Bar: 10 ␮m. (Courtesy of O. Ansorge.) B, Paraffin section of lumbar DRG from an autopsy of a patient with familial amyloidosis, stained with hematoxylin and eosin. The visible neurons (arrows) show features of chromatolysis (eccentric or displaced nuclei), and large deposits of amyloid are present throughout the DRG (*). Bar: 10 ␮m. C, Same case and DRG as B stained with Congo red and photographed under polarized light. Amyloid deposits surrounding a lone DRG neuron show up with a green birefringence in this preparation (arrows). Bar: 10 ␮m.

result in hypomyelinating/demyelinating neuropathies. There are few reports of pathologic alterations to neuronal perikarya in these diseases, although a loss of lumbar motor neurons has been reported in patients with HMSN I.449 Studies on the “trembler” mouse, a spontaneous mutant that has mutations in the P0 gene and is the most studied mouse model of HMSN I and III, failed to find any loss of sensory or motor neurons compared with normal mice.269 HMSN II appears to be a more heterogeneous genetic disease than HMSN I and III, with several loci linked to this neuropathy.99 Charcot-Marie-Tooth disease type X is an axonal neuropathy linked to mutations in the connexin 32 gene located on the X chromosome. The few published reports of autopsies of patients with HMSN II found degeneration and loss of DRG and motor neurons, as well as chromatolysis and atrophy of surviving motor neurons,23,524 and loss of large DRG and motor neurons in HMSN associated with deafness, mental retardation, and epilepsy.310

Hereditary Sensory and Autonomic Neuropathy Types I through V The genetic defects in hereditary sensory and autonomic neuropathy (HSAN) types I, II, and III have not been

identified at the time of this writing, and at least three loci have been found in genetic linkage studies of families with the HSAN I phenotype.99 An autopsy of a patient with a slowly progressive ataxic neuropathy of adult onset (which would probably be classified as HSAN I) revealed a slight decrease in the number of large DRG neurons and nodules of Nageotte in the cervical and lumbar DRGs, with no evidence of motor neuron loss.418 HSAN types IV and V (also known as congenital insensitivity to pain with anhidrosis) have been found to be the result of mutations in the TrkA NGF receptor.200,207,466 In transgenic TrkA knockout mice, the lack of functional TrkA leads to the death of non–TrkA-expressing smalldiameter DRG neurons and sympathetic neurons during embryonic development.415

Spinal Muscular Atrophy Spinal muscular atrophy (SMA) is a hereditary neurodegenerative disease affecting motor neurons mainly caused by homozygous deletions or mutations in the SMN gene, mapped to chromosome 5q12.2-q13.82,395 SMA spans a spectrum, from the severe SMA I (Werdnig-Hoffmann disease) to the mild SMA III (Kugelberg-Welander disease),

705

Pathology of Peripheral Neuron Cell Bodies

and pathologic findings from autopsies of 3- to 9-month-old infants with SMA I included chromatolysis and aberrant glycosylation/phosphorylation in spinal motor neurons.61 This appears to lead to the death of motor neurons by an apoptotic pathway, as revealed by TUNEL staining and ultrastructural examination.417

Hirschsprung’s Disease Also known as congenital megacolon, Hirschsprung’s disease is caused by a congenital absence of neural crest–derived enteric neurons in the myenteric plexus.32 The disease has a complex inheritance and is frequently associated with other congenital abnormalities. The abnormality can occur as a chromosome 2-to-13 translocation488 or as a deletion on chromosome 2.273,309

Giant Axonal Neuropathy Giant axonal neuropathy (GAN) is caused by mutations in the gigaxonin gene, localized to chromosome 16q24,122 and leads to accumulations of neurofilaments in axons that can be extremely large. The resultant neuropathy is often severe, leading to the death of affected individuals in the first or second decade. An autopsy of a 25-year-old man with GAN found accumulations of neurofilaments in spinal motor neurons that were slightly distended with central chromatolysis; coarsely clumped or a “dustlike” distribution of chromatin was also observed in motor neurons.343 Myenteric neurons in a rectal biopsy from an 8-year-old girl with GAN were reported to show neurofilament accumulation.133

Neuroaxonal Dystrophy Neuroaxonal dystrophy is characterized by swollen axons (smaller than those seen in GAN) and enlarged initial segments of axons or perikaryal projections in sensory ganglia, which may distort otherwise normal DRG neuron perikarya. Neuroaxonal dystrophy and GAN can be thought of as representing the two poles of a spectrum of neurofilamentrelated abnormalities, but whether this is due to mutations in the same genes is not known. Swellings similar to those seen in neuroaxonal dystrophy can develop as a function of aging and diabetes, and are composed of accumulations of phosphorylated neurofilaments; those seen in aged individuals are identical to those seen in diabetes.400

Defects in DNA Repair (Ataxia-Telangiectasia, Xeroderma Pigmentosum, Cockayne’s Syndrome) There are few reports of the peripheral neuron pathology of these diseases, which are usually caused by mutations in enzymes involved in DNA repair mechanisms. A progressive loss of DRG neurons was observed in an autopsy of a

49-year-old patient with xeroderma pigmentosum. The authors concluded that the disease is a primary neuronal degeneration that manifests first in the peripheral nervous system.369 NFTs have been observed in CNS neurons in isolated cases of Cockayne’s syndrome.446

Disorders Caused by Trinucleotide Repeat Instability Under this heading are included a number of genetically transmitted disorders previously classified in different groups. The first mutations to be described were X-linked SMA and bulbar muscular atrophy and fragile X syndrome, caused by expansions of trinucleotide repeats.252,482 Originally thought to be translated into polyamino acid tracts, the CGG repeat in fragile X syndrome was later identified as a nontranslated region of the FMR1 gene, showing that there are at least two types of trinucleotide repeat disorders251: ones in which the mutant triplet repeat is located in coding regions and ones in which it is located in noncoding regions of the gene.491 In the diseases in which the repeat is localized within the coding region of the gene involved, the repeat is translated into a stretch of polyglutamine, which may accumulate as neuronal intranuclear inclusions (NIIs). These may be seen with routine staining but can be specifically visualized using antibodies such as antiubiquitin (because the inclusions are always ubiquitinated), the 1C2 antibody that recognizes the expanded polyglutamine domain,469 or antibodies specific to each disorder. Indeed, the existence of NIIs in these diseases has only been recognized relatively recently, because it is difficult to see NIIs with normal histologic staining methods. These diseases can also be classified into four groups based on the nucleotide composition of the repeated triplet243: 1. Long cytosine-guanine-guanine (CGG) repeats in the two fragile X syndrome nucleotides, FRAXA and FRAXE 2. Long cytosine-thymidine-guanine (CTG) repeats in myotonic dystrophy 3. Long guanine-adenine-adenine (GAA) repeats in Friedreich’s ataxia 4. Short cytosine-adenine-guanine (CAG) repeat expansions implicated in at least 11 neurodegenerative disorders Transgenic mouse models of trinucleotide repeat disorders have been created, and NIIs have been observed in spinal motor, cranial sensory, and cranial motor neurons in some cases.18

Friedreich’s Ataxia Friedreich’s ataxia is the most prevalent inherited ataxia, with a frequency of 1 to 2/50,000 live births. It is an autosomal

706

Pathology of the Peripheral Nervous System

recessive disorder with onset in early childhood, although late-onset forms have been described. The disease is caused by the abnormal expansion of the GAA repeat in intron 1 of the FRDA gene on chromosome 9, which encodes frataxin, a protein that has been localized to mitochondria.35 This results in a severe and diffuse loss of large DRG neurons: A report of an autopsy found an increased number of Nageotte nodules with some of the remaining neurons showing chromatolytic changes.204 Motor neurons are not affected.

Spinocerebellar Ataxias Spinocerebellar ataxia (SCA) type 1 is an autosomal dominant disease with onset during midlife and is characterized by motor symptoms. It follows an expansion of the CAG repeat within a gene located in the short arm of chromosome 6; the gene product is called ataxin-1.335 A report of 11 autopsies of patients with SCA1 found an extensive loss of spinal motor neurons, with sparing of the nucleus of Onufrowicz; DRGs appeared to be normal.370 The mutation underlying SCA2 is a trinucleotide repeat (CAG) expansion within the coding sequence of the SCA2 gene, located in chromosome 12q24, which encodes a protein called ataxin-2. Autopsies of patients with SCA2 revealed a loss of motor neurons described either as mild334 or severe,112,339 and a moderate loss of DRG neurons.112 Intranuclear inclusions, described in the brain and brainstem, were not found in cranial ganglia.245 SCA3 (Machado-Joseph disease) is probably the most common dominantly inherited ataxia, and is caused by a CAG/polyglutamine repeat in a gene located to chromosome 14q that encodes a protein called ataxin-3.342 Autopsies of patients with SCA3 have revealed a loss of spinal motor507 and DRG neurons,74 loss of the latter being associated with nodules of Nageotte. NIIs in DRG and paravertebral sympathetic and celiac ganglion neurons have recently been reported in four of four autopsies of patients with the disease.515 These inclusions are similar to those seen in the CNS: dense, spherical bodies inside the nucleus as well as a diffuse, granular material present in the cytoplasm also visualized with the IC2 antibody. The genetic defect in SCA6 is located to chromosome 19p13; the CAG repeat is found within the gene for the human a1A voltage-dependent calcium channel unit that is expressed in Purkinje cells.529 This produces a purely cerebellar form of ataxia that progresses over 20 to 30 years. A loss of motor neurons in the motor nuclei of the medulla and spinal cord has been reported from autopsies of patients with SCA7,280 which is caused by a CAG repeat mapped to chromosome 3p12-13.83 The neuropathology of SCA7 is also characterized by the presence of NIIs195 that have been seen also in DRG neurons (Fig. 31–9A).

X-Linked Spinal and Bulbar Muscular Atrophy (Kennedy’s Disease) Kennedy’s disease is a motor neuronopathy of late onset and was the first disorder related to a CAG repeat expansion.252 The disease was mapped on the region of chromosome X where the gene for androgen receptor is located. Pathologic changes include loss of lower motor neurons in brainstem and spinal cord and the presence of NIIs, which are best visualized by antibodies to ubiquitin or to the abnormal protein. No obvious signs of neuronal degeneration or loss could be found in lumbar DRGs from autopsies of two patients with the disease, although a distal axonopathy was found.424

Neuronal Intranuclear Hyaline Inclusion Disease Neuronal intranuclear hyaline inclusion disease is a rare neurodegenerative disorder that has a heterogeneous clinical presentation and pathologic features and usually affects young patients, with death occurring by the third decade. The mode of inheritance is not clear, with both sporadic and familial cases having been described as well as cases in adults. In addition to nerve cell loss, the salient pathologic finding is the presence of NIIs. Because these inclusions share morphologic similarities with those observed in polyglutamine disorders, the disease is presented in this section. These inclusions are ubiquitin positive, react to a varying degree with the antibodies used to detect inclusions in the group presented above, and are ubiquitous in the nervous system. They are composed of aggregates of 8- to 9-nm diameter straight filaments that are autofluorescent, and have been reported in spinal motor,136,237,296,427,447 DRG,136,237,296,436,447 and autonomic436 neurons.

LYSOSOMAL STORAGE DISORDERS These disorders are caused by defects in single lysosomal enzymes. The genetic mutations that underlie a majority of them have been discovered, and the diseases themselves have been classified by detailed clinical and pathologic descriptions.442 This section summarizes any pathologic changes to peripheral neurons that have been described in these disorders.

Gaucher’s Disease Gaucher’s disease is an autosomal recessive disease usually caused by a deficient activity of glucosylceramidase, the gene for which is located in band q21 of chromosome 1.147 However, a deficiency of sphingolipid activator protein C also causes a variant of Gaucher’s disease. These mutations result in massive accumulation of glucocerebroside in

707

Pathology of Peripheral Neuron Cell Bodies

FIGURE 31–9 A, Paraffin section of postmortem lumbar spinal cord from a patient with progressive supranuclear palsy (PSP). Accumulations of tau protein and neurofibrillary tangles have been described in motor neurons of patients with PSP. These tau-immunoreactive deposits may coalesce into large clumps (arrow), displacing the nucleus of the motor neuron to the periphery. Bar: 10 ␮m. B, Luxol fast blue/cresyl fast violet–stained paraffin section of the ventral horn of lumbar spinal cord from an autopsy of a patient with chronic lymphocytic leukemia who developed a GuillainBarré–like syndrome. This motor neuron shows features of chromatolysis, with displacement of the nucleus (arrow) and loss of central basophilia. Bar: 10 ␮m. C, This chromatolytic motor neuron from the same case as B (nucleus indicated by arrow) also contains intracytoplasmic vacuoles (arrowheads). Bar: 10 ␮m.

reticuloendothelial cells, and only the type II and III forms of the disease are neuronopathic. An autopsy of an infant with severe type II Gaucher’s disease found an extensive loss of spinal motor neurons that the authors speculated may involve an apoptotic pathway.119 The sensory and motor neurons of transgenic mice with a targeted disruption of the glucocerebrosidase gene contained tubular Gaucher-type inclusions at the ultrastructural level.501

and have similar neurologic abnormalities and disease progression, showing neuronal accumulation of periodic acid–Schiff reagent (PAS)-positive material from 5 weeks of age in neurons of the brain and spinal cord.173,286 Suramin is an experimental chemotherapeutic agent that seems to produce an experimental GM1 gangliosidosis in cultured DRG neurons, which develop large multilamellar inclusion bodies composed of GM1 ganglioside.142

GM1 Gangliosidosis

GM2 Gangliosidosis

GM1 gangliosidosis is a progressive neurologic disorder caused by a deficiency of lysosomal acid ␤-galactosidase, the gene for which has been located to chromosome 3p21.33.448 There are two childhood forms (infantile type 1 and late infantile/juvenile type 2) and one adult form (type 3). In type 1, accumulation of ganglioside is seen in DRG neurons from 24 week-old human embryos28; subsequently the peripheral neuron pathology tends to be overshadowed by the CNS involvement. Neuronal involvement in the adult-onset form of the disease (type 3) is largely confined to the basal ganglia.441 Transgenic mice that have had the ␤-galactosidase gene deleted have been created

Formerly known as Tay-Sachs disease (a name retained for the infantile form of the disease), GM2 gangliosidosis is the result of a decreased activity of either ␤-hexosaminidase or GM2 activator protein resulting from deletion or mutation of the genes encoding the subunits of the enzyme or activator protein.68 The gene for the ␤-hexosaminidase a subunit is on chromosome 15; the genes for the ␤ subunit and GM2 activator protein are on chromosome 5. Mutations lead to a massive accumulation of GM2 ganglioside and other glycolipids in neuronal lysosomes, which appear as multilamellar structures known as membranous cytoplasmic bodies (MCBs). These have parallel or concentric

708

Pathology of the Peripheral Nervous System

lamellae when examined with the electron microscope,457 and may trigger apoptosis in the neurons affected.203 In the peripheral nervous system, accumulation of gangliosides has been found in enteric neurons in the rectal biopsies from nine adult patients ages 26 to 39 years.502 Transgenic mice with targeted mutations of ␤-hexosaminidase type B (Sandhoff variant) demonstrated extensive MCBs in neurons of the dorsal root and trigeminal ganglia, spinal cord, and myenteric plexus.381

in neurons of the brain and gastrointestinal (GI) tract. However, the stored material of Kufs’ disease is associated with normal lipofuscin, and to differentiate between the two is difficult in routine preparations. According to Lake,253 diagnostic rectal biopsy is not advisable because of the considerable lipofuscin deposition that takes place normally in GI neurons of the adult. Ultrastructurally, inclusions are not uniform and include curvilinear and fingerprint bodies and rectilinear profiles.

Neuronal Ceroid Lipofuscinosis (Batten Disease)

Sphingomyelin Lipidosis (Niemann-Pick Disease)

The term neuronal ceroid lipofuscinosis is incorrect because the stored material is neither ceroid nor lipofuscin but a protein, and also because storage takes place in all tissues and not just the nervous system. The various forms are related to lysosomal enzymes classified as CLN1 through CLN8, with at least six gene loci being implicated in their inheritance.134 In the infantile form (CLN1), the gene for the lysosomal enzyme palmitoyl protein thioesterase, located on chromosome 1p32,215 is mutated. This leads to the storage of PAS-positive material (in paraffin and frozen sections) in enteric neurons. Ultrastructural features are common to cells in all organs and consist of membrane-bound, granular, osmiophilic deposits forming globules 0.5 ␮m in diameter, or arranged in aggregates 3 ␮m in diameter. The late infantile form (CLN2) is caused by a mutation in a gene located to chromosome 11p15 encoding a pepstatin-insensitive lysosomal peptidase301 and is characterized by less severe changes. The enteric neurons of Auerbach’s plexus contain abnormal material arranged as curvilinear bodies that form the characteristic inclusions of this form of neuronal ceroid lipofuscinosis,97 and which are only weakly PAS positive. The normal function of the mutated genes responsible for CLN3, -4, -5, and -8 are unknown, although the genetic loci have largely been established301: juvenile Batten disease (CLN3) is caused by a mutation in a gene located at chromosome 16p12, and late infantile Batten disease (CLN5) is due to a mutation in a gene mapped to chromosome 13p22. It is worth noting that mutations in the CLN8 gene in mice produce the Mnd mouse.134 In this form of the disease, enteric neurons containing stored material similar to that seen in CNS neurons can be observed; this material can be stained with the PAS, Sudan black, and Luxol fast blue techniques, and shows strong autofluorescence. Ultrastructurally, these bodies are different from curvilinear bodies in having a fingerprint-like appearance; neurons may also contain MCBs. The pathologic features of CLN4 and -5 are identical to those of the juvenile form. The adult form of Batten disease (CLN4; Kufs’ disease) is also characterized by an excess of lipofuscin-like material

This group of autosomal recessive disorders is characterized by cholesterol accumulation in various cell types, the most common of which is Niemann-Pick disease type A/B resulting from deficient acid sphingomyelinase activity. The gene for this enzyme has been located to chromosome 11p15.1-p15.4.79 One mutation for Niemann-Pick disease type C1 has been located to chromosome 7p13,84 and another for the major form of types C and D to the centromeric region of chromosome 18q.159 In 1958, Croker and Farber75 proposed a subdivision into four groups (A through D). Subsequently two types, type 1 (groups A and B) and type 2 (C and D), were defined on the basis of the absence or presence of sphingomyelinase, respectively.106 Axonal neuropathy develops in advanced stages of the disease, and is similar to the pathologic alterations to axons in the CNS. “Ballooned” enteric neurons are found in the type 1 diseases, and ultrastructurally neuronal inclusions appear as loosely packed lipid lamellae enclosed within membranebound vacuoles measuring 1 to 2 ␮m in diameter. A transgenic mouse model of type 1 Niemann-Pick disease, produced by deletion of exon 3 of the sphingomyelinase gene, has histopathologic features similar to those of the human disease.250 In the group 2 diseases, ballooned neurons were reported in the anterior horn and in neurons of the myenteric and submucosal plexuses.172 The stored material in neurons from patients with group 2 diseases is weakly sudanophilic and PAS positive in frozen sections. The positive staining with ferric hematoxylin suggests the presence of phospholipids, and electron microscopy shows that the membrane-bound cytoplasmic bodies are polymorphous, are up to 3 ␮ in diameter, and contain loosely packed lamellae that are concentric in places. Mouse transgenic models have been produced by targeted deletion of the genes NPC1 or NPC2.331

Mucopolysaccharidosis The mucopolysaccharidoses form a group of disorders the majority of which are inherited as an autosomal recessive trait. An exception is represented by mucopolysaccharidosis

709

Pathology of Peripheral Neuron Cell Bodies

(MPS) type II (Hunter’s disease), which is an X-linked recessive disorder, although a few cases in females are recorded.308,503 These disorders are characterized by the accumulation of water-soluble mucopolysaccharides that are not retained in fixed material and are due to defects in the various mucopolysaccharide synthesis and storage genes. They have been classified by Lake253 on the basis of criteria that include eponyms, specific enzyme defects, identification of urinary mucopolysaccharides, genetic studies, and genetic/phenotype correlations. In cases of MPS IV, storage is seen in the cerebral hemispheres, including deep gray nuclei,145,244 but not in spinal motor neurons or in ganglia of the autonomic system. Neuronal storage of lipid (or ceroid) and/or membranous cytoplasmic bodies (“zebra bodies”) is present in motor neurons and autonomic ganglia in MPS IH, IH/S, II, III, and VII.178 However, an autopsy of a patient with MPS I found the spinal motor neurons to contain only lipid/ceroid and no lamellated inclusions.216

Mucolipidosis The mucolipidoses are classified into four categories; mucolipidosis (ML) I is now included under the neuraminidase deficiency disorders. ML II and III have the same enzyme defect (N-acetylglucosamine-1-phosphotransferase), the gene for which has been located to chromosome 16p.359 ML IV is caused by mutations in the MCOLN1 gene that encodes for mucolipin, a member of the transient receptor potential gene family,7 located on chromosome 19p.421 Neuronal storage of lipid-like material in spinal ganglia has been reported from an autopsy of a patient with ML II, which is also known as I-cell disease.318 However, autopsies of four patients with ML II found no inclusions in DRG or parasympathetic neurons, but did find basophilic stored material in motor neurons that was seen to be identical to zebra bodies under the electron microscope.279 Neuronal storage of PAS- and Sudan black–positive material in the central454 and peripheral150 nervous systems has been reported in cases of ML IV.

Mannosidosis Mannosidosis is divided into two categories according to whether it is due to a deficiency of either lysosomal ␣- or ␤-mannosidase. These enzymes degrade asparagine-linked carbohydrate cores of glycoproteins, and deficiencies in their activity results in intralysosomal accumulation of mannoserich oligosaccharides. The gene encoding ␣-mannosidase has been mapped to chromosome 19cenq12,322 and that for ␤-mannosidase to chromosome 4q22-25.6 An autopsy of a 3.5-year-old patient with mannosidosis found severely ballooned neurons (distended with large numbers of small vacuoles or vesicles) in the paravertebral

sympathetic, trigeminal, and dorsal root ganglia with occasional evidence of neuronal loss; spinal motor neurons appeared less severely affected. Ultrastructurally, the vacuoles/vesicles were often interdigitated or invaginated with each other, sometimes fusing or coalescing together and often contained small numbers of fine fibrils in stacks as well as lipid globules.439 Certain plant alkaloids, such as swainsonine (from Ipomoea carnea, Swainsonea, Astragalus, and Oxytropis species), inhibit mannosidase and produce similar effects upon cells and neurons in intoxicated animals.86,205 Intoxication produced characteristic accumulation of small vacuoles or vesicles in certain neuron populations in the CNS, and in lumbar DRG neurons with no evidence of loss.205 Only neurons unprotected by a blood-brain barrier were affected.

Fabry’s Disease Fabry’s disease is an X-linked genetic disease that results in a deficit of ␣-galactosidase activity238; the mutation responsible has been localized to chromosome Xq22.1.109 It leads to the accumulation of stored glycolipid, comprising globotriaosylceramide and galabiosylceramide,201 in DRG and sympathetic ganglion neurons, and in enteric neurons.412 Small-diameter, nociceptive neurons appear to be particularly affected,132,328 and a moderate loss of DRG neurons and a slight loss of sympathetic neurons was described in a report of an autopsy of a 50-year-old heterozygous woman.201 The stored material is Sudan black and PAS positive in frozen sections and is often birefringent in polarized light. It appears as myelinoid, lamellated, electrondense, intralysosomal, zebra-like bodies, 0.5 to 0.75 ␮m in diameter, under the electron microscope.445

Type II Glycogenosis (Pompe’s Disease) and Other Glycogen Storage Diseases Pompe’s disease was the first lysosomal enzyme deficiency to be recognized, and is the result of a mutation in the gene coding for acid maltase (acid ␣-1,4-glycosidase), located to chromosome 17q23.185,186 Spinal motor and DRG neurons and enteric neurons of the myenteric and submucosal plexuses show evidence of excessive glycogen storage in the infantile276 and childhood,287 but not adult,477 forms. The glycogen deposits appear under the light microscope as masses of PAS-positive granules in the cytoplasm of neurons, which often displace the nucleus and organelles to one end of the neuron.276 Defects in the glycogen branching enzyme cause glycogenosis type IV (Andersen’s disease), and other glycogenmetabolizing enzyme defects produce what are collectively now known as adult polyglucosan body diseases. These diseases lead to the formation of characteristic inclusions in neurons and glia of the central and peripheral nervous systems called corpora amylacea, or polyglucosan bodies

710

Pathology of the Peripheral Nervous System

(PGBs).48 These are strongly PAS positive and are usually found in axons rather than perikarya, but in at least one autopsy report of a patient in early infancy with glycogenosis type IV, PAS-positive PGBs were observed in spinal motor neurons.453 An autopsy of an infant with glycogen branching enzyme and phosphorylase deficiency found extensive storage of PAS-positive material in spinal and cranial motor neurons, DRG neurons, and enteric neurons.187 This material was present as membrane-bound granular aggregates around the Golgi apparatus and as large filamentous bodies similar to PGBs.

Chédiak-Higashi Syndrome Chédiak-Higashi syndrome is a multisystem disorder that is due to defective lysosomal trafficking regulator protein,489 the gene encoding which is located to chromosome 1q42q44.128 The disease is characterized pathologically by the occurrence of anomalous giant granules in leukocytes, as well as occasional lymphocytic infiltration of sympathetic and sensory ganglia. Pathologic changes to spinal motor neurons, such as chromatolysis and degeneration, appear to be related to the level of spinal cord examined: lumbosacral segments were the most affected, with little alteration seen at thoracic or cervical levels in two autopsies.438 Three types of inclusion body were observed in DRG and spinal motor neurons: spherical cytoplasmic bodies, “glomerate” bodies, and small discrete granules. Because the last two types of body were never seen in the same neurons as the large spherical bodies, the large spherical bodies may have been formed by fusion of the smaller ones. All types of body were strongly PAS positive and autofluorescent.

Leukodystrophies Metachromatic leukodystrophy is normally caused by a deficiency of activity of arylsulfatase, the gene for which is located in the region of chromosome 22q13.31.321 However, a deficiency of sphingolipid activator protein 1 (gene located to chromosome 10) produces a very similar syndrome. The only pathology described in the peripheral nervous system was the presence of PAS-positive MCBs in enteric neurons.393 Adrenomyeloneuropathy (also known as adrenoleukodystrophy) caused atrophy, but no significant loss, of large DRG neurons in three patients with this disease; many mitochondria contained “lipidic” inclusions.355,356 Globoid cell leukodystrophy (or Krabbe’s disease) is caused by a deficiency of galactocerebrosidase, an enzyme involved in myelin metabolism, the gene for which is located in the region of chromosome 14q31.42 This leads to myelin breakdown and a demyelinating neuropathy. Although few data are available on pathologic alterations in the human disease, membrane-bound bodies containing irregularly striated or lamellated material were described

in DRG neurons of the “twitcher” mouse, which has an identical enzyme deficiency.93

Refsum’s Disease Also known as heredopathia atactica polyneuritiformis, Refsum’s disease is a peroxisomal disorder that leads to the accumulation of phytanic acid throughout the body and subsequently ataxia, blindness, deafness, skeletal hyperostosis, and peripheral neuropathy. Around 45% of patients with Refsum’s disease were found to have mutations in the phytanoyl–coenzyme A hydroxylase gene located in the region of chromosome 10p13,497 indicating that it is a heterogeneous syndrome. There is at least one report of autopsy findings,52 and cultured rat DRG neurons exposed to phytanic acid accumulated non–membrane-bound osmiophilic inclusions at the periphery of perikarya and within mitochondria.91

Tangier Disease (Hereditary High-Density Lipoprotein Deficiency) Tangier disease is inherited as an autosomal codominant disease in which some heterozygotes have a peripheral neuropathy and premature coronary heart disease. Cholesterol ester deposition in various organs, particularly those of the reticuloendothelial system, as well as peripheral neuropathy, characterize homozygotes. It is caused by mutations in a gene that encodes an ATP-binding cassette transporter 1, localized to chromosome 9q31,39 and results in defective apolipoprotein-mediated efflux of cellular cholesterol and phospholipids and a marked deficiency of serum highdensity lipoprotein.375 In an autopsy study of a patient with a homozygous form of the disease, there was evidence of neuronal death in the L5 DRGs, with small-diameter neurons being particularly affected. Remaining neurons were found to contain large, membrane-bound lipid inclusions resembling giant lipofuscin granules. The neuronal inclusions were around 3 ␮m in diameter, were nonpigmented, had electron-dense and electron-lucent components, and displayed only weak autofluorescence.396 Spinal motor neurons also contained these inclusions, but there was no evidence of their loss at the sacral level. However, in another report of an autopsy of a patient with a homozygous form, the authors described a severe loss of motor neurons at the cervical level of the spinal cord and in the facial nucleus, but only a slight loss at lumbar levels.11

NEUROLOGIC MUTANTS These animals are the result of spontaneous mutations that produce a peripheral neuropathy, and are included where they have contributed to the understanding of equivalent human diseases.

711

Pathology of Peripheral Neuron Cell Bodies

Dystonia Musculorum The mutation is transmitted by a single autosomal recessive gene (dt), which has been mapped to mouse chromosome 6p12.40 It leads to alterations in the structure of dystonin, a cytoskeletal link protein forming bridges between F-actin and intermediate filaments.81 The high level of expression of dystonin in neurons of the cranial and spinal sensory ganglia matches those populations vulnerable to degeneration in dt mice.27 By the age of 3 to 4 weeks, the animal’s movements appear jerky and progressively less coordinated: The mice cannot now make purposeful movements, and affected animals usually die at about the time of weaning. Animals are born with fewer neurons in all spinal and cranial (5th, 7th, 9th, and 10th) ganglia examined, with the remaining neurons often appearing chromatolytic.95

The Sprawling Mouse and the Mutilated Foot Rat These two neurologic mutants have similar clinical and pathologic phenotypes, but do not totally match any known human neuropathy or transgenic mouse neuropathy. The most similar human disorders would be the ulcero-mutilating neuropathies HSAN I/HMSN IIB.99 The mouse mutation (Swl) is transmitted as an autosomal dominant trait, while the rat mutation (mf) is transmitted as an autosomal recessive trait, which has recently been found to be due to a mutation in a gene on chromosome 14 (M. J. Groves and F. Scaravilli, unpublished observations). Both the mouse and the rat manifest ataxia and loss of nociception in the limbs from birth: The ataxia progresses and the feet become ulcerated and eventually mutilated, although mutilation (not observed in the Swl mouse) can be seen occasionally at birth. There is neither muscle weakness nor wasting; animals that survive the early postnatal period usually live for a normal length of time. Pathologic findings shared by the adult forms of the two mutants consist of severe reduction in the number of both large- (with myelinated axons) and small- (with unmyelinated axons) diameter neurons of the lumbar and cervical DRGs. In the mf rat this reduction is around 50% in L4 and L5 DRGs, but some ganglia were reported as containing only 25% or less of the normal number of neurons.94,214 Motor neurons and axons are also decreased in number, although the reduction is less severe and mainly affects the smaller gamma motor neurons. Cranial and sympathetic ganglia and thoracic DRGs are reported as having normal numbers of neurons.214 The only morphologic difference between the surviving motor and sensory neurons of the mf rat and normal rats is that the remaining motor neurons are approximately 25% larger than the largest seen in normal littermates.388 A striking difference between the Swl mouse and the mf rat consists in the presence of light and electron microscopic features of

chromatolysis in some of the residual ganglion neurons in the Swl mouse. This loss of sensory neurons is due to events in development. As early as at the 11th embryonic day (E11), neurons with abnormally indented nuclei are present in the nascent DRGs of the Swl mouse. From E15 to E20–21, excessive numbers of immature DRG neurons die through apoptosis in the cervical and lumbar DRGs of both mf rats and Swl mice.58,386 This loss of sensory neurons at this time probably underlies the subsequent loss of gamma motor neurons330,388 and second-order sensory neurons in the gracile nucleus.58

The Wobbler Mouse This spontaneous mutation (gene symbol wr114) is located on chromosome 11 and is transmitted in an autosomal recessive manner.365 It leads to a progressive degeneration of motor neurons with a time course and pattern similar to those usually seen in human motor neuron disease. The earliest signs of the disease become evident during the fourth week, and progressive weakness, particularly in the forelimbs, soon becomes obvious and is linked to numbers of surviving motor neuron axons.351 Early morphologic changes in spinal and cranial motor neurons of the wobbler mouse include swelling and chromatolysis,31,96 and at later stages perikaryal vacuoles derived from dilated cisternae of the rough endoplasmic reticulum appear and may extend into the dendritic processes.8 Large motor neurons appear to be particularly affected, either by loss or lack of differentiation.31 These cellular changes seem to result in their death via an apoptotic pathway, involving the activation of protease-activated receptor 1, the incidence of which is greater at cervical than at lumbar levels of the spinal cord.118

Mouse Motor Neuron Degeneration This spontaneous mutation in C57Bl/6 mice follows an autosomal dominant inheritance pattern, and produces a disorder that begins by 6 months of age with limping of the hind limbs and progresses to spastic paralysis of the hind and, later, forelimbs.293 The mutation produces a progressively larger loss of spinal and cranial motor neurons, particularly those in the lumbosacral region, which undergo chromatolytic changes and accumulation of ubiquitinated and autofluorescent lipofuscinoid inclusions that contain various lysosomal enzymes.289,294 The apoptosis-associated enzyme transglutaminase type 2 is “superactivated” in the spinal cord at the same time that the motor neuron disease begins to manifest itself.196 It is now apparent that this mouse is a homologue196 of a human neuronal ceroid lipofuscinosis (progressive epilepsy with mental retardation) caused by a mutation in the CLN8 gene, which encodes for a membrane protein with as yet unknown functions.364

712

Pathology of the Peripheral Nervous System

Inherited Dog Neuropathies The neuropathy in the Brittany spaniel dog is an inherited motor neuron disease (hereditary canine SMA) that is transmitted as an autosomal dominant trait.377 The clinical course of the intermediate (and most common) heterozygous phenotype resembles that of the KugelbergWelander type of juvenile SMA. Furthermore, the accelerated, homozygous form of the disease is of interest in the context of both SMA and possible molecular mechanisms in motor neuron disease. Analysis of the ventral horn cholinergic neuron population showed that large motor neurons either fail to reach the correct size or undergo atrophy.69 Subsequent analysis has shown that alterations in cyclin-dependent kinase 5 expression precede altered expression of low-molecular-weight neurofilament protein155 and subsequent inhibition of neurofilament assembly or transport.311 An inherited sensory neuropathy has been reported in a litter of nine English Pointer dogs that presented as acral mutilation and analgesia78 very similar to that suffered by the mf rat (see above). Upon examination, the density of DRG neurons was reduced: Some vacuolated and chromatolytic neurons and some nodules of Nageotte were present, but the authors believed that much of the neuron loss (25% to 50% in DRGs) had probably occurred before birth and had particularly affected the large neuron population.

NEUROPATHIES PRODUCED BY SYSTEMIC METABOLIC DISORDERS Porphyria This group of conditions is due to defects in any of the enzymes and proteins involved in porphyrin metabolism, and only those forms caused by defects in hepatic heme synthesis produce a distal axonopathy. They are inherited in an autosomal dominant fashion. Autopsies have found either no evidence of loss of either spinal motor or DRG neurons49 or evidence of some loss of both. 188,460 Chromatolysis was seen in some of the larger spinal motor neurons in the cervical and lumbar enlargements as well as occasional vacuolated neurons in the lumbar region; chromatolysis was less evident in DRG neurons.49

Diabetes Mellitus This disease produces a loss of axons in sural nerve biopsies, but evidence that this is directly related to a loss of DRG neurons, at least initially in the progression of neuropathy, is by no means certain. However, pathologic alterations of spinal cord motor neurons, DRGs, and autonomic ganglia in autopsy specimens from diabetics have been described.

DRGs from a patient with a 50-year history of diabetes contained sparse infiltrates of mononuclear cells without any obvious neuron loss, as well as no obvious loss of spinal motor neurons.492 This case also demonstrated atrophy of the superior cervical ganglion with prominent neuron loss and lymphocytic infiltration, but a normal myenteric plexus in both stomach and intestine. Schmidt398 also found lymphocytic infiltration of autonomic ganglia, but found that it was not selectively increased in frequency and intensity compared to controls. An absence of significant changes in DRG and motor neurons of diabetic patients was also reported by Schmidt et al.400 However, DRG neuron vacuolation was a significant finding in three of five autopsies of diabetic patients, but only one or two of these cases showed evidence of either motor or sensory neuron loss.156 Pathologic alterations in autonomic neurons seem to be more significant: Abnormally enlarged sympathetic ganglion cells containing PAS-positive material have been observed.13,156 Infiltration of autonomic ganglia by inflammatory cells, vacuolated and enlarged sympathetic neurons, club-shaped enlargements of neuronal processes, and the formation of “neuromatous nodules” have been reported in each of five autopsies of diabetic patients with a severe autonomic neuropathy.92 Experimental work in rats has shown that hyperglycemia can induce apoptosis in DRG and Schwann cells,374 but studies of animal models of diabetes have found DRG neuron atrophy, rather than loss, a more significant feature.512,533 Further experimental work showed abnormal phosphorylation of neurofilaments of DRG cells both in spontaneous and streptozotocin-induced diabetic rats,117 as well as decreased expression of cytoskeletal proteins and their reduced incorporation into distal axons.405 A study of BB/Wor diabetic rats showed normal histologic appearances and numbers of motor neurons at spinal level L5.300

Amyloidosis The types, presentation, and treatment of amyloidoses are covered in Chapter 83. This section deals with any pathologic alterations to peripheral neuron perikarya described in amyloidosis. Amyloidosis is a disorder of protein conformation involving proteins normally present in the body fluids as soluble precursors. An alteration to the genetic coding for these proteins, or to other proteins involved in the formation of their tertiary and quaternary structure, can lead to their deposition as insoluble fibrils in tissues. The fibrils consist of self-assembled, low-molecular-weight mass peptides with beta pleated sheet staining characteristics,125 and seem to be directly toxic to neurons. Both peripheral nerves and ganglia are affected in cases of primary and familial amyloidoses. The pathologic description that follows is based on personal observations

713

Pathology of Peripheral Neuron Cell Bodies

on the autopsy of a patient, 49 years old, suffering from a familial form of type I amyloid polyneuropathy. In the spinal cord, a number of meningeal vessels showed focal deposits of amyloid; there were no obvious signs of degeneration of motor neurons, but some were chromatolytic. The dorsal columns showed striking pallor involving both the gracile and cuneate tracts. Several muscles from the upper and lower limbs were sampled; they showed a mild to moderate degree of denervation. DRGs (see Fig. 31–9B and C) were severely affected. Areas of amyloid formed irregular aggregates replacing and separating groups of neurons. The latter were severely reduced in number, as confirmed by the large number of nodules of Nageotte, and atrophy of the surviving neurons was suggested by hypertrophy of their surrounding satellite cells, which formed concentric layers around them. A number of the surviving neurons showed signs of degeneration and chromatolysis. In addition, amyloid was present infiltrating endoneurial blood vessels and was seen to continue into the roots and peripheral nerves, but there was no evidence of an inflammatory cell infiltrate. The superior cervical and celiac sympathetic ganglia showed comparable changes, with the nervous tissue being largely replaced by areas of amyloid deposition. Increased numbers of nodules of Nageotte, evidence of chromatolysis, and absence of inflammatory cells were features similar to those seen in the sensory ganglia. In striking contrast, the two gasserian ganglia of the same patient were minimally involved by the changes and showed a small number of nodules of Nageotte and no amyloid deposits. The changes described above are similar to those reported by a number of authors in patients with familial forms of amyloidosis.297,425,467 All observers agree that spinal cord motor neurons are not decreased in number, although they show evidence of chromatolysis, and that the posterior columns are pale. With regard to the severity of the changes in dorsal root and autonomic ganglia, unlike other groups, Misu et al.297 reported only moderate neuron loss and amyloid deposition in patients with familial amyloid polyneuropathy TTR Met 30.

INFECTIVE AND INFLAMMATORY NEUROPATHIES Varicella Zoster and Herpes Zoster Infection Following primary infection, varicella-zoster virus (VZV) establishes latency in DRG neurons and can reactivate years later to produce herpes zoster infection; studies in humans have confirmed that VZV DNA persists in DRG neurons and VZV proteins are expressed during the latent period,274 although there have been conflicting reports about the cellular localization of VZV within DRGs. In a study of DRGs of humans and rats, Annunziato et al.10 and

Lungu et al.271 detected VZV in both satellite and nerve cells; in contrast, Kennedy et al.230 found VZV DNA predominantly in the latter. When reactivated, the lesions are often localized to a dermatome, and pathologic changes such as hemorrhagic infarction, massive lymphocytic infiltration, and thrombosed small arteries may be found in the corresponding DRGs, roots, and spinal cord. Within these areas there is usually massive loss of DRG neurons, and those remaining show shrinkage, chromatolysis, and vacuolation.90 Motor symptoms may also complicate herpes zoster.140

Poliovirus Poliovirus is the etiologic agent that causes paralytic poliomyelitis by infecting and killing motor neurons, and that can produce additional neurologic problems (postpolio syndrome) years after the initial infection and disease. This latter effect may be due to persistence of the virus in neural and non-neural cells.67 Studies in monkeys show that infected motor neurons may increase in size303 before chromatolysis, nuclear pyknosis, and neuronophagia (see Fig. 31–4), leading to their loss.379

Cytomegalovirus Infection Cytomegalovirus (CMV) is another member of the Herpesviridae family and has been identified as a pathogen in several mammals, including humans. Cytomegalovirus involvement of the nervous system has become very frequent with the advent of human immunodeficiency virus (HIV) infection,387 and peripheral neuropathies resulting from CMV infection are becoming relatively common.129,307 Eidelberg and co-workers103 examined DRGs from patients with acquired immunodeficiency syndrome (AIDS) and CMV polyradiculopathy and found them to be unaffected, although spinal motor neurons often showed chromatolytic changes secondary to root damage. In a personally observed case, a patient was found to have florid lymphocytic ganglionitis with intranuclear inclusions in a small number of ganglion cells. A study by Nagano et al.317 found CMV antigen in the DRGs from 4 of 10 patients, but also in 1 of 5 control DRGs. However, no intranuclear or intracytoplasmic inclusions were seen by this group, and no relationship could be established between the expression of CMV antigen, density of nodules of Nageotte, and other pathologic changes.

Human Immunodeficiency Virus Peripheral neuropathies related to HIV and human T-lymphotropic virus type 1 (HTLV-1) infection are considered in Chapter 112, and generally present as distal sensory polyneuropathies.361 The main pathologic features seen in DRGs are inflammatory cell infiltration with loss of

714

Pathology of the Peripheral Nervous System

neurons and the formation of nodules of Nageotte; these are also seen in patients with antiretroviral drug toxic neuropathies.341 The inflammatory component in the ganglia consists of macrophages and a diffuse infiltration with T lymphocytes, described in seven patients with AIDS, two of whom also had evidence of CMV ganglionitis.389 In a study of sensory and sympathetic ganglia from 12 HIVpositive patients,111 many ganglia contained lymphocytes and macrophages as well as showed expression of major histocompatibility complex class II and occasional degenerating nerve cells. The gp41 HIV-1 antigen was found in larger amounts in sensory than in sympathetic ganglia.

DRGs160 showing collections of T lymphocytes around neurons. In milder cases there was degeneration of individual ganglion neurons; in an advanced case only a few ganglion cells remained. Inflammatory cell infiltrates have also been described in DRGs in Sjögren’s syndrome associated with lymphoma.197 The pathogenetic mechanism of neuron damage and loss in Sjögren’s syndrome remains unclear, but immunostaining of rat small sensory neurons by serum and cerebrospinal fluid (CSF) of a 59-year-old woman suffering from the syndrome suggests an autoimmune-related mechanism.383

Lymphoma Other Viruses Producing Peripheral Neuropathy A motor neuron disease in the mouse caused by “type C” retrovirus has been reported.135 The disorder is transmitted via the mother’s milk, begins at 6 months, and is characterized by hind limb paralysis and spongiform degeneration. It is clinically similar to that described in Mnd mice; however, the former is associated, in 10% of the animals, with lymphoma. HTLV-1 is the cause of endemic tropical spastic paraparesis (TSP) or HTLV-1–associated myelopathy. An autopsy of a TSP case found increased lipofuscin deposition in DRG neurons, but these were otherwise normal.262

Diphtheria The bacterium that causes this disease, Corynebacterium diphtheriae, produces a toxin that inactivates elongation factor 2, which is required for protein synthesis. When this toxin enters the systemic circulation, one of the clinical manifestations is a demyelinating neuropathy. When the purified toxin is injected into muscles, marked morphologic alterations to motor neurons occur, including the dilatation and fragmentation of rough endoplasmic reticulum similar to that produced by ricin.358

Leprosy There are few reports of pathologic alteration to peripheral neurons in leprosy, but experimental inoculation of nude mice with Mycobacterium leprae resulted in a progressive depletion of substance P and CGRP from sensory neurons, and a total loss of CGRP from motor neurons 12 months later.223

Sensory Ganglionitis in Sjögren’s Syndrome An ataxic sensory and autonomic neuropathy may develop in patients with Sjögren’s syndrome. The primary involvement of sensory neurons has been confirmed in biopsies of

There are occasional reports of peripheral neuropathy being a remote effect of lymphoma, as opposed to a neuropathy caused by lymphomatous infiltration of peripheral nerve. Autopsies of five patients with a subacute lower motor neuron syndrome found prominent motor neuron (see Fig. 31–8B and C) degeneration,402 and a severe depletion of DRG neurons with inflammatory changes was found in autopsies of patients with a subacute sensory neuronopathy.197

Paraneoplastic Neuropathy Paraneoplastic neuropathies are associated with systemic malignancies, and their presentation may precede the clinical appearance of the neoplasm. Their pathogenesis is not well established, although increasing evidence supports an autoimmune mechanism.385 Involvement of sensory and autonomic ganglia (ganglioradiculoneuritis) in paraneoplastic disorders may be focal or diffuse, and in the former it can be asymmetrical and localized to the cervical or lumbar region. The degree of severity of the lesions varies, and complete destruction of the affected ganglion can result. Disappearance of neurons is accompanied by proportional increase of nodules of Nageotte, whereas the remaining nerve cells may show cytoplasmic vacuolation, chromatolysis, and nuclear shrinkage. Inflammation can appear as diffuse lymphocytic infiltration or cuffing of the vessels, and may extend to the posterior roots. Henson and Urich182 questioned the existence of two separate subgroups of ganglionopathies, inflammatory and noninflammatory. They argued that the likelihood of finding inflammatory cells depends on the number of ganglia available and the duration of the illness, because the inflammatory process tends to fade and disappear with time. Disappearance of ganglion cells is followed by atrophy of the corresponding posterior roots, with the degree of posterior column pallor being proportional to the extent and severity of the neuronal loss. Involvement of the autonomic nervous system has also been described, although nerve cell loss in the sympathetic ganglia has not been reported to reach the severity of that

Pathology of Peripheral Neuron Cell Bodies

seen in sensory ganglia; this involvement may be accompanied by a lymphocyte infiltrate.219 The parasympathetic component can also be affected,426 and inflammation in the myenteric plexus has also been described.260 The existence of a pure form of motor neuron disease in paraneoplastic disorders, in addition to motor involvement as part of a poliomyelitis affecting both anterior and posterior horns, has been debated for a long time.385 Once the cases of paraneoplastic motor neuron disease associated with myelitis are excluded, a pure degenerative form of paraneoplastic motor neuron disease is very rare. However, Verma et al.483 detected anti-Hu antibodies in a 51-year-old man with small cell cancer of the lung who developed weakness and muscle wasting in the upper half of the body. Electromyography showed signs of denervation, and spinal motor neuron loss was found post mortem. The serum and CSF of a large number of patients suffering from the disorders described above contain polyclonal antibodies, called anti-Hu (also type 1 antineuronal nuclear antibodies, or ANNA-1).80 Virtually all patients belonging to this subgroup have (or have had) small cell lung carcinoma, while a small number suffer from neuroblastoma or breast or prostate carcinoma. The characteristic feature of these antibodies is that they stain the nuclei and, discretely, the cytoplasm of certain neurons of both the central and peripheral nervous systems.146

TOXINS Hexacarbons Exposure to toxic doses of hexacarbons such as n-hexane produces a distal axonopathy in humans by interfering with neurofilament structure and assembly. The neurotoxic metabolite 2,5-hexanedione is thought to interact directly with lysine residues, which are prevalent in neurofilament proteins, and n-hexane neuropathy has morphologic alterations to axons similar to those seen in GAN. Investigations into the effects of 2,5-hexanedione on cultured human DRG neurons found tangles of neurofilaments in the perikarya of neurons, as well as in their axons.306 Systemic administration of 2,5-hexanedione induces chromatolysis in rat DRG neurons over a time course of weeks, rather than days as seen following peripheral axotomy.432 Chronic dosing eventually led to the formation of nodules of Nageotte,431 and 2,5-hexanedione itself does not appear to be as acutely toxic to neurons as it is to glia.433

Metals “Heavy” metals (such as lead, mercury, arsenic, and cadmium) often have an affinity for nucleophilic moieties, particularly sulfhydryl groups, with intoxication tending to affect membrane transport processes. Organometal com-

715

pounds are generally far more toxic than inorganic salts, probably as a result of their greater lipid solubility and concomitantly more rapid and wider distribution in vivo. Jacobs et al.212,213 originally demonstrated the selective effect of methylmercury upon large-diameter rat and rabbit DRG neurons. A dose of 7.5 mg/kg/day for 1 to 4 (rabbits) or 8 (rats) days initially produced a loss of ribosomes from rough endoplasmic reticulum. This was followed by nuclear and perinuclear changes, cytoplasmic vesiculation, and necrosis of some neurons with evidence of neuronophagia.213 At a dose of 5 mg/kg/day for 10 days, largediameter rat DRG neurons demonstrated significant delays in action potential onset; vacuolation and occasional evidence of neuronophagia were also observed in DRGs.89 Methylmercury intoxication also affects motor neurons, producing vacuolation of this population in rats.434 Indeed, acute methylmercury intoxication leads to the death of rat spinal motor neurons (as indicated by evidence of neuronophagia) by the 16th day after 10 days of daily injections of 10 mg/kg, with all large motor neurons disappearing by 18 days.434 The organotin compounds triethyl tin and trimethyl tin are toxic to myelinating Schwann cells and neurotoxic, respectively: 6 mg/kg of trimethyl tin orally administered to rats produced vacuolar changes, accumulation of lysosomes, formation of myeloid bodies, and dissolution of the Nissl substance in DRG neurons (Fig. 31–10A) up to 2 weeks later.53 Vacuolation was also seen in motor neurons of gerbils and rats with trimethyl tin intoxication.325 Lead intoxication produces a differing pattern of neuropathy depending upon the species studied and dosing schedule used. One of the few reports of the histopathology of peripheral neurons in lead poisoning involved long-term intoxication in chickens, which produced degeneration of motor neurons that the authors suggested was similar to that seen in human motor neuron diseases.288 Platinum is particularly neurotoxic to sensory neurons, neuropathy often being the result of treatment with platinum-based cytotoxic drugs such as cisplatin. Cisplatin induces rat DRG neurons to undergo apoptosis in vitro and in vivo,120,143 and systemic exposure to 2 mg/kg twice weekly for 4.5 weeks produced reductions in perikaryal, nuclear, and nucleolar areas of DRG neurons in rats.468 “Necrotic” (meaning dying rather than true necrosis) neurons, nodules of Nageotte, and reduced perikaryal volumes have been reported in DRGs from autopsies of some patients treated with cisplatin and carboplatin.246 Studies using cultured DRG neurons showed that exposure to aluminum phosphate produced intraneuronal neurofibrillary spheroids,407 as well as a decreased expression of neurofilament proteins.141 All of these changes were reversible if dosing with aluminum was discontinued. Chronic administration of aluminum chloride to rabbits was reported to produce a loss of CNS and spinal motor neurons via a process of neurofibrillary degeneration.228

716

Pathology of the Peripheral Nervous System

FIGURE 31–10 A, A degenerating rat DRG neuron following acute intoxication with triethyl tin. A presumptive nucleus of a phagocytic cell (arrow) apparently within the degenerating perikaryon probably indicates the beginning of neuronophagia (see Fig. 31–4D). The apparent dissolution of the nucleus (arrowheads) and vesiculation of the cytoplasm indicates an acute, nonapoptotic form of neuron death probably involving loss of membrane integrity or homeostatic function. Bar: 10 ␮m. (Courtesy of J. M. Jacobs.) B, Photomicrograph of a toluidine blue–stained resin section of adult rat DRG shows chromatolytic large-diameter neurons (arrows) following intraperitoneal injection of pyridoxine. Excessive ingestion of pyridoxine (vitamin B6) is toxic to large sensory neurons in humans (and rats). Acutely toxic doses lead to loss of neurons while lower doses produce axonal atrophy. Bar: 10 ␮m. (Courtesy of J. M. Jacobs and M. Pires.) C and D, Amiodarone can produce a sensorimotor neuropathy of a mixed axonal and demyelinating type, which is associated with the presence of intracellular electron-dense, lamellated inclusions. C, Photomicrograph showing a toluidine blue–stained resin section of adult rat celiac ganglion following amiodarone intoxication. The osmiophilic inclusions (arrows) can be seen within the neuronal perikarya. Bar: 10 ␮m. (Courtesy of J. M. Jacobs.) D, Electron micrograph of an adult rat superior cervical ganglion neuron 2 weeks after administration of amiodarone. The electron-dense inclusions characteristic of amiodarone neuropathy can be seen in neurons and axons (arrows). Bar: 1 ␮m. (Courtesy of J. M. Jacobs.)

Cadmium is directly toxic to a wide variety of tissues, and also exerts secondary effects through its action as a very powerful vasoconstrictor. Acute intoxication has been reported by Gabbiani et al. to produce hemorrhagic lesions in dorsal root and fifth cranial nerve ganglia, but in addition chromatolytic and (apparently) necrotic neurons are clearly visible in their figures.130 Other metals, such as thallium and gold, which have been reported to produce neuropathy, have not had any

associated pathologic alterations to peripheral neurons described.

Vincristine Vincristine is a microtubule-depolymerizing drug used in the treatment of certain cancers that produces a painful sensory neuropathy in humans. When administered to rats, it induced the swelling of large-diameter DRG neurons,

Pathology of Peripheral Neuron Cell Bodies

possibly caused by the observed buildup of neurofilaments in their perikarya and proximal segments of axon.464

Pyridoxine Pyridoxine (vitamin B6), when taken in excessive amounts, is a sensory neurotoxin that interferes with sensory neuron (see Fig. 31–10B) cytoskeletal organization.304 A dose of 600 to 1200 mg/kg/day administered to rats produced necrosis of DRG neurons, particularly large-diameter neurons513; these authors also reported substantial species susceptibilities to pyridoxine neurotoxicity.

Amiodarone Amiodarone is a diiodinated benzofuran derivative used to treat cardiac arrhythmias, and has been associated with a number of cases of peripheral neuropathy. A dose of 50 mg/kg/day administered orally to rats induced accumulations of lipids within lysosomes, leading to the formation of cytoplasmic bodies in neurons from regions without bloodbrain or blood-nerve barriers. Small to medium-sized DRG and gasserian ganglion neurons, autonomic neurons, and enteric neurons subsequently showed inclusions (see Fig. 31–10C and D), with autonomic neurons the worst affected and showing evidence of degenerative changes.70

Acrylamide Chronic low-level exposure to acrylamide produces a distal axonopathy, and animals exposed to acute high doses (30 mg/kg/day) were found to have chromatolytic and vacuolated DRG neurons.220,430 Acrylamide does not appear to produce a loss of DRG neurons, just atrophy (by up to 28%) of the large-diameter subpopulation, indicating that alterations to neurofilament function may underlie its neurotoxicity.451

Organophosphates Exposure to organophosphates has been linked to a neuropathy termed organophosphate-induced delayed neuropathy, an axonopathy affecting neurons with long axons. Their neurotoxicity is thought to be due to interaction with “neuropathy target esterase,” which is located in neuronal cell bodies and axons.149 One of the effects of this interaction appears to be the hyperphosphorylation of tubulin and microtubule-associated protein-2, which in turn may cause destabilization of microtubule assemblies and axonal degeneration.62

REFERENCES 1. Acheson, A., and Lindsay, R. M.: Non-target-derived roles of the neurotrophins. Philos. Trans. R. Soc. Lond. B 351:417, 1995.

717

2. Adalbert, R., Engelhardt, J. I., and Siklos, L.: DLHomocysteic acid application disrupts calcium homeostasis and induces degeneration of spinal motor neurons in vivo. Acta Neuropathol. (Berl.) 103:428, 2002. 3. Aldskogius, H., and Arvidsson, J.: Nerve cell degeneration and death in the trigeminal ganglion of the adult rat following peripheral nerve transection. J. Neurocytol. 7:229, 1978. 4. Aldskogius, H., Arvidsson, J., and Grant, G.: Axotomyinduced changes in primary sensory neurons. In Scott, S. A. (ed.): Sensory Neurons. New York, Oxford University Press, p. 363, 1992. 5. Aldskogius, H., and Risling, M.: Effect of sciatic neurectomy on neuronal number and size distribution in the L7 ganglion of kittens. Exp. Neurol. 74:597, 1981. 6. Alkhayat, A. H., Kraemer, S. A., Leipprandt, J. R., et al.: Human beta-mannosidase cDNA characterization and first identification of a mutation associated with human betamannosidosis. Hum. Mol. Genet. 7:75, 1998. 7. Altarescu, G., Sun, M., Moore, D. F., et al.: The neurogenetics of mucolipidosis type IV. Neurology 59:306, 2002. 8. Andrews, J. M.: The fine structure of the cervical spinal cord, ventral root and brachial nerves in the wobbler (W2) mouse. J. Neuropathol. Exp. Neurol. 34:12, 1975. 9. Andrews, T. J.: Autonomic nervous system as a model of neuronal aging: the role of target tissues and neurotrophic factors. Microsc. Res. Tech. 35:2, 1996. 10. Annunziato, P. W., La Russa, P., Lee, P., et al.: Evidence of latent varicella-zoster virus in rat dorsal root ganglia. J. Infect. Dis. 178(Suppl. 1S):48, 1998. 11. Antoine, J. C., Tommasi, M., Boucheron, S., et al.: Pathology of roots, spinal cord and brainstem in syringomyelia-like syndrome of Tangier disease. J. Neurol. Sci. 106:179, 1991. 12. Apfel, S. C., Wright, D. E., Wiideman, A. M., et al.: Nerve growth factor regulates the expression of brain-derived neurotrophic factor mRNA in the peripheral nervous system. Mol. Cell. Neurosci. 7:134, 1996. 13. Appenzeller, O., and Richardson, E. P.: The sympathetic chain in patients with diabetic and alcoholic polyneuropathy. Neurology (Minneap.) 16:1205, 1966. 14. Arvidsson, J., Ygge, J., and Grant, G.: Cell loss in lumbar dorsal root ganglia and transganglionic degeneration after sciatic nerve resection in the rat. Brain Res. 373:15, 1986. 15. Arvidsson, U., Johnson, H., Piehl, F., et al.: Peripheral nerve section induces increased levels of calcitonin generelated peptide (CGRP)-like immunoreactivity in axotomized motoneurons. Exp. Brain Res. 79:212, 1990. 16. Bahadori, M. H., Al-Tiraihi, T., and Valojerdi, M. R.: Sciatic nerve transection in neonatal rats induces apoptotic neuronal death in L5 dorsal root ganglion. J. Neurocytol. 30:125, 2001. 17. Barrett, G. L., and Bartlett, P. F.: The p75 nerve growth factor receptor mediates survival or death depending on the stage of sensory neuron development. Proc. Natl. Acad. Sci. U. S. A. 91:6501, 1994. 18. Bates, G. P., Mangiarini, L., and Davies, S. W.: Transgenic mice in the study of polyglutamine repeat expansion diseases. Brain Pathol. 8:699, 1998.

718

Pathology of the Peripheral Nervous System

19. Beaver, D. L., Moses, H. L., and Ganote, C. E.: Electron microscopy of the trigeminal ganglion. II. Autopsy study of human ganglia. Arch. Pathol. 79:557, 1965. 20. Beiswanger, C. M., Roscoe-Graessle, T. L., Zerbe, N., et al.: 3-acetylpyridine-induced degeneration in the dorsal root ganglia: involvement of small diameter neurons and influence of axotomy. Neuropathol. Appl. Neurobiol. 19:164, 1993. 21. Bendotti, C., Calvaresi, N., Chiveri, L., et al.: Early vacuolization and mitochondrial damage in motor neurons of FALS mice are not associated with apoptosis or with changes in cytochrome oxidase histochemical reactivity. J. Neurol. Sci. 191:25, 2001. 22. Bennett, D. L., Averill, S., Clary, D. O., et al.: Postnatal changes in the expression of the trkA high-affinity NGF receptor in primary sensory neurons. Eur. J. Neurosci. 8:2204, 1996. 23. Berciano, J., Combarros, O., Figols, J., et al.: Hereditary motor and sensory neuropathy, type II. Clinicopathological study of a family. Brain 109:897, 1986. 24. Bergman, E., Fundin, B. T., and Ulfhake, B.: Effects of aging and axotomy on the expression of neurotrophin receptors in primary sensory neurons. J. Comp. Neurol. 410:368, 1999. 25. Berlit, P., and Schwechheimer, K.: Neuropathological findings in radiation myelopathy of the lumbosacral cord. Eur. Neurol. 27:29, 1987. 26. Bernal, F., Graus, F., Pifarre, A., et al.: Immunohistochemical analysis of anti-Hu-associated paraneoplastic encephalomyelitis. Acta Neuropathol. (Berl.) 103:509, 2002. 27. Bernier, G., Brown, A., Dalpe, G., et al.: Dystonin expression in the developing nervous system predominates in the neurons that degenerate in dystonia musculorum mutant mice. Mol. Cell. Neurosci. 6:509, 1995. 28. Bieber, F. R., Mortimer, G., Kolodny, E. H., and Driscoll, S. G.: Pathologic findings in fetal GM1 gangliosidosis. Arch. Neurol. 43:736, 1986. 29. Bigotte, L., and Olsson, Y.: Cytotoxic effects of Adriamycin on mouse hypoglossal neurons following retrograde axonal transport from the tongue. Acta Neuropathol. (Berl.) 61:161, 1983. 30. Bledsoe, A. W., Jackson, C. A., McPherson, S., and Morrow, C. D.: Cytokine production in motor neurons by poliovirus replicon vector gene delivery. Nat. Biotechnol. 18:964, 2000. 31. Blondet, B., Carpentier, G., Ait-Ikhlef, A., et al.: Motoneuron morphological alterations before and after the onset of the disease in the wobbler mouse. Brain Res. 930:53, 2002. 32. Bodian, M., Stephens, F. D., and Ward, B. L. H.: Hirschsprung’s disease and idiopathic megacolon. Lancet 1:6, 1949. 33. Boettger, M. K., Till, S., Chen, M. X., et al.: Calciumactivated potassium channel SK1- and IK1-like immunoreactivity in injured human sensory neurons and its regulation by neurotrophic factors. Brain 125:252, 2002. 34. Bondok, A. A., and Sansone, F. M.: Retrograde and transganglionic degeneration of sensory neurons after a peripheral nerve lesion at birth. Exp. Neurol. 86:322, 1984. 35. Bradley, J. L., Blake, J. C., Chamberlain, S., et al.: Clinical, biochemical and molecular genetic correlations in Friedreich’s ataxia. Hum. Mol. Gen. 9:275, 2000.

36. Bredesden, D. E., and Rabizadeh, S.: p75NTR and apoptosis: trk-dependent and trk-independent effects. TINS 20:287, 1997. 37. Bridge, P. M., Ball, D. J., Mackinnon, S. E., et al.: Nerve crush injuries—a model for axonotmesis. Exp. Neurol. 127:284, 1994. 38. Brorson, J. R., Schumacker, P. T., and Zhang, H.: Nitric oxide acutely inhibits neuronal energy production. J. Neurosci. 19:147, 1999. 39. Brousseau, M. E., Schaefer, E. J., Dupuis, J., et al.: Novel mutations in the gene encoding ATP-binding cassette 1 in four Tangier disease kindreds. J. Lipid Res. 41:433, 2000. 40. Brown, A., Lemieux, N., Rossant, J., and Kothary, R.: Human homolog of the mouse sequence from the dystonia musculorum locus is on chromosome 6p12. Mamm. Genome 5:434, 1994. 41. Bussmann, K. A., and Sofroniew, M. V.: Re-expression of p75NTR by adult motor neurons after axotomy is triggered by retrograde transport of a positive signal from axons regrowing through damaged or denervated peripheral nerve tissue. Neuroscience 91:273, 1999. 42. Cannizzaro, L. A., Chen, Y. Q., Rafi, M. A., and Wenger, D. A.: Regional mapping of the human galactocerebrosidase gene (GALC) to 14q31 by in situ hybridisation. Cytogenet. Cell Genet. 66:244, 1994. 43. Carlson, C. D., Bai, Y., Ding, M., et al.: Interleukin-1 involvement in the induction of leukemia inhibitory factor mRNA expression following axotomy of sympathetic ganglia. J. Neuroimmunol. 70:181, 1996. 44. Carmel, P. W., and Stein, B. M.: Cell changes in sensory ganglia following proximal and distal nerve section in the monkey. J. Comp. Neurol. 135:145, 1969. 45. Carter, B. D., Kaltschmidt, C., Kaltschmidt, B., et al.: Selective activation of NF-kB by nerve growth factor through the neurotrophin receptor p75. Science 272:542, 1996. 46. Casanovas, A., Ribera, J., Hager, G., et al.: C-jun regulation in rat neonatal motoneurons postaxotomy. J. Neurosci. Res. 63:469, 2001. 47. Casanovas, A., Ribera, J., Hukkanen, M., et al.: Prevention by lamotrigine, MK-801 and N-omega-nitro-L-arginine methyl ester of motoneuron cell death after neonatal axotomy. Neuroscience 71:313, 1996. 48. Cavanagh, J. B.: Corpora-amylacea and the family of polyglucosan diseases. Brain Res. Rev. 29:265, 1999. 49. Cavanagh, J. B., and Mellick, R. S.: On the nature of the peripheral nerve lesions associated with acute intermittent porphyria. J. Neurol. Neurosurg. Psychiatry 28:320, 1965. 50. Cavanaugh, M. W.: Quantitative effects of the peripheral innervation area on nerves and spinal ganglion cells. J. Comp. Neurol. 94:181, 1951. 51. Celik, M., Gokmen, N., Erbayraktar, S., et al.: Erythropoietin prevents motor neuron apoptosis and neurologic disability in experimental spinal cord ischemic injury. Proc. Natl. Acad. Sci. U. S. A. 99:2258, 2002. 52. Cervos-Navarro, J.: Heredopathia atactica polyneuritiformis (Refsum’s disease). Histol. Histopathol. 5:439, 1990. 53. Chang, L. W., and Dyer, R. S.: Trimethyltin induced pathology in sensory neurons. Neurobehav. Toxicol. Teratol. 5:673, 1983.

Pathology of Peripheral Neuron Cell Bodies 54. Chang, Q., and Balice-Gordon, R. J.: Gap junctional communication among developing and injured motor neurons. Brain Res. Brain Res. Rev. 32:242, 2000. 55. Chang, Q., Pereda, A., Pinter, M. J., and Balice-Gordon, R. J.: Nerve injury induces gap junctional coupling among axotomized adult motor neurons. J. Neurosci. 20:674, 2000. 56. Chao, M. V., and Bothwell, M.: Neurotrophins: to cleave or not to cleave. Neuron 3:9, 2002. 57. Chard, P. S., Bleakman, D., Savidge, J. R., and Miller, R. J.: Capsaicin-induced neurotoxicity in cultured dorsal root ganglion neurons: involvement of calcium-activated proteases. Neuroscience 65:1099, 1995. 58. Chimelli, L., and Scaravilli, F.: The abnormal development of the gracile nucleus in the neurological mutant rat mf. Dev. Brain Res. 29:193, 1986. 59. Cho, H. J., Kim, S. Y., Park, M. J., et al.: Expression of mRNA for brain-derived neurotrophic factor in the dorsal root ganglion following peripheral inflammation. Brain Res. 749:358, 1997. 60. Chou, S. M., and Payne, W. N.: Vacuolation and chromatolysis of lower motoneurons in tetanus. A case report and review of the literature. Cleve. Clin. Q. 49:255, 1982. 61. Chou, S. M., and Wang, H. S.: Aberrant glycosylation/phosphorylation in chromatolytic motoneurons of WerdnigHoffmann disease. J. Neurol. Sci. 152:198, 1997. 62. Choudhary, S., Joshi, K., and Gill, K. D.: Possible role of enhanced microtubule phosphorylation in dichlorvos induced delayed neurotoxicity in rat. Brain Res. 897:60, 2001. 63. Clarke, P. G., and Clarke, S.: Nineteenth century research on naturally occurring cell death and related phenomena. Anat. Embryol. 193:81, 1996. 64. Clarke, P. G. H.: Apoptosis versus necrosis: how valid a dichotomy for neurons? In Koliatsos, V. E., and Ratan, R. R. (eds.): Cell Death and Diseases of the Nervous System. Totowa, NJ, Humana Press, p. 3, 1999. 65. Clatterbuck, R. E., Price, D. L., and Koliatsos, V. E.: Further characterization of the effects of brain-derived neurotrophic factor and ciliary neurotrophic factor on axotomized neonatal and adult mammalian motor neurons. J. Comp. Neurol. 342:45, 1994. 66. Coggeshall, R. E.: A consideration of neural counting methods. TINS 15:9, 1992. 67. Colbere-Garapin, F., Duncan, G., Paviao, N., et al.: An approach to understanding the mechanisms of poliovirus persistence in infected cells of neural and non-neural origin. Clin. Diagn. Virol. 9:107, 1998. 68. Cordeiro, P., Hechtman, P., and Kaplan, F.: The GM2 gangliosidoses databases: allelic variation at the HEXA, HEXB, and GM2A gene loci. Genet. Med. 2:319, 2000. 69. Cork, L. C., Altschuler, R. J., Bruha, P. J., et al.: Changes in neuronal size and neurotransmitter marker in hereditary canine spinal muscular atrophy. Lab. Invest. 61:69, 1989. 70. Costa-Jussà, F. R., and Jacobs, J. M.: The pathology of amiodarone neurotoxicity. Brain 108:735, 1985. 71. Costigan, M., Mannion, R. J., Kendall, G., et al.: Heat shock protein 27: developmental regulation and expression after peripheral nerve injury. J. Neurosci. 18:5891, 1998. 72. Cottrell, D. F., McGorum, B. C., and Pearson, G. T.: The neurology and enterology of equine grass sickness: a review of basic mechanisms. Neurogastroenterol. Motil. 11:79, 1999.

719

73. Coulson, E. J., Reid, K., Barrett, G. L., and Bartlett, P. F.: p75 neurotrophin receptor-mediated neuronal death is promoted by Bcl-2 and prevented by Bcl-xL. J. Biol. Chem. 274:16387, 1999. 74. Coutinho, P., Guimarães, A., and Scaravilli, F.: The pathology of Machado-Joseph disease: report of a possible homozygous case. Acta Neuropathol. (Berl.) 58:48, 1982. 75. Croker, A. C., and Farber, S.: Niemann-Pick disease: a review of eighteen patients. Medicine 37:1, 1958. 76. Cowan, T., Johnson, R. J., Soubeyre, V., and Santer, R. M.: Restricted diet rescues rat enteric motor neurons from age related cell death. Gut 47:653, 2000. 77. Coward, K., Jowett, A., Plumpton, C., et al.: Sodium channel beta1 and beta2 subunits parallel SNS/PN3 alphasubunit changes in injured human sensory neurons. Neuroreport 12:483, 2001. 78. Cummings, J. F., de Lahunta, A., and Winn, S. S.: Acral mutilation and nociceptive loss in English Pointer dogs. Acta Neuropathol. (Berl.) 53:119, 1981. 79. da Veiga Pereira, L., Desnick, R. J., Adler, D. A., et al.: Regional assignment of the human acid sphingomyelinase gene (SMPD1) by PCR analysis of somatic cell hybrids and in situ hybridisation to 11p15.1-p15.4. Genomics 9:229, 1991. 80. Dalmau, J., Graus, F., Rosenblum, M. K., and Posner, J. B.: Anti-Hu-associated paraneoplastic encephalomyelitis/sensory neuropathy: a clinical study of 71 patients. Medicine 71:59, 1992. 81. Dalpe, G., Leclerc, N., Vallee, A., et al.: Dystonin is essential for maintaining neuronal cytoskeleton organization. Mol. Cell. Neurosci. 10:243, 1998. 82. Daniels, R. J., Suthers, G. K., Morrison, K. E., et al.: Prenatal prediction of spinal muscular atrophy. J. Med. Genet. 29:165, 1992. 83. David, G., Abbas, N., Stevanin, G., et al.: Cloning of the SCA7 gene reveals a highly unstable CAG repeat expansion. Nat. Genet. 17:65, 1997. 84. Davies, J. P., Levy, B., and Ioannou, Y. A.: Evidence for a Niemann-Pick C (NPC) gene family: identification and characterization of NPC1L1. Genomics 65:137, 2000. 85. Dawson, V. L.: Nitric oxide: role in neurotoxicity. Clin. Exp. Pharmacol. Physiol. 22:305, 1995. 86. De Balogh, K. K., Dimande, A. P., van der Lugt, J. J., et al.: A lysosomal storage disease induced by Ipomoea carnea in goats in Mozambique. J. Vet. Diagn. Invest. 11:266, 1999. 87. De Bilbao, F., Giannakopoulos, P., Srinivasan, A., and Dubois-Dauphin, M.: In vivo study of motoneuron death induced by nerve injury in mice deficient in the caspase 1/interleukin-1 beta-converting enzyme. Neuroscience 98:573, 2000. 88. De Castro, F.: Sensory ganglia of the cranial and spinal nerves: normal and pathological. In Penfield, W. (ed.): Cytology and Cellular Pathology of the Nervous System. New York, Hoeber, p. 32, 1932. 89. Delio, D. A., Reuhl, K. R., and Lowndes, H. E.: Ectopic impulse generation in dorsal root ganglion neurons during methylmercury intoxication: an electrophysiological and morphological study. Neurotoxicology 13:527, 1992. 90. Denny-Brown, D., Adams, R. D., and Fitzgerald, P. J.: Pathologic features of herpes zoster: a note on “geniculate herpes.” AMA Arch. Neurol. Psychiatry 51:216, 1944.

720

Pathology of the Peripheral Nervous System

91. Dubois-Dalcq, M., Menu, R., and Buyse, M.: Influence of fatty acids on fine structure of cultured neurons: an experimental approach to Refsum’s disease. J. Neuropathol. Exp. Neurol. 31:645, 1972. 92. Duchen, L. W., Anjorin, S., Watkins, P. J., and McKay, J. D.: Pathology of autonomic neuropathy in diabetes. Ann. Intern. Med. 92:301, 1980. 93. Duchen, L. W., Eicher, E. M., Jacobs, J. M., et al.: Hereditary leukodystrophy in the mouse: the new mutant twitcher. Brain 103:695, 1980. 94. Duchen, L. W., and Scaravilli, F.: Quantitative and electron microscopic studies of sensory ganglion cells of the Sprawling mouse. J. Anat. 123:763, 1977. 95. Duchen, L. W., and Strich, S. J.: Clinical and pathological studies of an hereditary neuropathy in mice (dystonia musculorum). Brain 87:367, 1964. 96. Duchen, L. W., Strich, S. J., and Falconer, D. S.: An hereditary motor neurone disease with progressive denervation of muscle in the mouse: the mutant “wobbler”. J. Neurol. Neurosurg. Psychiatry 31:535, 1968. 97. Duffy, P. E., Kornfeld, M., and Suzuki, K.: Neurovisceral storage disease with curvilinear bodies. J. Neuropathol. Exp. Neurol. 27:351, 1968. 98. D’Urso, D., Ehrhardt, P., and Muller, H. W.: Peripheral myelin protein 22 and protein zero: a novel association in peripheral nervous system myelin. J. Neurosci. 19:3396, 1999. 99. Dyck, P. J., Dyck, P. J. B., and Schaid, D. J.: Genetic heterogeneity in hereditary sensory and autonomic neuropathies: the need for improved ascertainment. Muscle Nerve 23:1453, 2000. 100. Dykens, J. A.: Free radicals and mitochondria dysfunction in excitotoxicity and neurodegenerative disease. In Koliatsos, V. E., and Ratan, R. R. (eds.): Cell Death and Diseases of the Nervous System. Totowa, NJ, Humana Press, p. 45, 1999. 101. Edwards, S. N., Buckmaster, A. E., and Tolkowsky, A. M.: The death programme in cultured sympathetic neurons can be suppressed at the posttranslational level by nerve growth factor, cyclic AMP, and depolarization. J. Neurochem. 57:2140, 1991. 102. Edwards, S. N., and Tolkovsky, A. M.: Characterization of apoptosis in cultured rat sympathetic neurons after nerve growth factor withdrawal. J. Cell Biol. 124:537, 1994. 103. Eidelberg, D., Sotrel, A., Vogel, H., et al.: Progressive polyradiculopathy in acquired immune deficiency syndrome. Neurology 36:912, 1986. 104. Ekblad, E., Mulder, H., and Sundler, F.: Vasoactive intestinal peptide expression in enteric neurons is upregulated by both colchicine and axotomy. Regul. Pept. 63:113, 1996. 105. Ekström, P.: Neurones and glial cells of the mouse sciatic nerve undergo apoptosis after injury in vivo and in vitro. Neuroreport 6:1029, 1995. 106. Elleder, M., and Jirasek, A.: Niemann-Pick disease: report on a symposium held in Hlava’s Institute of Pathology, Charles University, Prague, September 1982. Acta Univ. Carol. Med. Praha 29:259, 1983. 107. Elliott, J. L., and Snider, W. D.: Axotomy-induced motor neuron death. In Koliatsos, V. E., and Ratan, R. R. (eds.): Cell Death and Diseases of the Nervous System. Totowa, NJ, Humana Press, p. 181, 1999.

108. Emery, J. L., and Singhal, R.: Changes associated with growth in the cells of the dorsal root ganglion in children. Dev. Med. Child Neurol. 15:460, 1973. 109. Eng, C. M., and Desnick, R. J.: Molecular basis of Fabry disease: mutations and polymorphisms in the human alphagalactosidase A gene. Hum. Mutat. 3:103, 1994. 110. Ernfors, P., Rosario, C. M., Merlio, J. P., et al.: Expression of mRNAs for neurotrophin receptors in the dorsal root ganglion and spinal cord during development and following peripheral or central axotomy. Mol. Brain Res. 17:217, 1993. 111. Esiri, M. M., Morris, C. S., and Millard, P. R.: Sensory and sympathetic ganglia in HIV infection: immunocytochemical demonstration of HIV-1 viral antigens, increased MHC class II antigens expression and mild reactive inflammation. J. Neurol. Sci. 114:178, 1993. 112. Estrada, R., Galaraga, J., Orozco, G., et al.: Spinocerebellar ataxia 2 (SCA2): morphometric analyses in 11 autopsies. Acta Neuropathol. (Berl.) 97:306, 1999. 113. Fadok, V. A., Voelker, D. R., Campbell, P. A., et al.: Exposure of phosphatidylserine on the surface of apoptotic lymphocytes triggers specific recognition and removal by macrophages. J. Immunol. 148:2207, 1992. 114. Falconer, D. S.: Wobbler (rat) mouse. Mouse Newslett. 15:23, 1956. 115. Fernandes, K. J., Fan, D. P., Tsui, B. J., et al.: Influence of the axotomy to cell body distance in rat rubrospinal and spinal motoneurons: differential regulation of GAP-43, tubulins, and neurofilament-M. J. Comp. Neurol. 414:495, 1999. 116. Fernandes, K. J., Kobayashi, N. R., Jasmin, B. J., and Tetzlaff, W.: Acetylcholinesterase gene expression in axotomized rat facial motoneurons is differentially regulated by neurotrophins: correlation with trkB and trkC mRNA levels and isoforms. J. Neurosci. 18:9936, 1998. 117. Fernyhough, P., Gallagher, A., and Averill, S. A.: Aberrant neurofilament phosphorylation in sensory neurons in rats with diabetic neuropathy. Diabetes 48:881, 1999. 118. Festoff, B. W., D’Andrea, M. R., Citron, B. A., et al.: Motor neuron cell death in wobbler mutant mice follows overexpression of the G-protein-coupled, protease-activated receptor for thrombin. Mol. Med. 6:410, 2000. 119. Finn, L. S., Zhang, M., Chen, S. H., and Scott, C. R.: Severe type II Gaucher disease with ichthyosis, arthrogryposis and neuronal apoptosis: molecular and pathological analyses. Am. J. Med. Genet. 91:222, 2000. 120. Fischer, S. J., McDonald, E. S., Gross, L., and Windebank, A. J.: Alterations in cell cycle regulation underlie cisplatin induced apoptosis of dorsal root ganglion neurons in vivo. Neurobiol. Dis. 8:1027, 2001. 121. Fitzgerald, M., Woolf, C. J., Gibson, S. J., and Mallaburn, P. S.: Alterations in the structure, function, and chemistry of C fibres following local application of vinblastine to the sciatic nerve of the rat. J. Neurosci. 4:430, 1984. 122. Flanigan, K. M., Crawford, T. O., Griffin, J. W., et al.: Localization of the giant axonal neuropathy gene to chromosome 16q24. Ann. Neurol. 43:143, 1998. 123. Fletcher, G. C., Xue, L., Passingham, S. K., and Tolkovsky, A. M.: Death commitment point is advanced by axotomy in sympathetic neurons. J. Cell Biol. 150:741, 2000.

Pathology of Peripheral Neuron Cell Bodies 124. Frade, J. M., Rodriguez-Tébar, A., and Barde, Y.-A.: Induction of cell death by endogenous nerve growth factor through its p75 receptor. Nature 383:166, 1996. 125. Frangione, B., Vidal, R., Rostagno, A., and Ghiso, J.: Familial cerebral amyloid angiopathies and dementia. Alzheimer Dis. Assoc. Disord. 14 Suppl. 1S:25, 2000. 126. Friede, R. L., and Johnstone, M. A.: Responses of thymidine labeling of nuclei in gray matter and nerve following sciatic transection. Acta Neuropathol. (Berl.) 7:218, 1967. 127. Friede, R. L., and Samorajski, T.: Axon caliber related to neurofilaments and microtubules in sciatic nerve fibers of rats and mice. Anat. Rec. 167:379, 1970. 128. Fukai, K., Oh, J., Karim, M. A., et al.: Homozygosity mapping of the gene for Chediak-Higashi syndrome to chromosome 1q42-q44 in a segment of conserved synteny that includes the mouse beige locus (bg). Am. J. Hum. Genet. 59:620, 1996. 129. Fuller, G. N., and Jacobs, J. M.: Peripheral nerve and muscle disease in HIV infection. In Scaravilli, F. (ed.): The Neuropathology of HIV Infection. London, Springer, p. 215, 1993. 130. Gabbiani, G., Gregory, A., and Baic, D.: Cadmium-induced selective lesions of sensory ganglia. Exp. Neurol. 26:498, 1967. 131. Gabella, G.: Structure of the Autonomic Nervous System. London, Chapman & Hall, 1976. 132. Gadoth, N., and Sandbank, U.: Involvement of dorsal root ganglia in Fabry’s disease. J. Med. Genet. 20:309, 1983. 133. Gambarelli, D., Hassoun, J., Pellissier, J. F., et al.: Giant axonal neuropathy: involvement of peripheral nerve, myenteric plexus and extra-neuronal area. Acta Neuropathol. (Berl.) 39:261, 1977. 134. Gardiner, R. M.: The molecular genetic basis of the neuronal ceroid lipofuscinoses. Neurol. Sci. 21:S15, 2000. 135. Gardner, M. B., Rashid, S., Klement, V., et al.: Lower motor neuron disease in wild mice caused by indigenous type C virus and search for a similar etiology in human amyotrophic lateral sclerosis. In Andrews, J. M., Johnson, R. T., and Brazier, M. B. (eds.): Amyotrophic Lateral Sclerosis. New York, Academic Press, p. 217, 1976. 136. Garen, P. D., Powers, J. M., Young, G. F., and Lee, V.: Neuronal intranuclear hyaline inclusion disease in a nine year old. Acta Neuropathol. (Berl.) 70:327, 1986. 137. Gavrieli, A., Sherman, Y., and Ben-Sasson, S. A.: Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. J. Cell Biol. 119:493, 1992. 138. Gilad, G. M., and Gilad, V. H.: Increased choline kinase activity in the rat superior cervical ganglion after axonal injury. Brain Res. 220:420, 1981. 139. Gilad, G. M., and Gilad, V. H.: Polyamine biosynthesis is required for survival of sympathetic neurons after axonal injury. Brain Res. 273:191, 1983. 140. Gilbert, G. J.: Herpes zoster ophthalmicus and delayed contralateral hemiparesis. JAMA 229:302, 1974. 141. Gilbert, M. R., Harding, B. L., Hoffman, P. N., et al.: Aluminum-induced neurofilamentous changes in cultured rat dorsal root ganglia explants. J. Neurosci. 12:1763, 1992. 142. Gill, J. S., Hobday, K. L., and Windebank, A. J.: Mechanism of suramin toxicity in stable myelinating dorsal root ganglion cultures. Exp. Neurol. 133:113, 1995.

721

143. Gill, J. S., and Windebank, A. J.: Cisplatin-induced apoptosis in rat dorsal root ganglion neurons is associated with attempted entry into the cell cycle. J. Clin. Invest. 101:2842, 1998. 144. Gillardon, F., Klimaschewski, L., Wickert, H., et al.: Expression pattern of candidate cell death effector proteins Bax, Bcl-2, Bcl-X, and c-Jun in sensory and motor neurons following sciatic nerve transection in the rat. Brain Res. 739:244, 1996. 145. Gilles, G. H., and Deuel, R. K.: Neuronal cytoplasmic globules in the brain in Morquio’s syndrome. Arch. Neurol. 25:393, 1971. 146. Giometto, B., Scaravilli, T., Nicolao, P., et al.: Detection of paraneoplastic anti-neuronal autoantibodies on paraffinembedded tissues. Acta Neuropathol. (Berl.) 92:435, 1996. 147. Glenn, D., Gelbart, T., and Beutler, E.: Tight linkage of pyruvate kinase (PKLR) and glucocerebrosidase (GBA) genes. Hum. Genet. 93:635, 1994. 148. Gloster, A., Wu, W., Speelman, A., et al.: The T alpha 1 alpha-tubulin promoter specifies gene expression as a function of neuronal growth and regeneration in transgenic mice. J. Neurosci. 14:7319, 1994. 149. Glynn, P., Holton, J. L., Nolan, C. C., et al.: Neuropathy target esterase: immunolocalization to neuronal cell bodies and axons. Neuroscience 83:295, 1998. 150. Goebel, H. H., Kohlschutter, A., and Lenard, H. G.: Morphologic and chemical biopsy findings in mucolipidosis IV. Clin. Neuropathol. 1:73, 1982. 151. Gold, M. S., Reichling, D. B., Hampl, K. F., et al.: Lidocaine toxicity in primary afferent neurons from the rat. J. Pharmacol. Exp. Ther. 285:413, 1998. 152. Goldstein, M. E., Cooper, H. S., Bruce, J., et al.: Phosphorylation of neurofilament proteins and chromatolysis following transection of rat sciatic nerve. J. Neurosci. 7:1586, 1987. 153. Gotz, J., Tolnay, M., Barmettler, R., et al.: Oligodendroglial tau filament formation in transgenic mice expressing G272V tau. Eur. J. Neurosci. 13:2131, 2001. 154. Grafstein, B.: The nerve cell body response to axotomy. Exp. Neurol. 48:32, 1975. 155. Green, S. L., Vulliet, P. R., Pinter, M. J., and Cork, L. C.: Alterations in cyclin-dependent protein kinase 5 (CDK5) protein levels, activity and immunocytochemistry in canine motor neuron disease. J. Neuropathol. Exp. Neurol. 57:1070, 1998. 156. Greenbaum, D., Richardson, P. C., Salmon, M. V., and Urich, H.: Pathological observations on six cases of diabetic neuropathy. Brain 87:201, 1964. 157. Greenberg, S. G., and Lasek, R. J.: Neurofilament protein synthesis in DRG neurons decreases more after peripheral axotomy than after central axotomy. J. Neurosci. 8:1739, 1988. 158. Greensmith, L., Hasan, H. I., and Vrbova, G.: Nerve injury increases the susceptibility of motoneurons to N-methyl-Daspartate-induced neurotoxicity in the developing rat. Neuroscience 58:727, 1994. 159. Greer, W. L., Riddell, D. C., Byers, D. M., et al.: Linkage of Niemann-Pick disease type D to the same region of human chromosome 18 as Niemann-Pick disease type C. Am. J. Hum. Genet. 61:139, 1997.

722

Pathology of the Peripheral Nervous System

160. Griffin, J. W., Cornblath, D. R., Alexander, E., et al.: Ataxic sensory neuropathy and dorsal root ganglionitis associated with Sjögren’s syndrome. Ann. Neurol. 27:304, 1990. 161. Griffin, J. W., Li, C. Y., Ho, T. W., et al.: Pathology of the motor-sensory axonal Guillain-Barré syndrome. Ann. Neurol. 39:17, 1996. 162. Groves, M. J., An, S.-F., Giometto, B., and Scaravilli, F.: Inhibition of sensory neuron apoptosis and prevention of loss by NT-3 administration following axotomy. Exp. Neurol. 155:284, 1999. 163. Groves, M. J., Christopherson, T., Giometto, B., and Scaravilli, F.: Axotomy-induced apoptosis in adult rat sensory neurons. J. Neurocytol. 26:615, 1997. 164. Groves, M. J., Giometto, B. and Scaravilli, F.: Axotomyinduced vacuolation of primary sensory neurons and effect of administered neurotrophic factors: a morphometric, immunocytochemical and ultrastructural study. Prim. Sensory Neuron 2:111, 1997. 165. Groves, M. J., Ng, Y.-W., Ciardi, A., and Scaravilli, F.: Sciatic nerve injury in the adult rat: comparison of effects on oligosaccharide, CGRP, and GAP43 immunoreactivity in primary afferents following two types of trauma. J. Neurocytol. 26:219, 1996. 166. Guegan, C., Vila, M., Rosoklija, G., et al.: Recruitment of the mitochondrial-dependent apoptotic pathway in amyotrophic lateral sclerosis. J. Neurosci. 21:6569, 2001. 167. Gunasekar, P. G., Kanthasamy, A. G., Borowitz, J. L., and Isom, G. E.: NMDA receptor activation produces concurrent generation of nitric oxide and reactive oxygen species: implication for cell death. J. Neurochem. 65:2016, 1995. 168. Guntinas-Lichius, O., Schulte, E., Stennert, E., and Neiss, W. F.: The use of texture analysis to study the time course of chromatolysis. J. Neurosci. Methods 78:1, 1997. 169. Guo, A., Vulchanova, L., Wang, J., et al.: Immunocytochemical localization of the vanilloid receptor 1 (VR1): relationship to neuropeptides, the P2X3 purinoceptor and IB4 binding sites. Eur. J. Neurosci. 11:946, 1999. 170. Haas, C. A., Hofmann, H. D., and Kirsch, M.: Expression of CNTF/LIF-receptor components and activation of STAT3 signaling in axotomized facial motoneurons: evidence for a sequential postlesional function of the cytokines. J. Neurobiol. 41:559, 1999. 171. Hafer-Macko, C., Hsieh, S.-T., Li, C. Y., et al.: Acute motor axonal neuropathy: an antibody-mediated attack on axolemma. Ann. Neurol. 40:635, 1996. 172. Hagberg, B., Haltia, M., Sourander, P., et al.: Neurovisceral storage disorder simulating Niemann-Pick disease: a new form of oligosaccharidosis? Neuropädiatrie 9:59, 1978. 173. Hahn, C. N., del Pilar Martin, M., Schröder, M., et al.: Generalised CNS disease and massive GM1-ganglioside accumulation in mice defective in lysosomal acid ␤gangliosidase. Hum. Mol. Genet. 6:205, 1997. 174. Hahn, C. N., Mayhew, I. G., and de Lahunta, A.: Central neuropathology of equine grass sickness. Acta Neuropathol. (Berl.) 102:153, 2001. 175. Hammarberg, H., Piehl, F., Cullheim, S., et al.: GDNF mRNA in Schwann cells and DRG satellite cells after chronic sciatic nerve injury. Neuroreport 7:857, 1996. 176. Hammarberg, H., Piehl, F., Risling, M., and Cullheim, S.: Differential regulation of trophic factor receptor mRNAs in

177.

178.

179.

180.

181.

182. 183.

184.

185.

186.

187.

188. 189.

190.

191. 192.

spinal motoneurons after sciatic nerve transection and ventral root avulsion in the rat. J. Comp. Neurol. 426:587, 2000. Harvey, J. S., Davis, C., James, I. F., and Burgess, G. M.: Activation of protein kinase C by the capsaicin analogue resiniferatoxin in sensory neurones. J. Neurochem. 65:1309, 1995. Haust, M. D., and Gordon, B. A.: Ultrastructural and biochemical aspects of the Sanfilippo syndrome—type III genetic mucopolysaccharidosis. Connect. Tissue Res. 15:57, 1986. Hayashi, T., Sakurai, M., Abe, K., et al.: Apoptosis of motor neurons with induction of caspases in the spinal cord after ischemia. Stroke 29:1007, 1998. Helliwell, R. J., McLatchie, L. M., Clarke, M., et al.: Capsaicin sensitivity is associated with the expression of the vanilloid (capsaicin) receptor (VR1) mRNA in adult rat sensory ganglia. Neurosci. Lett. 258:77, 1998. Hendry, I. A., and Campbell, J.: Morphometric analysis of rat superior cervical ganglion after axotomy and nerve growth factor treatment. J. Neurocytol. 5:351, 1976. Henson, R. A., and Urich, H.: Cancer and the Nervous System. London, Blackwell Scientific, 1982. Herdegen, T., Fiallos-Estrada, C. E., Schmid, W., et al.: The transcription factors c-JUN, JUN D and CREB, but not FOS and KROX-24, are differentially regulated in axotomized neurons following transection of the rat sciatic nerve. Brain Res. Mol. Brain Res. 14:155, 1992. Herdegen, T., and Leah, J. D.: Inducible and constitutive transcription factors in the mammalian nervous system: control of gene expression by Jun, Fos, and Krox, and KREB/ATF proteins. Brain Res. Brain Res. Rev. 28:370, 1998. Hermans, M. M., de Graaff, E., Kroos, M. A., et al.: Identification of a point mutation in the human lysosomal alpha-glucosidase gene causing infantile glycogenosis type II. Biochem. Biophys. Res. Commun. 179:919, 1991. Hermans, M. M., Kroos, M. A., de Graaff, E., et al.: Two mutations affecting the transport and maturation of lysosomal alpha-glucosidase in an adult case of glycogen storage disease type II. Hum. Mutat. 2:268, 1993. Herrick, M. K., Twiss, J. L., Vladutiu, G. D., et al.: Concomitant branching enzyme and phosphorylase deficiencies: an unusual glycogenosis with extensive neuronal polyglucosan storage. J. Neuropathol. Exp. Neurol. 53:239, 1994. Hierons, R.: Changes in the nervous system in acute porphyria. Brain 80:176, 1957. Himes, B. T., and Tessler, A.: Death of some dorsal root ganglion neurons and plasticity of others following sciatic nerve section in adult and neonatal rats. J. Comp. Neurol. 284:215, 1989. Hirosawa, K.: Electron microscopic studies on pigment granules in the substantia nigra and locus coeruleus in Japanese monkey (Macaca fuscate yakui). Z. Zellforsch. Mikrosk. Anat. 88:187, 1968. Hiura, A.: Neuroanatomical effects of capsaicin on the primary afferent neurons. Arch. Histol. Cytol. 63:199, 2000. Hiura, A., and Ishizuka, H.: Changes in features of degenerating primary sensory neurons with time after capsaicin treatment. Acta Neuropathol. (Berl.) 78:35, 1989.

Pathology of Peripheral Neuron Cell Bodies 193. Hoffman, P. N., Cleveland, D. W., Griffin, J. W., et al.: Neurofilament gene expression: a major determinant of axonal caliber. Proc. Natl. Acad. Sci. U. S. A. 84:3472, 1987. 194. Hol, E. M., Schwaiger, F. W., Werner, A., et al.: Regulation of the LIM-type homeobox gene islet-1 during neuronal regeneration. Neuroscience 88:917, 1999. 195. Holmberg, M., Duyckaerts, C., Dürr, A., et al.: Spinocerebellar ataxia type 7 (SCA7): a neurodegenerative disorder with neuronal intranuclear inclusions. Hum. Mol. Genet. 7:913, 1998. 196. Holmes, F. E., and Haynes, L. W.: Superactivation of transglutaminase type 2 without change in enzyme level occurs during progressive neurodegeneration in the mnd mouse mutant. Neurosci. Lett. 213:185, 1996. 197. Horwich, M. S., Cho, L., Porro, R. S., and Posner, J. B.: Subacute sensory neuropathy: a remote effect of carcinoma. Ann. Neurol. 2:7, 1977. 198. Hou, X. E., Lundmark, K., and Dahlstrom, A. B.: Cellular reactions to axotomy in rat superior cervical ganglia include apoptotic cell death. J. Neurocytol. 27:441, 1998. 199. Houenou, L. J., Li, L., Lo, A. C., et al.: Naturally occurring and axotomy-induced motoneuron death and its prevention by neurotrophic agents: a comparison between chick and mouse. Prog. Brain Res. 102:217, 1994. 200. Houlden, H., King, R. H., Hashemi-Nejad, A., et al.: A novel TRK A (NTRK1) mutation associated with hereditary sensory and autonomic neuropathy type V. Ann. Neurol. 49:521, 2001. 201. Hozumi, I., Nishizawa, M., Ariga, T., et al.: Accumulation of glycosphingolipids in spinal and sympathetic ganglia of a symptomatic heterozygote of Fabry’s disease. J. Neurol. Sci. 90:273, 1989. 202. Hu, P., and McLachlan, E. M.: Macrophage and lymphocyte invasion of dorsal root ganglia after peripheral nerve lesions in the rat. Neuroscience 112:23, 2002. 203. Huang, J. Q., Trasler, J. M., Igdoura, S., et al.: Apoptotic cell death in mouse models of GM2 gangliosidosis and observations on human Tay-Sachs and Sandhoff diseases. Hum. Mol. Genet. 6:1879, 1997. 204. Hughes, J. T., Brownell, B., and Hewer, R. L.: The peripheral sensory pathway in Friedreich’s ataxia: an examination by light and electron microscopy of the posterior nerve roots, posterior root ganglia and peripheral sensory nerves in cases of Friedreich’s ataxia. Brain 91:803, 1968. 205. Huxtable, C. R., and Dorling, P. R.: Mannoside storage and axonal dystrophy in sensory neurones of swainsoninetreated rats: morphogenesis of lesions. Acta Neuropathol. (Berl.) 68:65, 1985. 206. Inaishi, Y., Kashihara, Y., Sakaguchi, M., et al.: Cooperative regulation of calcitonin gene-related peptide levels in rat sensory neurons via their central and peripheral processes. J. Neurosci. 12:518, 1992. 207. Indo, Y., Tsuruta, M., Hayashida, Y., et al.: Mutations in the TRKA/NGF receptor gene in patients with congenital insensitivity to pain with anhidrosis. Nat. Genet. 13:485, 1996. 208. Iwahashi, Y., Furuyama, T., Inagaki, S., et al.: Distinct regulation of sodium channel types I, II and III following nerve transection. Brain Res. Mol. Brain Res. 22:341, 1994. 209. Jaatinen, P., and Hervonen, A.: Reactions of rat sympathetic neurons to ethanol exposure are age-dependent. Neurobiol. Aging 15:419, 1994.

723

210. Jaatinen, P., Riihioja, P., Haapalinna, A., et al.: Prevention of ethanol-induced sympathetic overactivity and degeneration by dexmedetomidine. Alcohol 12:439, 1995. 211. Jacob, J. M.: Lumbar motor neuron size and number is affected by age in male F344 rats. Mech. Ageing Dev. 106:205, 1998. 212. Jacobs, J. M., Carmichael, N., and Cavanagh, J. B.: Ultrastructural changes in the dorsal root and trigeminal ganglia of rats poisoned with methyl mercury. Neuropathol. Appl. Neurobiol. 1:1, 1975. 213. Jacobs, J. M., Carmichael, N., and Cavanagh, J. B.: Ultrastructural changes in the nervous system of rabbits poisoned with methyl mercury. Toxicol. Appl. Pharmacol. 39:249, 1977. 214. Jacobs, J. M., Scaravilli, F., Duchen, L. W., and Mertin, J.: A new neurological rat mutant “mutilated foot.” J. Anat. 132:525, 1981. 215. Jarvela, I., Vesa, J., Santavuori, P., et al.: Molecular genetics of neuronal ceroid lipofuscinoses. Pediatr. Res. 32:645, 1992. 216. Jellinger, K., Paulus, W., Grisold, W., and Paschke, E.: New phenotype of adult alpha-L-iduronidase deficiency (mucopolysaccharidosis I) masquerading as Friedreich’s ataxia with cardiopathy. Clin. Neuropathol. 9:163, 1990. 217. Jessell, T. M., Tsunoo, A., Kanazawa, I., and Otsuka, M.: Substance P: depletion in the dorsal horn of rat spinal cord after section of the peripheral processes of primary sensory neurons. Brain Res. 168:247, 1979. 218. Ji, R.-R., Zhang, Q., Zhang, X., et al.: Prominent expression of bFGF in dorsal root ganglia after axotomy. Eur. J. Neurosci. 7:2458, 1995. 219. Johnson, P. C., Roluk, L. A., Hamilton, R. H., and Laguna, J. F.: Paraneoplastic vasculitis of nerve: a remote effect of cancer. Ann. Neurol. 5:437, 1979. 220. Jones, H. B., and Cavanagh, J. B.: Cytoplasmic vacuoles. Ultrastruct. Pathol. 6:359, 1984. 221. Jones, K. J., and Lavelle, A.: Differential effects of axotomy on immature and mature hamster facial neurons: a time course study of initial nucleolar and nuclear changes. J. Neurocytol. 15:197, 1986. 222. Kami, K., Morikawa, Y., Kawai, Y., and Senba, E.: Leukemia inhibitory factor, glial cell line-derived neurotrophic factor, and their receptor expressions following muscle crush injury. Muscle Nerve 22:1576, 1999. 223. Karanth, S. S., Springall, D. R., Kar, S., et al.: Time-related decrease of substance P and CGRP in central and peripheral projections of sensory neurons in Mycobacterium leprae infected nude mice: a model for lepromatous leprosy in man. J. Pathol. 160:335, 1990. 224. Kashiba, H., Hyon, B., and Senba, E.: Glial cell line-derived neurotrophic factor and nerve growth factor mRNAs are expressed in distinct subgroups of dorsal root ganglion neurons and are differentially regulated by peripheral axotomy in the rat. Neurosci. Lett. 252:107, 1998. 225. Kashiba, H., Senba, E., Yoshihiro, U., and Tohyama, M.: Co-localized but target-unrelated expression of vasoactive intestinal polypeptide and galanin in rat dorsal root ganglion neurons after peripheral nerve crush injury. Brain Res. 582:47, 1992. 226. Kashihara, Y., Kuno, M., and Miyata, Y.: Cell death of axotomized motoneurons in neonatal rats, and its prevention

724

227.

228.

229.

230.

231.

232.

233.

234.

235.

236.

237.

238. 239.

240.

241. 242. 243.

244.

Pathology of the Peripheral Nervous System by peripheral reinnervation. J. Physiol. (Lond.) 386:135, 1987. Kato, T., Hirano, A., Weinberg, M. N., and Jacobs, A. K.: Spinal cord lesions in progressive supranuclear palsy: some new observations. Acta Neuropathol. (Berl.) 71:11, 1986. Katsetos, C. D., Savory, J., Herman, M. M., et al.: Neuronal cytoskeletal lesions induced in the CNS by intraventricular and intravenous aluminium maltol in rabbits. Neuropathol. Appl. Neurobiol. 16:511, 1990. Kawasaki, H., Murayama, S., Tomonaga, M., et al.: Neurofibrillary tangles in human upper cervical ganglia: morphological study with immunohistochemistry and electron microscopy. Acta Neuropathol. (Berl.) 75:156, 1987. Kennedy, P. G., Grinfeld, E., and Gow, J. W.: Latent varicella-zoster virus is located predominantly in neurons in human trigeminal ganglia. Proc. Natl. Acad. Sci. U. S. A. 95:4658, 1998. Kerenyi, T., Havanhyl, L., and Huttner, I.: Investigations on experimentally produced age-pigment in the nervous system. Exp. Gerontol. 3:155, 1968. Kerezoudi, E., King, R. H. M., Muddle, J. R., et al.: Influence of age on the retrograde effects of sciatic nerve section in the rat. J. Anat. 187:27, 1995. Kerezoudi, E., and Thomas, P. K.: Influence of age on regeneration in the peripheral nervous system. Gerontology 45:301, 1999. Kerr, J. F. R., Wyllie, A. H., and Currie, A. R.: Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. Br. J. Cancer 26:239, 1972. Kikuchi, H., Doh-ura, K., Kira, J., and Iwaki, T.: Preferential neurodegeneration in the cervical spinal cord of progressive supranuclear palsy. Acta Neuropathol. (Berl.) 97:577, 1999. Kim, D. S., Yoon, C. H., Lee, S. J., et al.: Changes in voltage-gated calcium channel alpha(1) gene expression in rat dorsal root ganglia following peripheral nerve injury. Brain Res. Mol. Brain Res. 96:151, 2001. Kimber, T. E., Blumbergs, P. C., Rice, J. P., et al.: Familial neuronal intranuclear inclusion disease with ubiquitin positive inclusions. J. Neurol. Sci. 160:33, 1998. Kint, J. A.: Fabry’s disease: ␣-galactosidase deficiency. Science 167:1268, 1970. Klimaschewski, L., Obermuller, N., Majewski, M., et al.: Increased expression of nitric oxide synthase in a subpopulation of rat sympathetic neurons after axotomy—correlation with vasoactive intestinal peptide. Cell Tissue Res. 285:419, 1996. Kobayashi, N. R., Bedard, A. M., Hincke, M. T., and Tetzlaff, W.: Increased expression of BDNF and trkB mRNA in rat facial motoneurons after axotomy. Eur. J. Neurosci. 8:1018, 1996. Koliatsos, V. E., and Ratan, R. R.: Cell Death and Diseases of the Nervous System. Totowa, NJ, Humana Press, 1999. Koltzenburg, M., Wall, P. D., and McMahon, S. B.: Does the right side know what the left is doing? TINS 22:122, 1999. Koshy, B. T., and Zoghbi, H. Y.: The CAG/polyglutamine tract diseases: gene products and molecular pathogenesis. Brain Pathol. 7:927, 1997. Koto, A., Horwitz, A. L., Suzuki, K., and Tiffany, C. W.: The Morquio syndrome: neuropathology and biochemistry. Ann. Neurol. 4:26, 1978.

245. Koyano, S., Uchihara, T., Fujigasaki, H., et al.: Neuronal intranuclear inclusions in spinocerebellar ataxia type 2: triple labelling immunofluorescent study. Neurosci. Lett. 273:117, 1999. 246. Krarup-Hansen, A., Rietz, B., Krarup, C., et al.: Histology and platinum content of sensory ganglia and sural nerves in patients treated with cisplatin and carboplatin: an autopsy study. Neuropathol. Appl. Neurobiol. 25:29, 1999. 247. Krebs, C., Neiss, W. F., Streppel, M., et al.: Axotomy induces transient calbindin D28K immunoreactivity in hypoglossal motoneurons in vivo. Cell Calcium 22:367, 1997. 248. Kristensson, K., and Olsson, Y.: Retrograde transport of horseradish peroxidase in transected axons. 1. Time relationships between transport and induction of chromatolysis. Brain Res. 79:101, 1974. 249. Kroemer, G., Zamzami, N., and Susin, S. A.: Mitochondrial control of apoptosis. Immunol. Today 18:44, 1997. 250. Kuemmel, T. A., Schroeder, R., and Stoffel, W.: Light and electron microscopic analysis of the central and peripheral nervous systems of acid sphingomyelinase-deficient mice resulting from gene targeting. J. Neuropathol. Exp. Neurol. 56:171, 1997. 251. La Spada, A. R., Paulson, H. L., and Fishbeck, K. H.: Trinucleotide repeat expansion in neurological disease. Ann. Neurol. 36:814, 1994. 252. La Spada, A. R., Wilson, E. M., Lubahn, D. B., et al.: Androgen receptor gene mutations in X-linked spinal and bulbar muscular atrophy. Nature 352:77, 1991. 253. Lake, B.: Lysosomal and peroxisomal disorders. In Graham, D. I., and Lantos, P. L. (eds.): Greenfield’s Neuropathology. London, Arnold, p. 657, 1997. 254. LaMotte, C. C., and Kapadia, S. E.: Deafferentationinduced alterations in the rat dorsal horn: II. Effects of selective poisoning by pronase of the central processes of a peripheral nerve. J. Comp. Neurol. 266:198, 1987. 255. Lawson, S. J., and Lowrie, M. B.: The role of apoptosis and excitotoxicity in the death of spinal motoneurons and interneurons after neonatal nerve injury. Neuroscience 87:337, 1998. 256. Lawson, S. N.: Morphological and biochemical types of sensory neurons. In Scott, S. A. (ed.): Sensory Neurons. Oxford, UK, Oxford University Press, p. 27, 1992. 257. Lee, M., Hyun, D. H., Halliwell, B., and Jenner, P.: Effect of overexpression of wild-type and mutant Cu/Zn-superoxide dismutases on oxidative stress and cell death induced by hydrogen peroxide, 4-hydroxynonenal or serum deprivation: potentiation of injury by ALS-related mutant superoxide dismutases and protection by Bcl-2. J. Neurochem. 78:209, 2001. 258. Lee, P., Zhuo, H., and Helke, C. J.: Axotomy alters neurotrophin and neurotrophin receptor mRNAs in the vagus nerve and nodose ganglion of the rat. Brain Res. Mol. Brain Res. 87:31, 2001. 259. Lewis, S. E., Mannion, R. J., White, F. A., et al.: A role for HSP27 in sensory neuron survival. J. Neurosci. 19:8945, 1999. 260. Lhermitte, F., Gray, F., Lyon-Caen, O., et al.: Paralysie du tube digestif avec lesions des plexus myentériques: nouveau syndrome paranéoplasique. Rev. Neurol. (Paris) 136:825, 1980.

Pathology of Peripheral Neuron Cell Bodies 261. Li, L., Houenou, L. J., Wu, W., et al.: Characterization of spinal motoneuron degeneration following different types of peripheral nerve injury in neonatal and adult mice. J. Comp. Neurol. 396:158, 1998. 262. Liberski, P. P., Buczynski, J., Yanagihara, R., et al.: Ultrastructural pathology of a Chilean case of tropical spastic paraparesis/human T-cell lymphotropic type 1-associated myelopathy (TSP/HAM). Ultrastruct. Pathol. 23:157, 1999. 263. Lieberman, A. R.: The axon reaction: a review of the principal features of perikaryal responses to axon injury. Int. Rev. Neurobiol. 14:49, 1971. 264. Lieberman, A. R.: Some factors affecting retrograde neuronal responses to axonal lesions. In Bellairs, R., and Gray, E. G. (eds.): Essays on the Nervous System. Oxford, UK, Clarendon Press, p. 71, 1974. 265. Lieberman, A. R.: Sensory ganglia. In Landon, D. N. (ed.): The Peripheral Nerve. London, Chapman and Hall, p. 188, 1976. 266. Liu, R. Y., and Snider, W. D.: Different signaling pathways mediate regenerative versus developmental sensory axon growth. J. Neurosci. 21:RC164, 2001. 267. Liu, Z., and Martin, L. J.: Motor neurons rapidly accumulate DNA single-strand breaks after in vitro exposure to nitric oxide and peroxynitrite and in vivo axotomy. J. Comp. Neurol. 432:35, 2001. 268. Ljungberg, C., Novikov, L., Kellerth, J. O., et al.: The neurotrophins NGF and NT-3 reduce sensory neuronal loss in adult rat after peripheral nerve lesion. Neurosci. Lett. 262:29, 1999. 269. Low, P. A.: Hereditary hypertrophic neuropathy in the trembler mouse. Part 1. Histopathological studies: light microscopy. J. Neurol. Sci. 30:327, 1976. 270. Lu, X., and Richardson, P. M.: Inflammation near the nerve cell body enhances axonal regeneration. J. Neurosci. 11:972, 1991. 271. Lungu, O., Annunziato, P. W., Gershon, A., et al.: Reactivated and latent varicella-zoster virus in human dorsal root ganglia. Proc. Natl. Acad. Sci. U. S. A. 92:10980, 1995. 272. Luo, X. G., Zhou, X. F., and Rush, R. A.: Ultrastructural changes of sympathetic neurons following neurotrophin-3 antiserum treatment in young rat. Exp. Neurol. 147:401, 1997. 273. Lurie, I. W., Supowitz, K. R., Rosenblum-Vos, L. S., and Wulfsberg, E. A.: Phenotypic variability of del (2) (q22-q23): report of a case with a review of the literature. Genet. Couns. 5:11, 1994. 274. Mahalingam, R., Wellish, M., Cohrs, R., et al.: Expression of protein encoded by varicella-zoster virus open reading frame 63 in latently infected human ganglionic neurons. Proc. Natl. Acad. Sci. U. S. A. 93:2122, 1996. 275. Majno, G., and Joris, I.: Apoptosis, oncosis, and necrosis: an overview of cell death. Am. J. Pathol. 146:3, 1995. 276. Mancall, E. L., Aponte, G. E., and Berry, R. G.: Pompe’s disease (diffuse glycogenosis) with neuronal storage. J. Neuropathol. Exp. Neurol. 24:85, 1965. 277. Mariotti, R., Peng, Z.-C., Kristensson, K., and Bentivoglio, M.: Age-dependent induction of nitric oxide synthase activity in facial motoneurons after axotomy. Exp. Neurol. 145:361, 1997.

725

278. Martin, D. P., Schmidt, R. E., DiStefano, P. S., et al.: Inhibitors of protein synthesis and RNA synthesis prevent neuronal death caused by nerve growth factor deprivation. J. Cell Biol. 106:829, 1988. 279. Martin, J. J., Leroy, J. G., van Eygen, M., and Ceuterick, C.: I-cell disease: a further report on its pathology. Acta Neuropathol. (Berl.) 64:234, 1984. 280. Martin, J. J., Van Regemorter, N., Krols, L., et al.: On an autosomal dominant form of retinal-cerebellar degeneration: an autopsy study of five patients in one family. Acta Neuropathol. (Berl.) 88:277, 1994. 281. Martin, L. J.: Neuronal death in amyotrophic lateral sclerosis is apoptosis: possible contribution of a programmed cell death mechanism. J. Neuropathol. Exp. Neurol. 58:459, 1999. 282. Martin, L. J., Al-Abdulla, N. A., Brambrink, A. M., et al.: Neurodegeneration in excitotoxicity, global cerebral ischemia, and target deprivation: a perspective on the contributions of apoptosis and necrosis. Brain Res. Bull. 46:281, 1998. 283. Martin, L. J., Kaiser, A., and Price, A. C.: Motor neuron degeneration after sciatic nerve avulsion in adult rat evolves with oxidative stress and is apoptosis. J. Neurobiol. 40:185, 1999. 284. Martin, L. J., and Liu, Z.: Injury-induced spinal motor neuron apoptosis is preceded by DNA single-strand breaks and is p53- and Bax-dependent. J. Neurobiol. 50:181, 2002. 285. Martin, L. J., Price, A. C., Kaiser, A., et al.: Mechanisms for neuronal degeneration in amyotrophic lateral sclerosis and in models of motor neuron death [review]. Int. J. Mol. Med. 5:3, 2000. 286. Matsuda, J., Suzuki, O., Oshima, A., et al.: Neurological manifestations of knockout mice with b-galactosidase deficiency. Brain Dev. 19:19, 1997. 287. Matsuishi, T., Teresawa, K., Yoshida, I., et al.: Vacuolar myopathy with type 2 fibre atrophy and type 2B fibre deficiency: a case of childhood form of acid ␣-1,4-glucosidase deficiency. Neuropediatrics 13:173, 1982. 288. Mazliah, J., Barron, S., Bental, E., et al.: The effects of longterm lead intoxication on the nervous system of the chicken. Neurosci. Lett. 101:253, 1989. 289. Mazurkiewicz, J. E., Callahan, L. M., Swash, M., et al.: Cytoplasmic inclusions in spinal neurons of the motor neuron degeneration (mnd) mouse. I. Light microscopic analysis. J. Neurol. Sci. 116:59, 1993. 290. McKay-Hart, A., Brannstrom, T., Wiberg, M., and Terenghi, G.: Primary sensory neurons and satellite cells after peripheral axotomy in the adult rat. Exp. Brain Res. 142:308, 2002. 291. Meller, K.: Chromatolysis of dorsal root ganglion cells studied by cryofixation. Cell Tissue Res. 256:283, 1989. 292. Mentis, G. Z., Greensmith, L., and Vrbova, G.: Motoneurons destined to die are rescued by blocking N-methyl-D-aspartate receptors by MK-801. Neuroscience 54:283, 1993. 293. Messer, A., and Flaherty, L.: Autosomal dominance in a lateonset motor neuron disease in the mouse. J. Neurogenet. 3:345, 1986. 294. Messer, A., and Plummer, J.: Accumulating autofluorescent material as a marker for early changes in the spinal cord of the mnd mouse. Neuromuscul. Disord. 3:129, 1993. 295. Michael, G. J., Averill, S., Shortland, P. J., et al.: Axotomy results in major changes in BDNF expression by dorsal root

726

296.

297.

298.

299.

300.

301. 302.

303.

304.

305.

306.

307.

308.

309.

310.

Pathology of the Peripheral Nervous System ganglion cells: BDNF expression in large trkB and trkC cells, in pericellular baskets, and in projections to deep dorsal horn and dorsal column nuclei. Eur. J. Neurosci. 11:3539, 1999. Michaud, J., and Gilbert, J. J.: Multiple system atrophy with neuronal intranuclear hyaline inclusions: report of a new case with light and electron microscopic studies. Acta Neuropathol. (Berl.) 54:113, 1981. Misu, K., Hattori, N., Nagamatsu, M., et al.: Late onset familial amyloid polyneuropathy type 1 (transthyretin Met 30-associated familial amyloid polyneuropathy) unrelated to endemic focus in Japan: clinicopathological and genetic features. Brain 122:1951, 1999. Miyagishi, T., Takahata, N., and Izuka, R.: Electron microscopic studies on the lipopigments in the cerebral cortex nerve cells of senile and vitamin E deficient rats. Acta Neuropathol. (Berl.) 9:7, 1967. Mohiuddin, L., Delcroix, J. D., Fernyhough, P., and Tomlinson, D. R.: Focally administered nerve growth factor suppresses molecular regenerative responses of axotomized peripheral afferents in rats. Neuroscience 91:265, 1999. Mohseni, S.: Hypoglycaemic neuropathy in diabetic BB/Wor rats treated with insulin implants affected ventral root axons but not dorsal root axons. Acta Neuropathol. (Berl.) 100: 415, 2000. Mole, S., and Gardiner, M.: Molecular genetics of the neuronal ceroid lipofuscinoses. Epilepsia 40:S329, 1999. Monney, L., Otter, I., Olivier, R., et al.: Defects in the ubiquitin pathway induce caspase-independent apoptosis blocked by Bcl-2. J. Biol. Chem. 273:6121, 1998. Montgomery, M. M., Wood, A., Stott, E. J., et al.: Changes in neuron size in cynomolgus macaques infected with various immunodeficiency viruses and poliovirus. Neuropathol. Appl. Neurobiol. 24:468, 1998. Montpetit, V. J., Clapin, D. F., Tryphonas, L., and Dancea, S.: Alteration of neuronal cytoskeletal organization in dorsal root ganglia associated with pyridoxine neurotoxicity. Acta Neuropathol. (Berl.) 76:71, 1988. Moran, L. B., Kosel, S., Spitzer, C., et al.: Expression of alpha-synuclein in non-apoptotic, slowly degenerating facial motoneurones. J. Neurocytol. 30:515, 2001. Moretto, G., Monaco, S., Passarin, M. G., et al.: Cytoskeletal changes induced by 2,5-hexanedione on developing human neurons in vitro. Arch. Toxicol. 65:409, 1991. Morgello, S., Cho, E.-S., Nielsen, S., et al.: Cytomegalovirus encephalitis in patients with acquired immunodeficiency syndrome: an autopsy study of 30 cases and a review of the literature. Hum. Pathol. 18:289, 1987. Mossman, J., Blunt, S., Stephens, R., et al.: Hunter’s disease in a girl: association with X:5 chromosomal translocation disrupting the Hunter gene. Arch. Dis. Child. 58:911, 1983. Mowat, D. R., Croaker, G. D., Cass, D. T., et al.: Hirschsprung disease, microcephaly, mental retardation and characteristic facial features: delineation of a new syndrome and identification of a locus at chromosome 2q22-q23. J. Med. Genet. 35:617, 1998. Müller, H. D., Mugler, M., Ramaekers, V. T., and Schröder, J. M.: Hereditary sensory and motor neuropathy with absence of large myelinated fibers due to absence of large neurons in dorsal root ganglia and anterior horns, clinically associated

311.

312.

313.

314.

315.

316.

317.

318.

319.

320. 321.

322.

323.

324.

325.

326.

327.

with deafness, mental retardation, and epilepsy (HMSNADM). J. Peripher. Nerv. Syst. 5:147, 2000. Muma, N. A., and Cork, L. C.: Alterations in neurofilament mRNA in hereditary canine spinal muscular atrophy. Lab. Invest. 69:436, 1993. Muma, N. A., Hoffman, P. N., Slunt, H. H., et al.: Alterations in levels of mRNAs coding for neurofilament protein subunits during regeneration. Exp. Neurol. 107:230, 1990. Murashov, A. K., Haq, I. U., Hill, C., et al.: Crosstalk between p38, HSP25 and Akt in spinal motor neurons after sciatic nerve injury. Mol. Brain Res. 93:199, 2001. Musaka, T., Urase, K., Momoi, Y. M., et al.: Specific expression of CPP32 in sensory neurons of mouse embryos and activation of CPP32 in the apoptosis induced by a withdrawal of NGF. Biochem. Biophys. Res. Commun. 231:770, 1997. Nacimiento, W., Schlozer, B., Brook, G. A., et al.: Transient decrease of acetylcholinesterase in ventral horn neurons caudal to a low thoracic spinal cord hemisection in the adult rat. Brain Res. 714:177, 1996. Nagamatsu, M., Terao, S., Misu, K., et al.: Axonal and perikaryal involvement in chronic inflammatory demyelinating polyneuropathy. J. Neurol. Neurosurg. Psychiatry 66:727, 1999. Nagano, I., Shapshak, P., Yoshioka, M., et al.: Increased NADPH-diaphorase reactivity and cytokine expression in dorsal root ganglia in acquired immunodeficiency syndrome. J. Neurol. Sci. 136:117, 1996. Nagashima, K., Sakakibara, K., Endo, H., et al.: I-cell disease (mucolipidosis II): pathological and biochemical studies of an autopsy case. Acta Pathol. Jpn. 27:251, 1977. Nandy, K.: Further effects of centrophenoxine on the lipofuscin pigment in the neurons of senile guinea pigs. J. Gerontol. 23:82, 1968. Nandy, K.: Properties of neuronal lipofuscin pigment in mice. Acta Neuropathol. (Berl.) 19:25, 1971. Narahara, K., Takahashi, Y., Murakami, M., et al.: Terminal 22q deletion associated with a partial deficiency of arylsulphatase A. J. Med. Genet. 29:432, 1992. Nebes, V. L., and Schmidt, M. C.: Human lysosomal alphamannosidase: isolation and nucleotide sequence of the full-length cDNA. Biochem. Biophys. Res. Commun. 200:239, 1994. Nishimura, M., Namba, Y., Ikeda, K., et al.: Neurofibrillary tangles in the neurons of spinal dorsal root ganglia of patients with progressive supranuclear palsy. Acta Neuropathol. (Berl.) 85:453, 1993. Nittono, K.: On bilateral effects from the unilateral section of branches of the nervous trigeminus in the albino rat. J. Comp. Neurol. 35:133, 1923. Nolan, C. C., Brown, A. W., and Cavanagh, J. B.: Regional variations in nerve cell responses to trimethyltin intoxication I. Mongolian gerbils and rats: further evidence for involvement of the Golgi apparatus. Acta Neuropathol. (Berl.) 81:204, 1990. Novikov, L., Novikova, L., and Kellerth, J. O.: Brain-derived neurotrophic factor promotes survival and blocks nitric oxide synthase expression in adult rat spinal motoneurons after ventral root avulsion. Neurosci. Lett. 200:45, 1995. Oblinger, M. M., and Lasek, R. J.: Axotomy-induced alterations in the synthesis and transport of neurofilaments and

Pathology of Peripheral Neuron Cell Bodies

328. 329.

330.

331.

332. 333. 334.

335.

336.

337.

338.

339.

340.

341.

342.

343.

344.

345.

microtubules in dorsal root ganglion cells. J. Neurosci. 8:1747, 1988. Ohnishi, A., and Dyck, P. J.: Loss of small peripheral sensory neurons in Fabry disease. Arch. Neurol. 31:120, 1974 Ohtake, K., Matsumoto, M., Wakamatsu, H., et al.: Glutamate release and neuronal injury after intrathecal injection of local anaesthetics. Neuroreport 11:1105, 2000. Okado, N., and Oppenheim, R. W.: Cell death of motoneurons in the chick embryo spinal cord. IX. The loss of motoneurons following removal of afferent inputs. J. Neurosci. 4:1639, 1984. Ong, W. Y., Kumar, U., Switzer, R. C., et al.: Neurodegeneration in Niemann-Pick type C disease mice. Exp. Brain Res. 141:218, 2001. Oorschot, D. E., and McLennan, I. S.: The trophic requirements of mature motoneurons. Brain Res. 789:315, 1998. Oppenheim, R. W.: Cell death during development of the nervous system. Annu. Rev. Neurosci. 14:453,1991. Orozco, G., Estrada, R., Perry, T., et al.: Dominantly inherited olivopontocerebellar atrophy from Eastern Cuba: clinical, neuropathological and biochemical findings. J. Neurol. Sci. 93:37, 1989. Orr, H. T., Chung, M.-Y., Banfi, S., et al.: Expansion of an unstable trinucleotide CAG repeat in spinocerebellar ataxia type 1. Nat. Genet. 4:221, 1993. O’Toole, D., Ingram, J., Welch, V., et al.: An inherited lower motor neuron disease of pigs: clinical signs in two litters and pathology of an affected pig. J. Vet. Diagn. Invest. 6:62, 1994. Otto, D., Unsicker, K., and Grothe, C.: Pharmacological effects of nerve growth factor and fibroblast growth factor applied to the transectioned sciatic nerve on neuron death in adult rat dorsal root ganglia. Neurosci. Lett. 83:156, 1987. Oyanagi, K., Makifuchi, T., Ohtoh, T., et al.: Amyotrophic lateral sclerosis of Guam: the nature of the neuropathological findings. Acta Neuropathol. (Berl.) 88:405, 1994. Pang, J., Giunti, P., Chamberlain, S., et al.: Neuronal intranuclear inclusions in SCA2: a genetic, morphological and immunohistochemical study of two cases. Brain 125:656, 2002. Pannese, E.: Investigations on the ultrastructural changes of the spinal ganglion neurons in the course of axon regeneration and cell hypertrophy. I. Changes during axon regeneration. Z. Zellforsch. Mikroskop. Anat. 60:711, 1963. Pardo, C. A., McArthur, J. C., and Griffin, J. W.: HIV neuropathy: insight in the pathology of HIV peripheral nerve disease. J. Peripher. Nerv. Syst. 6:21, 2001. Paulson, H. L., Perez, M. K., Trottier, Y., et al.: Intranuclear inclusions of expanded polyglutamine protein in spinocerebellar ataxia type 3. Neuron 19:333, 1997. Peiffer, J., Schlote, W., Bischoff, A., et al.: Generalized giant axonal neuropathy: a filament-forming disease of neuronal, endothelial, glial, and Schwann cells in a patient without kinky hair. Acta Neuropathol. (Berl.) 40:213, 1977. Petit, P. X., Goubern, M., Diolez, P., et al.: Disruption of the outer mitochondrial membrane as a result of large amplitude swelling: the impact of irreversible permeability transition. FEBS Lett. 426:111, 1998. Pettmann, B., and Henderson, C. E.: Neuronal cell death. Neuron 20:633, 1998.

727

346. Peyronnard, J.-M., Charron, L., Messier, J.-P., and Lavoie, J.: Differential effects of distal and proximal nerve lesions on carbonic anhydrase activity in rat primary sensory neurons, ventral and dorsal root axons. Exp. Brain Res. 70:550, 1988. 347. Peyronnard, J.-M., Charron, L., Messier, J.-P., et al.: Changes in lectin binding of lumbar dorsal root ganglia neurons and peripheral axons after sciatic and spinal nerve injury in the rat. Cell Tissue Res. 257:379, 1989. 348. Picklo, M. J., Wiley, R. G., Lappi, D. A., and Robertson, D.: Noradrenergic lesioning with an anti-dopamine betahydroxylase immunotoxin. Brain Res. 666:195, 1994. 349. Piehl, F., Arvidsson, U., Johnson, H., et al.: GAP43, aFGF, CCK, ␣ and ␤ CGRP in rat spinal motoneurons subjected to axotomy and/or dorsal root severance. Eur. J. Neurosci. 5:1321, 1993. 350. Piehl, F., Tabar, G., and Cullheim, S.: Expression of NMDA receptor mRNAs in rat motoneurons is down regulated after axotomy. Eur. J. Neurosci. 7:2101, 1995. 351. Pollin, M. M., McHanwell, S., and Slater, C. R.: Loss of motor neurons from the median nerve motor nucleus of the mutant mouse ‘wobbler.’ J. Neurocytol. 19:29, 1990. 352. Popken, G. J., and Farel, P. B.: Sensory neuron number in neonatal and adult rats estimated by means of stereologic and profile-based methods. J. Comp. Neurol. 386:8, 1997. 353. Portera-Cailliau, C., Price, D. L., and Martin, L. J.: NonNMDA and NMDA receptor-mediated excitotoxic neuronal deaths in adult brain are morphologically distinct: further evidence for an apoptosis-necrosis continuum. J. Comp. Neurol. 378:88, 1997. 354. Powers, B. E., Beck, E. R., Gillette, E. L., et al.: Pathology of radiation injury to the canine spinal cord. Int. J. Radiat. Oncol. Biol. Phys. 23:539, 1992. 355. Powers, J. M.: Normal and defective neuronal membranes: structure and function. Neuronal lesions in peroxisomal disorders. J. Mol. Neurosci. 16:285, 2001. 356. Powers, J. M., DeCiero, D. P., Cox, C., et al.: The dorsal root ganglia in adrenomyeloneuropathy: neuronal atrophy and abnormal mitochondria. J. Neuropathol. Exp. Neurol. 60:493, 2001. 357. Price, D. L., and Porter, K. R.: The response of ventral horn neurons to axonal transection. J. Cell Biol. 53:24, 1972. 358. Pullen, A. H.: Morphometric evidence from c-synapses for phased Nissl body response in alpha-motoneurones retrogradely intoxicated with diphtheria toxin. Brain Res. 509:8, 1990. 359. Raas-Rothschild, A., Cormier-Daire, V., Bao, M., et al.: Molecular basis of variant pseudo-Hurler polydystrophy (mucolipidosis IIIC). J. Clin. Invest. 105:673, 2000. 360. Raivich, G., Hellweg, R., and Kreutzberg, G. W.: NGF receptor-mediated reduction in axonal NGF uptake and retrograde transport following sciatic nerve injury and during regeneration. Neuron 7:151,1991. 361. Rance, N. E., McArthur, J. C., Cornblath, D. R., et al.: Gracile tract degeneration in patients with sensory neuropathy and AIDS. Neurology 38:265, 1988. 362. Ranson, S. W.: Retrograde degeneration in the spinal nerves. J. Comp. Neurol. Psychol. 16:265, 1906. 363. Ranson, S. W.: Alterations in the spinal ganglion cells following axotomy. J. Comp. Neurol. 19:125, 1909.

728

Pathology of the Peripheral Nervous System

364. Ranta, S., Savukoski, M., Santavuori, P., and Haltia, M.: Studies of homogenous populations: CLN5 and CLN8. Adv. Genet. 45:123, 2001. 365. Resch, K., Korthaus, D., Wedemeyer, N., et al.: Homology between human chromosome 2p13.3 and the wobbler critical region on mouse chromosome 11: comparative highresolution mapping of STS and EST loci on YAC/BAC contigs. Mamm. Genome 9:893, 1998. 366. Rich, K. M., Disch, S. P., and Eichler, M. E.: The influence of regeneration and nerve growth factor on the neuronal cell body reaction to injury. J. Neurocytol. 18:569, 1989. 367. Rich, K. M., Luszczynski, J. R., Osborne, P. A., and Johnson, E. M.: Nerve growth factor protects adult sensory neurons from cell death and atrophy caused by nerve injury. J. Neurocytol. 16:261, 1987. 368. Risling, M., Aldskogius, H., and Hildebrand, C.: Effects of sciatic nerve crush on the L7 spinal roots and dorsal root ganglia in kittens. Exp. Neurol. 79:176, 1983. 369. Robbins, J. H., Kraemer, K. H., Merchant, S. N., and Brumback, R. A.: Adult-onset xeroderma pigmentosum neurological disease—observations in an autopsy case. Clin. Neuropathol. 21:18, 2002. 370. Robitaille, Y., Schut, L., and Kish, S. J.: Structural and immunocytochemical features of olivopontocerebellar atrophy caused by the spinocerebellar ataxia type 1 (SCA-1) mutation define a unique phenotype. Acta Neuropathol. (Berl.) 90:572, 1995. 371. Romanes, G. J.: Motor localization and the effects of nerve injury on the ventral horn cells of the spinal cord. J. Anat. 42:117, 1946. 372. Rossiter, J. P., Riopelle, R. J., and Bisby, M. A.: Axotomyinduced apoptotic cell death of neonatal rat facial motoneurons: time course analysis and relation to NADPH-diaphorase activity. Exp. Neurol. 138:33, 1996. 373. Rowland, L. P., Defendini, R., Sherman, W., et al.: Macroglobulinemia with peripheral neuropathy simulating motor neuron disease. Ann. Neurol. 11:532, 1982. 374. Russell, J. W., Sullivan, K. A., Windebank, A. J., et al.: Neurons undergo apoptosis in animal and cell culture models of diabetes. Neurobiol. Dis. 6:347, 1999. 375. Rust, S., Rosier, M., Funke, H., et al.: Tangier disease is caused by mutations in the gene encoding ATP-binding cassette transporter 1. Nat. Genet. 22:352, 1999. 376. Rydh-Rinder, M., Holmberg, K., Elfvin, L.-G., et al.: Effects of peripheral axotomy on neuropeptides and nitric oxide synthase in dorsal root ganglia and spinal cord of the guinea pig: an immunohistochemical study. Brain Res. 707:180, 1996. 377. Sack, G. H., Cork, L. C., Morris, J. M., et al.: Autosomal dominant inheritance of hereditary canine spinal muscular dystrophy. Ann. Neurol. 15:369, 1984. 378. Saito, Y., and Murayama, S.: Expression of tau immunoreactivity in the spinal motor neurons in Alzheimer’s disease. Neurology 55:1727, 2000. 379. Samuel, B. U., Ponnuraj, E., Rajasingh, J., and John, T. J.: Experimental poliomyelitis in bonnet monkey: clinical features, virology and pathology. Dev. Biol. Stand. 78:71, 1993. 380. Sanchez-Vives, M. V., Valdeolmillos, M., Martinez, S., and Gallego, R.: Axotomy-induced changes in Ca2⫹ homeostasis in rat sympathetic ganglion cells. Eur. J. Neurosci. 6:9, 1994.

381. Sango, K., Yamanaka, S., Hoffmann, A., et al.: Mouse models of Tay-Sachs and Sandhoff diseases differ in neurologic phenotype and ganglioside metabolism. Nat. Genet. 11:170, 1995. 382. Sasaki, S., Shibata, N., Komori, T., and Iwata, M.: iNOS and nitrotyrosine immunoreactivity in amyotrophic lateral sclerosis. Neurosci. Lett. 291:44, 2000. 383. Satake, M., Yoshimura, T., Iwaki, T., et al.: Anti-dorsal root ganglion neuron antibody in a case of dorsal root ganglionitis associated with Sjögren syndrome. J. Neurol. Sci. 132:122, 1995. 384. Scaravilli, F.: Changes in neuronal structure and cell populations with ageing. In Thomas, P. K. (ed.): Peripheral Nerve Changes in the Elderly (New Issues in Neurosciences). Amsterdam, Wiley, p. 95, 1988. 385. Scaravilli, F., An, S.-F., Groves, M., and Thom, M.: The neuropathology of paraneoplastic syndromes. Brain Pathol. 9:251, 1999. 386. Scaravilli, F., and Duchen, L. W.: Electron microscopic and quantitative studies of cell necrosis in developing sensory ganglia in normal and sprawling mutant mice. J. Neurocytol. 9:373, 1980. 387. Scaravilli, F., Gray, F., Mikol, J., and Sinclair, E.: Pathology of the nervous system. In Scaravilli, F. (ed.): The Neuropathology of HIV Infection. London, Springer, p. 99, 1993. 388. Scaravilli, F., and Jacobs, J. M.: Quantitative studies of motor, spino-cerebellar and secondary sensory neurons in the mutilated foot mutant rat. Neuroscience 6:1663, 1981. 389. Scaravilli, F., Sinclair, E., Arango, J. C., et al.: The pathology of the posterior root ganglia in AIDS and its relationship to the pallor of the gracile tract. Acta Neuropathol. (Berl.) 84:163, 1992. 390. Scaravilli, T., Pramstaller, P. P., Salerno, A., et al.: Neuronal loss in Onuf’s nucleus in three patients with progressive supranuclear palsy. Ann. Neurol. 48:97, 2000. 391. Schaumberg, H. H., and Berger, A. R.: Human toxic neuropathy due to industrial agents. In Dyck, P. J., Thomas, P. K., Griffin, J. W., et al. (eds.): Peripheral Neuropathy, 3rd ed. Philadelphia, W. B. Saunders, p. 1533, 1993. 392. Schinder, V., Govrin-Lippmann, R., Cohen, S., et al.: Structural basis of sympathetic-sensory coupling in rat and human dorsal root ganglia following peripheral nerve injury. J. Neurocytol. 28:743, 1999. 393. Schlote, W., Harzer, K., Christomanou, H., et al.: Sphingolipid activator protein 1 deficiency in metachromatic leukodystrophy with normal arylsulphatase A activity: a clinical, morphological, biochemical, and immunological study. Eur. J. Pediatr. 150:584, 1991. 394. Schmalbruch, H.: Loss of sensory neurons after sciatic nerve section in the rat. Anat. Rec. 219:323, 1987. 395. Schmalbruch, H., and Haase, G.: Spinal muscular atrophy: present state. Brain Pathol. 11:231, 2001. 396. Schmalbruch, H., Stender, S., and Boysen, G.: Abnormalities in spinal neurons and dorsal root ganglion cells in Tangier disease presenting with a syringomyelia-like syndrome. J. Neuropathol. Exp. Neurol. 46:533, 1987. 397. Schmidt, M. L., Zhukareva, V., Perl, D. P., et al.: Spinal cord neurofibrillary pathology in Alzheimer’s disease and Guam

Pathology of Peripheral Neuron Cell Bodies

398. 399.

400.

401.

402.

403.

404.

405.

406.

407.

408.

409.

410.

411.

412.

413.

414.

parkinsonism-dementia complex. J. Neuropathol. Exp. Neurol. 60:1075, 2001. Schmidt, R. E.: Neuropathology of human sympathetic autonomic ganglia. Microsc. Res. Tech. 35:107, 1996. Schmidt, R. E., Beaudet, L. N., Plurad, S. B., and Dorsey, D. A.: Axonal cytoskeletal pathology in aged and diabetic human sympathetic autonomic ganglia. Brain Res. 769:375, 1997. Schmidt, R. E., Dorsey, D., Parvin, C. A., et al.: Dystrophic axonal swellings develop as a function of age and diabetes in human dorsal root ganglia. J. Neuropathol. Exp. Neurol. 56:1028, 1997. Schmidt, R. E., Plurad, S. B., Parvin, C. A., and Roth, K. A.: Effect of diabetes and ageing on human sympathetic autonomic ganglia. Am. J. Pathol. 143:143, 1993. Schold, S. C., Cho, E. S., Somasundaram, M., and Posner, J. B.: Subacute motor neuronopathy: a remote effect of lymphoma. Ann. Neurol. 5:271, 1979. Schroer, J. A., Plurad, S. B., and Schmidt, R. E.: Fine structure of presynaptic axonal terminals in sympathetic autonomic ganglia of ageing and diabetic human subjects. Synapse 12:1, 1992. Schwab, C., DeMaggio, A. J., Ghoshal, N., et al.: Casein kinase 1 delta is associated with pathological accumulation of tau in several neurodegenerative diseases. Neurobiol. Aging 21:503, 2000. Scott, J. N., Clark, A. W., and Zochodne, D. W.: Neurofilament and tubulin gene expression in progressive experimental diabetes: failure of synthesis and export by sensory neurons. Brain 122:2019, 1999. Sebert, M. E., and Shooter, E. M.: Expression of mRNA for neurotrophic factors and their receptors in the rat dorsal root ganglion and sciatic nerve following nerve injury. J. Neurosci. Res. 36:357, 1993. Seil, F. J., Lampert, P. W., and Klatzo, I.: Neurofibrillary spheroids induced by aluminum phosphate in dorsal root ganglia neurons in vitro. J. Neuropathol. Exp. Neurol. 28:74, 1969. Sekhon, S. S., and Maxwell, D. S.: Ultrastructural changes in neurons of the spinal anterior horn of ageing mice with particular reference to the accumulation of lipofuscin pigment. J. Neurocytol. 3:59, 1974. Shadiack, A. M., Sun, Y., and Zigmond, R. E.: Nerve growth factor antiserum induces axotomy-like changes in neuropeptide expression in intact sympathetic and sensory neurons. J. Neurosci. 21:363, 2001. Shaw, C. E., al-Chalabi, A., and Leigh, N.: Progress in the pathogenesis of amyotrophic lateral sclerosis. Curr. Neurol. Neurosci. Rep. 1:69, 2001. Shen, H., Chung, J. M., Coggeshall, R. E., and Chung, K.: Changes in trkA expression in the dorsal root ganglion after peripheral nerve injury. Exp. Brain Res. 127:141, 1999. Sheth, K. J., Werlin, S. L., Freeman, M. E., and Hodach, A. E.: Gastrointestinal structure and function in Fabry’s disease. Am. J. Gastroenterol. 76:246, 1981. Shi, T. J., Tandrup, T., Bergman, E., et al.: Effect of peripheral nerve injury on dorsal root ganglion neurons in the C57 BL/6J mouse: marked changes both in cell numbers and neuropeptide expression. Neuroscience 105:249, 2001. Siconolfi, L. B., and Seeds, N. W.: Induction of the plasminogen activator system accompanies peripheral nerve regeneration after sciatic crush. J. Neurosci. 21:4336, 2001.

729

415. Silos-Santiago, I., Molliver, D. C., Ozaki, S., et al.: Non-trkAexpressing small DRG neurons are lost in TrkA deficient mice. J. Neurosci. 15:5929, 1995. 416. Silverman, J. D., and Kruger, L.: Selective neuronal glycoconjugate expression in sensory and autonomic ganglia: relation of lectin reactivity to peptide and enzyme markers. J. Neurocytol. 19:789, 1990. 417. Simic, G., Seso-Simic, D., Lucassen, P. J., et al.: Ultrastructural analysis and TUNEL demonstrate motor neuron apoptosis in Werdnig-Hoffmann disease. J. Neuropathol. Exp. Neurol. 59:398, 2000. 418. Simon, L. T., Ricaurte, G. A., and Forno, L. S.: Chronic idiopathic ataxic neuropathy: neuropathology of a case. Acta Neuropathol. (Berl.) 79:104, 1989. 419. Singer, P. A., Mehler, S., and Fernandez, H. L.: Blockade of retrograde axonal transport delays the onset of metabolic and morphologic changes induced by axotomy. J. Neurosci. 2:1299, 1982. 420. Singer, P. A., Mehler, S., and Fernandez, H. L.: Effect of extracts of injured nerve on initiating the regenerative response in the hypoglossal nucleus in the rat. Neurosci. Lett. 84:155, 1988. 421. Slaugenhaupt, S. A., Acierno, J. S. Jr., Helbling, L. A., et al.: Mapping of the mucolipidosis type IV gene to chromosome 19p and definition of founder haplotypes. Am. J. Hum. Genet. 65:773, 1999. 422. Soares, H. D., Chen, S. C., and Morgan, J. I.: Differential and prolonged expression of Fos-lacZ and jun-lacZ in neurons, glia, and muscle following sciatic nerve damage. Exp. Neurol. 167:1, 2001. 423. Sobkowicz, H. M., Slapnick, S. M., and August, B. K.: Apoptosis of inner hair cells caused by laser ablation of their spiral ganglion neurons in cultures of the mouse organ of Corti. J. Neurocytol. 28:939, 1999. 424. Sobue, G., Hashizume, Y., Mukai, E., et al.: X-linked recessive bulbospinal neuronopathy. Brain 112:209, 1989. 425. Sobue, G., Nakao, N., Murakami, K., et al.: Type 1 familial amyloid polyneuropathy. Brain 113:903, 1990. 426. Sodhi, N., Camilleri, M., Camoriano, J. K., et al.: Autonomic function and motility in intestinal pseudo-obstruction caused by paraneoplastic syndrome. Dig. Dis. Sci. 34:1937, 1989. 427. Soffer, D.: Neuronal intranuclear hyaline inclusion disease presenting as Friedreich’s ataxia. Acta Neuropathol. (Berl.) 65:322, 1985. 428. Sommervaille, T., Molander, C., and Woolf, C. J.: Timedependent differences in the increase in GAP-43 expression in dorsal root ganglion cells after peripheral axotomy. Neuroscience 45:213, 1991. 429. Sosunov, A. A., Krugliakov, P. P., Shvalev, V. N., et al.: Age related changes in the autonomic ganglia [in Russian]. Arkh. Pathol. 59:32, 1997. 430. Sterman, A. B.: Acrylamide induces early morphologic reorganization of the neuronal cell body. Neurology 32:1023, 1982. 431. Sterman, A. B.: Cell body remodeling during dying-back axonopathy: DRG changes during advanced disease. J. Neuropathol. Exp. Neurol. 41:400, 1982. 432. Sterman, A. B., and Delannoy, M. R.: Cell body responses to axonal injury: traumatic axotomy versus toxic neuropathy. Exp. Neurol. 89:408, 1985.

730

Pathology of the Peripheral Nervous System

433. Stoltenburg-Didinger, G., Boegner, F., Gruning, W., et al.: Specific neurotoxic effects of different organic solvents on dissociated cultures of the nervous system. Neurotoxicology 13:161, 1992. 434. Su, M., Wakabayashi, K., Kakita, A., et al.: Selective involvement of large motor neurons in the spinal cord of rats treated with methylmercury. J. Neurol. Sci. 156:12, 1998. 435. Sun, Y., and Zigmond, R. E.: Regulation of neuropeptide expression in sympathetic neurons: paracrine and retrograde influences. Ann. N. Y. Acad. Sci. 814:181, 1997. 436. Sung, J. H.: Light, fluorescence, and electron microscopic features of neuronal intranuclear hyaline inclusions associated with multisystem atrophy. Acta Neuropathol. (Berl.) 50:115, 1980. 437. Sung, J. H., and Mastri, A. R.: Spinal autonomic neurons in Werdnig-Hoffmann disease, mannosidosis, and Hurler’s syndrome: distribution of autonomic neurons in the sacral spinal cord. J. Neuropathol. Exp. Neurol. 39:441, 1980. 438. Sung, J. H., Meyers, J. P., Stadlan, E. M., et al.: Neuropathological changes in Chediak-Higashi disease. J. Neuropathol. Exp. Neurol. 28:86, 1969. 439. Sung, J. O., Hayano, M., and Desnick, R. J.: Mannosidosis: pathology of the nervous system. J. Neuropathol. Exp. Neurol. 36:807, 1977. 440. Suzuki, H., Oyanagi, K., Takahashi, H., et al.: A quantitative pathological investigation of the cervical cord, roots and ganglia after long-term amputation of the unilateral upper arm. Acta Neuropathol. (Berl.) 85:666, 1993. 441. Suzuki, K.: Neuropathology of late onset gangliosidosis: a review. Dev. Neurosci. 13:205, 1991. 442. Suzuki, K., and Suzuki, K.: Lysosomal diseases. In Graham, D. I., and Lantos, P. L. (eds.): Greenfield’s Neuropathology. London, Arnold, p. 653, 2002. 443. Svensson, M., and Aldskogius, H.: The effect of axon injury on microtubule-associated proteins MAP2, 3 and 5 in the hypoglossal nucleus of the adult rat. J. Neurocytol. 21:222, 1992. 444. Swett, J. E., Hong, C.-Z., and Miller, P. G.: Most dorsal root ganglion neurons of the adult rat survive nerve crush injury. Somatosens. Motor Res. 12:177, 1995. 445. Tabira, T., Goto, I., and Kuroiwa, Y.: Neuropathological and biochemical studies in Fabry’s disease. Acta Neuropathol. (Berl.) 30:345, 1974. 446. Takada, K., and Becker, L. E.: Cockayne’s syndrome: report of two autopsy cases associated with neurofibrillary tangles. Clin. Neuropathol. 5:64, 1986. 447. Takahashi, J., Tanaka, J., Arai, K., et al.: Recruitment of nonexpanded polyglutamine proteins to intranuclear aggregates in neuronal intranuclear hyaline inclusion disease. J. Neuropathol. Exp. Neurol. 60:369, 2001. 448. Takano, T., and Yamanouchi, Y.: Assignment of human betagalactosidase-A gene to 3p21.33 by fluorescence in situ hybridization. Hum. Genet. 92:403, 1993. 449. Takase, Y., Takahashi, K., Takada, K., et al.: Hereditary motor and sensory neuropathy type 1 (HMSN 1) associated with cranial neuropathy: an autopsy case report. Acta Neurol. Scand. 82:368, 1990. 450. Tanabe, K., Kiryu-Seo, S., Nakamura, T., et al.: Alternative expression of Shc family members in nerve-injured motoneurons. Brain Res. Mol. Brain Res. 53:291, 1998.

451. Tandrup, T., and Braendgaard, H.: Number and volume of rat dorsal root ganglion cells in acrylamide intoxication. J. Neurocytol. 23:242, 1994. 452. Tandrup, T., Woolf, C. J., and Coggeshall, R. E.: Delayed loss of small dorsal root ganglion cells after transection of the rat sciatic nerve. J. Comp. Neurol. 422:172, 2000. 453. Tang, T. T., Segura, A. D., Chen, Y. T., et al.: Neonatal hypotonia and cardiomyopathy secondary to type IV glycogenosis. Acta Neuropathol. (Berl.) 87:531, 1994. 454. Tellez Nagel, I., Rapin, I., Iwamoto, T., et al.: Mucolipidosis IV: clinical, ultrastructural, histochemical and chemical studies of a case, including a brain biopsy. Arch. Neurol. 33:828, 1976. 455. Tenser, R. B.: Sequential changes of sensory neuron (fluoride resistant) acid phosphatase in dorsal root ganglion neurons following neurectomy and rhizotomy. Brain Res. 332:386, 1985. 456. Terenghi, G.: Peripheral nerve regeneration and neurotrophic factors. J. Anat. 194:1, 1999. 457. Terry, R. D., and Weiss, M.: Studies in Tay-Sachs disease. II. Ultrastructure of the cerebrum. J. Neuropathol. Exp. Neurol. 22:18, 1963. 458. Tetzlaff, W., Alexander, S. W., Miller, F. D., and Bisby, M. A.: Response of facial and rubrospinal neurons to axotomy: changes in mRNA expression for cytoskeletal proteins and GAP-43. J. Neurosci. 11:2528, 1991. 459. Tetzlaff, W., and Bisby, M. A.: Cytochemical protein synthesis and regulation of nerve regeneration in PNS and CNS neurons of the rat. Restor. Neurol. Neurosci. 1:189, 1990. 460. Thorner, P. S., Bilbao, J. M., Sima, A. A. F., and Briggs, S.: Porphyric neuropathy: an ultrastructural and quantitative case study. Can. J. Neurol. Sci. 8:281, 1981. 461. Tolkovsky, A.: Neurotrophic factors in action—new dogs and new tricks. TINS 20:1, 1997. 462. Tong, J. X., and Rich, K. M.: Diphenylpiperazines enhance regeneration after facial nerve injury. J. Neurocytol. 26:339, 1997. 463. Tonra, J. R., Curtis, R., Wong, V., et al.: Axotomy upregulates the anterograde transport and expression of brain-derived neurotrophic factor by sensory neurons. J. Neurosci. 18:4374, 1998. 464. Topp, K. S., Tanner, K. D., and Levine, J. D.: Damage to the cytoskeleton of large diameter sensory neurons and myelinated axons in vincristine-induced painful peripheral neuropathy in the rat. J. Comp. Neurol. 424:563, 2000. 465. Tornquist, E., Liu, L., Mattsson, P., and Svensson, M.: Response of glial cells and activation of complement following motorneuron degeneration induced by toxic ricin. Neurosci. Res. 28:167, 1997. 466. Toscano, E., della Casa, R., Mardy, S., et al.: Multisystem involvement in congenital insensitivity to pain with anhidrosis (CIPA), a nerve growth factor receptor (trk A)-related disorder. Neuropediatrics 31:39, 2000. 467. Toyooka, K., Fujimura, H., Ueno, S., et al.: Familial amyloid polyneuropathy associated with transthyretin Gly42 mutation: a qualitative light and electron microscopic study of the peripheral nervous system. Acta Neuropathol. (Berl.) 90:516, 1995. 468. Tredici, G., Tredici, S., Fabbrica, D., et al.: Experimental cisplatin neuronopathy in rats and the effect of retinoic acid administration. J. Neurooncol. 36:31, 1998.

Pathology of Peripheral Neuron Cell Bodies 469. Trottier, Y., Lutz, Y., Stevanin, G., et al.: Polyglutamine expansion as a pathological epitope in Huntington’s disease and four dominant cerebellar ataxias. Nature 378:403, 1995. 470. Troy, C. M., Muma, N. A., Greene, L. A., et al.: Regulation of peripherin and neurofilament expression in regenerating rat motor neurons. Brain Res. 529:232, 1990. 471. Tsujino, H., Kondo, E., Fukuoka, T., et al.: Activating transcription factor 3 (ATF3) induction by axotomy in sensory and motoneurons: a novel neuronal marker of nerve injury. Mol. Cell Neurosci. 15:170, 2000. 472. Tsuzuki, K., Kondo, E., Fukuoka, T., et al.: Differential regulation of P2X(3) mRNA expression by peripheral nerve injury in intact and injured neurons in the rat sensory ganglia. Pain 91:351, 2001. 473. Umahara, T., Hirano, A., Kato, S., et al.: Demonstration of neurofibrillary tangles and neuropil thread-like structures in spinal cord white matter in Parkinson-dementia complex of Guam and in Guamanian amyotrophic lateral sclerosis. Acta Neuropathol. (Berl.) 88:180, 1994. 474. Urushanti, M., and Shimohama, S.: The role of nitric oxide in amyotrophic lateral sclerosis. Amyotroph. Lateral Scler. Other Motor Neuron Disord. 2:71, 2001. 475. Van Den Bosch, L., Schwaller, B., Vleminckx, V., et al.: Protective effect of parvalbumin on excitotoxic motor neuron death. Exp. Neurol. 174:150, 2002. 476. Van den Bosch de Aguilar, P. H., and Goemare-Vanneste, J.: Paired helical filaments in spinal ganglion neurons of elderly rats. Virchows Arch. B 47:217, 1984. 477. van der Walt, J. D., Swash, M., Leake, J., and Cox, E. L.: The pattern of involvement of adult-onset acid maltase deficiency at autopsy. Muscle Nerve 10:272, 1987. 478. Vandenberghe, W., Robberecht, W., and Brorson, J. R.: AMPA receptor calcium permeability, GluR2 expression, and selective motoneuron vulnerability. J. Neurosci. 20:123, 2000. 479. Vanderluit, J. L., McPhail, L. T., Fernandes, K. J., et al.: Caspase-3 is activated following axotomy of neonatal facial motoneurons and caspase-3 gene deletion delays axotomyinduced cell death in rodents. Eur. J. Neurosci. 12:3469, 2000. 480. Verge, V., Zhang, X., Xu, X.-J., et al.: Marked increase in nitric oxide synthase mRNA in rat dorsal root ganglia after peripheral axotomy: in situ hybridization and functional studies. Proc. Natl. Acad. Sci. USA 89:11617, 1992. 481. Verge, V. M. K., Richardson, P. M., Wiesenfeld-Hallin, Z., and Hökfelt, T.: Differential influence of nerve growth factor on neuropeptide expression in vivo: a novel role in peptide suppression in adult sensory neurons. J. Neurosci. 15:2081, 1995. 482. Verkerk, A. J., Pieretti, M., Sutcliffe, J. S., et al.: Identification of a gene (FMR-1) containing a CGG repeat coincident with a breakpoint cluster region exhibiting length variation in fragile X syndrome. Cell 65:905, 1991. 483. Verma, A., Berger, J. R., Snodgrass, S., and Petito, C.: Motor neuron disease: a paraneoplastic process associated with anti-Hu antibody and small cell lung carcinoma. Ann. Neurol. 40:112, 1996. 484. Vestergaard, S., Tandrup, T., and Jakobsen, J.: Effect of permanent axotomy on number and volume of dorsal root ganglion cell bodies. J. Comp. Neurol. 388:307, 1997. 485. Villar, M. J., Cortés, R., Theodorsson, E., et al.: Neuropeptide expression in rat dorsal root ganglion cells and spinal cord

486.

487.

488.

489.

490.

491. 492.

493.

494.

495.

496.

497.

498.

499.

500.

501.

502.

731

after peripheral nerve injury with special reference to galanin. Neuroscience 33:587, 1989. Vitaliani, R., Scaravilli, T., Egarter-Vigl, E., et al.: The pathology of the spinal cord in progressive supranuclear palsy. J. Neuropathol. Exp. Neurol. 61:268, 2002. Wakabayashi, K., Kobayashi, H., Kawasaki, S., et al.: Autosomal recessive spastic paraplegia with hypoplastic corpus callosum, multisystem degeneration and ubiquinated eosinophilic granules. Acta Neuropathol. (Berl.) 101:69, 2001. Wakamatsu, N., Yamada, Y., Yamada, K., et al.: Mutations in SIP1, encoding Smad interacting protein-1, cause a form of Hirschsprung disease. Nat. Genet. 27:369, 2001. Wang, X., Herberg, F. W., Laue, M. M., et al.: Neurobeachin: a protein kinase A-anchoring, beige/Chediak-Higashi protein homolog implicated in neuronal membrane traffic. J. Neurosci. 20:8551, 2000. Wang, Z. M., Dai, C. F., Kanoh, N., et al.: Apoptosis and expression of Bcl-2 in facial motoneurons after facial nerve injury. Otol. Neurotol. 23:397, 2002. Warren, S. T.: The expanding world of trinucleotide repeats. Science 271:1374, 1996. Watkins, P. J., Gayle, C., Alsanjari, N., et al.: Severe sensoryautonomic neuropathy and endocrinopathy in insulindependent diabetes. Q. J. Med. 88:795, 1995. Wells, M. R.: Changes of ornithine decarboxylase activity in dorsal root ganglion cells after axon injury: possible relationship to alterations in neuronal chromatin. Exp. Neurol. 95:313, 1987. Wells, M. R., and Vaidya, U.: Morphological alterations in dorsal root ganglion neurons after peripheral axon injury: association with changes in metabolism. Exp. Neurol. 104:32, 1989. White, C. M., Greensmith, L., and Vrbova, G.: Repeated stimuli for axonal growth causes motoneuron death in adult rats: the effect of botulinum toxin followed by partial denervation. Neuroscience 95:1101, 2000. White, F. A., Bennett-Clarke, C. A., Macdonald, G. J., et al.: Neonatal infraorbital nerve transection in the rat: comparison of effects on substance P immunoreactive primary afferents and those recognized by the lectin Bandierea simplicifolia-I. J. Comp. Neurol. 300:249, 1990. Wierzbicki, A. S., Mitchell, J., Lambert-Hammill, M., et al.: Identification of genetic heterogeneity in Refsum’s disease. Eur. J. Hum. Genet. 8:649, 2000. Wijsman, J. H., Jonker, J. R., Keijzer, R., et al.: A new method to detect apoptosis in paraffin sections: in situ end-labeling of fragmented DNA. J. Histochem. Cytochem. 41:7, 1993. Wiley, R. G., Blessing, W. W., and Reis, D. J.: Suicide transport: destruction of neurons by retrograde transport of ricin, abrin, and modeccin. Science 216:889, 1982. Wiley, R. G., and Stirpe, F.: Neuronotoxicity of axonally transported toxic lectins, abrin, modeccin and volkensin in rat peripheral nervous system. Neuropathol. Appl. Neurobiol. 13:39, 1987. Willemsen, R., Tybulewicz, E., Sidransky, E., et al.: A biochemical and ultrastructural evaluation of the type 2 Gaucher mouse. Mol. Chem. Neuropathol. 24:179, 1995. Willner, J. P., Grabowski, G. A., Gordon, R. E., et al.: Chronic GM2 gangliosidosis masquerading as atypical Friedreich’s ataxia: clinical, morphologic and biochemical studies of nine cases. Neurology 31:787, 1981.

732

Pathology of the Peripheral Nervous System

503. Winchester, B., Young, E., Geddes, S., et al.: Female twin with Hunter disease due to non-random inactivation of X-chromosome: a consequence of twinning. Am. J. Med. Genet. 44:834, 1992. 504. Wong, J., and Oblinger, M. M.: A comparison of peripheral and central axotomy effects on neurofilament and tubulin gene expression in rat dorsal root ganglion neurons. J. Neurosci. 10:2215, 1990. 505. Wong, J., and Oblinger, M. M.: Differential regulation of peripherin and neurofilament gene expression in regenerating rat DRG neurons. J. Neurosci. Res. 27:332, 1990. 506. Wood, K. A., and Youle, R. J.: The role of free radicals and p53 in neuron apoptosis in vivo. J. Neurosci. 15:5851, 1995. 507. Woods, B., and Schaumburg, H. H.: Nigrospinodental degeneration with nuclear ophthalmoplegia: an unique and partially treatable clinicopathological entity. J. Neurol. Sci. 17:149, 1972. 508. Worrall, B. B., Rowland, L. P., Chin, S. S., and Mastrianni, J. A.: Amyotrophy in prion diseases. Arch. Neurol. 57:33, 2000. 509. Wyllie, A. H.: Glucocorticoid-induced thymocyte apoptosis is associated with endogenous endonuclease activation. Nature 284:555, 1980. 510. Wyllie, A. H., Morris, R. G., Smith, A. L., and Dunlop, D.: Chromatin cleavage in apoptosis: association with condensed chromatin morphology and dependence on macromolecular synthesis. J. Pathol. 142:67, 1984. 511. Xian, C. J., and Zhou, X.-F.: Neuronal-glial differential expression of TGF-␣ and its receptor in the dorsal root ganglia in response to sciatic nerve lesion. Exp. Neurol. 157:317, 1999. 512. Xu, G., Pierson, C. R., Murakawa, Y., and Sima, A. A. F.: Altered tubulin and neurofilament expression and impaired axonal growth in diabetic nerve regeneration. J. Neuropathol. Exp. Neurol. 61:164, 2002. 513. Xu, Y., Sladky, J. T., and Brown, M. J.: Dose-dependent expression of neuronopathy after experimental pyridoxine intoxication. Neurology 39:1077, 1989. 514. Yamada, E., Kataoka, H., Isozumi, T., and Hazama, F.: Increased expression of phosphotyrosine after axotomy in the dorsal motor nucleus of the vagus nerve and the hypoglossal nucleus. Acta Neuropathol. (Berl.) 88:14, 1994. 515. Yamada, M., Hayashi, S., Tsuji, S., and Takahashi, H.: Involvement of the cerebral cortex and autonomic ganglia in Machado-Joseph disease. Acta Neuropathol. (Berl.) 101:140, 2001. 516. Yamamoto, T., Iwasaki, Y., Konno, H., and Kudo, H.: Primary degeneration of motor neurons by toxic lectins conveyed from the peripheral nerve. J. Neurol. Sci. 70:327, 1985. 517. Yan, Q., Elliott, J., and Snider, W. D.: Brain-derived neurotrophic factor rescues spinal motor neurons from axotomyinduced cell death. Nature 360:753, 1992. 518. Ygge, J.: Neuronal loss in lumbar dorsal root ganglia after proximal compared to distal sciatic nerve resection: a quantitative study in the rat. Brain Res. 478:193, 1989.

519. Yiangou, Y., Facer, P., Baecker, P. A., et al.: ATP-gated ion channel P2X(3) is increased in human inflammatory bowel disease. Neurogastroenterol. Motil. 13:365, 2001. 520. Yiangou, Y., Facer, P., Birch, R., et al.: P2X3 receptor in injured human sensory neurons. Neuroreport 11:993, 2000. 521. Yiangou, Y., Facer, P., Smith, J. A., et al.: Increased acidsensing ion channel ASIC-3 in inflamed human intestine. Eur. J. Gastroenterol. Hepatol. 13:891, 2001. 522. Yokota, T., Uchihara, T., Kumagai, J., et al.: Postmortem study of ataxia with retinitis pigmentosa by mutation of the alpha-tocopherol transfer protein gene. J. Neurol. Neurosurg. Psychiatry 68:521, 2000. 523. Yoneda, E., Inagaki, S., Hayashi, Y., et al.: Differential regulation of manganese and copper/zinc superoxide dismutases by the facial nerve transection. Brain Res. 582:342, 1992. 524. Yoshimura, I., Yoshimura, N., Hanazono, T., et al.: An autopsy case of neuronal type Charcot-Marie-Tooth disease (HMSN type II) with nerve deafness and psychiatric symptoms [in Japanese]. No To Shinkei 44:571, 1992. 525. Zhang, X., Ju, G., Elde, R., and Hökfelt, T.: Effect of peripheral nerve cut on neuropeptides in dorsal root ganglia and the spinal cord of monkey with special reference to galanin. J. Neurocytol. 22:342, 1993. 526. Zhang, X., Verge, V., Wiesenfeld-Hallin, Z., et al.: Nitric oxide synthase-like immunoreactivity in lumbar dorsal root ganglia and spinal cord of rat and monkey and effect of peripheral axotomy. J. Comp. Neurol. 335:563, 1993. 527. Zhou, X.-F., Deng, Y.-S., Chie, E., et al.: Satellite-cell-derived nerve growth factor and neurotrophin-3 are involved in noradrenergic sprouting in the dorsal root ganglia following peripheral nerve injury in the rat. Eur. J. Neurosci. 11:1711, 1999. 528. Zhou, X.-F., Rush, R. A., and McLachlan, E. M.: Differential expression of the p75 nerve growth factor receptor in glia and neurons of the rat dorsal root ganglia after peripheral nerve transection. J. Neurosci. 16:2901, 1996. 529. Zhuchenko, O., Bailey, J., Bonnen, P., et al.: Autosomal dominant cerebellar ataxia (SCA6) associated with small polyglutamine expansions in the ␣ 1A-voltage-dependent calcium channel. Nat. Genet. 15:62, 1997. 530. Zigmond, R. E.: LIF, NGF, and the cell body response to axotomy. Neuroscientist 3:176, 1997. 531. Ziv, N. E., and Spira, M. E.: Axotomy induces a transient and localized elevation of the free intracellular calcium concentration to the millimolar range. J. Neurophysiol. 74:2625, 1995. 532. Ziv, N. E., and Spira, M. E.: Localized and transient elevations of intracellular Ca2⫹ induce the dedifferentiation of axonal segments into growth cones. J. Neurosci. 17:3568, 1997. 533. Zochodne, D. W., Verge, V. M. K., Cheng, C., et al.: Does diabetes target ganglion neurons? Progressive sensory neuron involvement in long-term experimental diabetes. Brain 124:2319, 2001.

32 Pathologic Alterations of Nerves PETER J. DYCK, P. JAMES B. DYCK, AND JANEAN ENGELSTAD

Role of Pathologic Study of Peripheral Nerves Nerve Biopsy Indications for Nerve Biopsy Nerve Biopsy versus Other Neuropathic Evaluations Peripheral Nerves That May Be Biopsied Symptoms after Nerve Biopsy Postmortem Spinal Cord and Nerve Removal and Processing Histologic Methods For Nerve Biopsy For Postmortem Tissue Teased Fiber History Uses Histologic Technique Teasing Technique Fixation and Teasing Artifacts Sampling Evaluation of Teased Fibers Classification of Teased Fibers Comparison of the Frequency of Teased Fiber Abnormalities Electron Microscopic Assessment of Teased Fibers Electron Microscopic Correlation with Teased Fiber Abnormality Morphometry of Neuron Somas Motor Neuron Columns Spinal Ganglia

Intermediolateral Column Sympathetic Trunk Ganglion Morphometry of Peripheral Nerves Histologic Processing Section Thickness and Axon-Myelin Relationship Attribute of Shape of Sampled Sections of Internodes That Best Estimates Average Diameter Myelinated Fiber Composition: Photographic Method Myelinated Fiber Composition: Computer Imaging Method Normal Numbers of Fibers in Control Human Nerves Artifacts of Surgical Removal and of Histologic Preparation Histologic Structures with and without Pathologic Significance Renaut Corpuscles Schwann Cell Inclusions Fiber Alterations Neuronal Degeneration Wallerian Degeneration Axonal Dystrophy Axonal Atrophy, Myelin Remodeling (Secondary Demyelination) and Degeneration Segmental Demyelination and Remyelination Myelin Remodeling in Normal Nerves

ROLE OF PATHOLOGIC STUDY OF PERIPHERAL NERVES The studies possible on nerves taken at biopsy are different from those possible on nerves taken at necropsy. Both may provide important information about the kind and severity of nerve fiber damage, or lack of it, and of

Differences from Secondary Demyelination Debris Removal from the Endoneurium Nerve Regeneration Microfascicular Regeneration Interstitial Pathology of Nerves Edema and Hemorrhage Compression Ischemia Intraneural Injection of a Foreign Substance Inflammatory Demyelinating Disorders Chronic Immune Sensory Polyradiculopathy Multifocal Motor Neuropathy Leprosy Sarcoidosis Amyloidosis Tumor Perineurial Pathology Parenchymatous Pathologic Alterations Neuronal (Axonal) Pathologic Alterations Classification of Neuronal (Axonal) Pathologic Alterations Schwann Cell Pathologic Alterations

interstitial changes, but there are important advantages and limitations of each. Only on biopsied nerve is it possible to do functional, metabolic, or high-quality ultrastructural studies. Also on biopsy specimens it is possible to study a specific stage of disease. Even with biopsy, a degree of autolysis occurs, but it is much less than in necropsy specimens. Furthermore, insights from biopsy 733

734

Pathology of the Peripheral Nervous System

may be used for the diagnosis and management of the patient from whom it was obtained, obviously not possible on specimens taken at death. However, only at necropsy is it possible to simultaneously sample different rostral-to-caudal segmental levels of the peripheral nervous system, such as different levels of the spinal cord; the roots, spinal ganglia, segmental nerves, and plexuses; and different proximal-to-distal levels of limb nerves. Additionally, large samples of tissue are possible. Pathologic study of nerve focuses on alteration from normal of nerve fibers (parenchymatous change) and on interstitial change. It is possible to determine whether nerve fibers are normal in number or decreased (as a result of failure of development, axonal degeneration, or atrophy and degeneration) or increased (e.g., from regeneration). Next, it is possible to determine whether selective classes of fibers are increased or decreased. The changes that fibers may undergo may indicate underlying pathologic mechanisms of disease (e.g., compression, ischemia, metabolic derangement, storage, infection, inflammation, neoplasia, or other). Nerve biopsy is used to diagnose infections such as leprosy and certain inflammatory-immune reactions: inflammatory demyelination, granulomas, infections, necrotizing vasculitis, and tumor. Some authors think that there is less need for nerve biopsy than formerly. However, in our medical practice and with the development of better imaging of focal nerve enlargement and enhancement, we are demonstrating new needs for and uses of nerve biopsy (Figs. 32–1 and 32–2).

NERVE BIOPSY Indications for Nerve Biopsy It may be easier to rationalize when nerve biopsy should not be done than when it should be done. It should not be done (1) before adequate characterization of the neuropathy by history, neurologic examination, and other test approaches have been done and thought about; (2) when the cause of the neuropathic process is known; (3) when the cost and side effects of the procedure cannot be justified by the information that is likely to be obtained; (4) when the disease process appears to be self-limited or improving; and (5) when, in previous cases of a similar kind, nerve biopsies have repeatedly not been helpful. Nerve biopsy may be considered when the patient: (1) appears to have an interstitial process and one of the following diagnoses is suspected but other less invasive tests have been negative or inconclusive: tumor, leprosy, necrotizing vasculitis, granuloma, amyloidosis, lysosomal storage diseases, or other diseases recognizable by their tissue alterations; (2) has an undiagnosed symptomatic and progressive focal nerve lesion and a putative focal nerve enlargement or enhancement of an accessible nerve (for biopsy) has been identified; and (3) has neuropathic symptoms or impairments and determining neuropathic activity is important. Nerve biopsy can be a very helpful diagnostic procedure, but it should be reserved for cases that have been carefully characterized by other clinical approaches first,

FIGURE 32–1 T2-weighted transverse (left) and sagittal (middle) and postgadolinium sagittal (right) magnetic resonance images of the lumbosacral spine from a patient with a polyradiculoneuropathy of unknown cause. The arrows show thickened and enlarged nerve roots, which enhance with contrast. One of these roots was later biopsied subsequently and diagnosed as sarcoidosis (see Fig. 32–2).

735

Pathologic Alterations of Nerves

FIGURE 32–2 Serial transverse paraffin sections of a dorsal lumbar rootlet from a patient with a polyradiculoneuropathy of unknown cause whose magnetic resonance imaging study showed enlarged nerve roots (shown in Fig. 32–1). Top, Granuloma formations (nests of macrophages, arrow) surrounded by intense mononuclear inflammatory infiltrates (hematoxylin and eosin). Middle, The CD68 (KP-1) immunoreactivity confirms that the granuloma cells are macrophages (arrow). Bottom, The CD45 (leukocyte common antigen) immunoreactivity indicates that mononuclear cells are lymphocytes. Based on these biopsy results, the patient was diagnosed as having sarcoidosis. See Color Plate

and these have failed to provide a definite answer and informed judgment is made that nerve biopsy may be useful. It should be done by surgeons experienced in taking such tissue and be evaluated in laboratories that are able to evaluate teased fibers, cryostat sections, and paraffin and epoxy sections and have available facilities to do histochemistry, electron microscopy, and morphometry. The availability of pathologists expert in hematopathology and tumor pathology is also a necessary requirement.

100μ

Nerve Biopsy versus Other Neuropathic Evaluations The role of nerve biopsy cannot be discussed without considering other neuropathic evaluations. Evaluations useful in detecting, characterizing, and judging the severity of peripheral neuropathy are the neurologic history (including assessment of symptoms), neurologic deficits, electromyography (EMG) (nerve conduction and needle

736

Pathology of the Peripheral Nervous System

EMG), quantitative sensation testing (QST), quantitative autonomic examination (QAE), and a variety of other laboratory, roentgenologic, and imaging tests. In recent publications, we have assessed the comparative value of these six evaluations (symptoms, deficits, nerve conduction/EMG, QSE, QAE, and nerve biopsy) in the detection, characterization, and assessment of diabetic polyneuropathy.31,41,52 Is the information provided by these evaluations simply confirmatory of neuropathy or complementary, providing useful information for differential diagnosis? The answer is that the tests provide overlapping but different characterizing information useful in differential diagnosis. Symptoms must be assessed because they reflect the subjective experience of patients. Symptoms (1) are unpleasant, painful, disabling, or worrisome to patients; (2) bring patients to doctors for relief; (3) cannot be inferred adequately from other tests; and (4) may convey the patient’s reaction to primary experiences. Neurologic deficit can be assessed as disparate information that the neurologist uses in formulating an opinion as to kind and severity, as a summated score of neurologic deficit (e.g., the Neuropathy Impairment Score), or by a scale of impairments in acts of daily living. Abnormalities of nerve conduction and EMG are known to be sensitive and specific in the detection of peripheral neuropathy in diabetes and in inherited neuropathy, and they appear to correlate with severity. Electrodiagnostic studies provide considerable information about focal or generalized nerve conduction abnormalities and about degenerative and regenerative events of large-diameter fibers. Electrodiagnostic studies can be used to infer fiber degeneration and loss, fiber regeneration, generalized or focal demyelination, collateral sprouting, reorganization of motor units, and such phenomena as fasciculations and cramps. Generally electrodiagnostic studies are not reliable for diagnosing a specific variety of neuropathy, inferring symptoms, predicting global neurologic deficits, detecting and characterizing small-fiber neuropathy, characterizing specific pathologic alterations of fibers, or characterizing interstitial pathologic alteration of nerve. QST may be used to detect and characterize sensory abnormality at specific sites. It is useful in recognizing abnormality of small-diameter sensory fibers not well assessed by present electrophysiologic techniques. QST is especially useful for recognizing small fiber dysfunction and such phenomena as thermal hyperalgesia. QAE is used to detect and characterize autonomic nerve fiber dysfunction of eye, heart, vessel, gut, sphincters, and sweat glands. In a recent study of diabetics with and without neuropathy, we found that nerve conduction velocities (NCVs) were especially sensitive and objective (uninfluenced by the will of the patient) and therefore were useful as minimal criteria for the diagnosis of polyneuropathy. Vibratory detection

threshold using the 4, 2, 1 stepping algorithm with null stimuli and the CASE IV system (WR Medical Electronics, Stillwater, MN) was also a rapid, nonpainful, and accurate method of determining whether polyneuropathy was present. Assessment of symptoms, summated neurologic deficit (e.g., The Neuropathy Impairment Score), QST, and the summated compound muscle action potential of ulnar, peroneal, and tibial nerves were best for assessing staged severity of neuropathy. Nerve biopsy is useful in detecting and characterizing pathologic changes in nerve fibers and Schwann cells (myelin) as described above. Nerve biopsy may also be used to discriminate monophasic from continuing axonal degeneration, axonal atrophy from loss of large fibers, and primary from secondary demyelination. Nerve biopsy can be used to decide whether fiber loss is focal, multifocal, or diffuse. Nerve biopsy is especially useful in recognizing interstitial pathologic alterations. The neuropathy of leprosy is suspected when focal epithelioid and lymphocyte infiltration of the endoneurium is found and by visualization of Mycobacterium leprae organisms in acid-fast stains. In herpes, poliomyelitis, cytomegalovirus, and human immunodeficiency virus infections, the pattern and localization of the inflammation and cell injury may suggest the diagnosis. Inflammation in nerve may be categorized by the level of nerve affected (ventral horn columns, roots, ganglia, plexuses, or nerves), compartment of nerve involved (endoneurium, perineurium, or epineurium), relation to certain cells’ or tissues’ pathologic reaction (vasculitis, inflammatory demyelination, granuloma, or other), and identity of the inflammatory cells. Nerve biopsy is frequently done for the purpose of diagnosing amyloid deposition. Lysosomal storage diseases may be recognized by characteristic organelles in the tissue.25 Various alterations of vessels and patterns of ischemic injury may be recognized in nerve biopsy specimens, and such information is diagnostically useful and has treatment implications. Necrotizing vasculitis of several types may occur. Vasculitic disorders are classified by the vessels involved (arteries, arterioles, venules, microvessels), by the distribution of involvement, by the inflammatory reaction, and by mechanism. Nerve biopsy may show bleeding or characteristic patterns of ischemic injury of the nerve (see Interstitial Pathology of Nerves below). Nerve biopsy may be needed to define the pathologic basis of focal nerve enlargement or enhancement on magnetic resonance imaging (MRI). Nerve biopsy tissue is of limited or no value in explaining such symptoms as paresthesia, pain, cramps, and other manifestations of neural hyperactivity. Study of nerve tissue must be limited to a small sample of accessible tissue, and it can be done only on a single or at most a few occasions.

Pathologic Alterations of Nerves

Often, detailed study is possible only at death on poorly preserved tissue.

Peripheral Nerves That May Be Biopsied Nerve biopsies may be categorized into those done from sites of focal nerve enlargement or enhancement, as seen on MRI, and those done from sites thought to be affected by disease but not known to show enlargement or enhancement. Most nerve biopsies fall into the latter category. For the first group, the focal nerve enlargement and enhancement determine the site of the biopsy. For the latter category, it is important to choose a nerve that is involved with the neuropathy to be studied, contains sufficient tissue for study, is readily accessible, and the removal of which will not cause excessive side effects, considering the information obtained. Biopsy at Sites of Nerve Enlargement or Enhancement Biopsy of nerve should be considered when progressive focal or multifocal neuropathic symptoms and signs appear to be attributable to focal nerve enlargement or enhancement on MRI, the cause of which remains undiagnosed and for which information about diagnosis, prognosis, and treatment depends on a pathologic diagnosis. The putative cause may be focal infection (leprosy), inflammation (inflammatory demyelination, vasculitis, or granuloma), tumor (lymphoma, perineurioma, neurofibroma, Schwannoma, or other), or other (see Figs. 32–1 and 32–2). As for other biopsies, the physician must decide whether the potential benefits outweigh the morbidity, subsequent symptoms and impairments, and costs. This procedure should only be done at medical centers with the necessary surgical experience and clinical and pathologic facilities to perform these studies adequately. Sites for biopsy may include nerve rootlets, portions of spinal ganglia, mixed segmental nerve, and various proximal-to-distal levels of nerve. Because these fascicular biopsies are taken from mixed limb nerves, some increased weakness, sensory loss, and positive neuropathic sensory symptoms may ensue. Fortunately, these complications are usually minimal. Nevertheless, the patient needs to be adequately forewarned, permission must be obtained, and only a minimal amount (but sufficient for pathologic study) of tissue should be procured. Ideally, two to three rootlets or fascicles are taken together with intervening tissue spanning from above the point of enlargement to within the enlargement. Cutaneous Nerve Biopsied without Focal MRI Enlargement or Enhancement In cases of apparent disease but without focal enlargement or enhancement on MRI, the nerves that may be considered for biopsy include the sural, superficial peroneal,

737

saphenous, cutaneous branch of the radial, medial or lateral cutaneous nerves of the forearm, greater auricular, and other nerves. For multifocal neuropathy (e.g., leprosy, necrotizing vasculitis, amyloidosis, or other interstitial processes), we prefer using a leg sensory nerve (the sural or the superficial peroneal nerve) that is known to be affected. We have generally preferred the sural nerve at the ankle level as the nerve of choice for biopsy because it usually has a generous epineurium and usually contains several small arteries or arterioles. Others have favored the superficial peroneal nerve, often combining the procedure of nerve biopsy with muscle biopsy (of the peroneus longus or brevis) (see Chapter 106). We have had concern about biopsying a large piece of the peroneus brevis muscle because it is an important everter of the foot at the ankle. A whole or a fascicular cutaneous biopsy can be taken. We tend to take whole cutaneous biopsies because whole nerves contain much more epineurium, which contains the larger nerve vessels and therefore provides a better sample for recognizing interstitial events. The nerve that is chosen for biopsy should be shown to be affected by the disease process. This can be determined from the clinical findings, study of nerve conduction, or subjective alterations of skin sensation in the cutaneous distribution of the nerve (e.g., asleep-numbness or prickling when the examining hand is passed over the cutaneous distribution of the nerve). Sural Nerve Biopsy. Sural nerve biopsy is performed under sterile conditions in the operating room, with the extremity shaved, scrubbed, disinfected, and draped as in any general surgical procedure. The surgical team wears gowns, masks, and gloves. Surgical instruments must be delicate and in good working order. Fine suture material, 4-0 or 5-0, and atraumatic needles are used to minimize foreign body reactions. Magnifying glasses or a surgical dissecting microscope are essential for visualizing the nerve fascicles and for accurate detection of bleeding vessels. Because most patients are ambulatory outpatients, preanesthetic sedation is avoided. We prefer local infiltration anesthesia with 0.5% lidocaine (Xylocaine) without epinephrine. General anesthesia is rarely used, being reserved for the hyperactive young child. The sural nerve is most readily exposed at the ankle level. The patient lies prone on the operating table with the ankle, slightly everted, resting on a pillow so that the foot assumes a 90-degree angle with the leg. This places the sural nerve at normal tension. When the calf is compressed, the lesser saphenous vein becomes distended and may become visible or palpable in the trough between the external malleolus and the Achilles tendon. The saphenous vein is the most reliable anatomic landmark for locating the sural nerve, which lies immediately adjacent or deep to the vein at this level.

738

Pathology of the Peripheral Nervous System

Eight to 10 mL of 0.5% lidocaine is infiltrated into the skin and subcutaneous tissue, extending from just behind the external malleolus for a distance of 10 cm proximally. The incision is placed to overlie the course of the saphenous vein. As the tough, semitransparent Scarpa’s fascia is exposed, the saphenous vein can be seen beneath it. When the Scarpa’s fascia directly over the vein is divided and the vein is retracted gently, the sural nerve comes into view, most commonly on the medial and usually deep side of the vein. The bleeding points are then ligated. Occasionally, a tributary vein to the lesser saphenous at the malleolus requires division and ligation because it crosses over the sural nerve. When the lesser saphenous vein goes into spasm, it may be mistaken for the sural nerve because it becomes white like a nerve. This error can be avoided by knowing that the tributaries of the vein come off of the lesser saphenous vein at right angles like the branches of an oak tree, whereas the branches of nerve come off at acute (much smaller than 90-degree) angles, like the branches of an inverted elm tree. After exposure of the sural nerve for 10 cm, a small amount of local anesthetic is injected directly into the sural nerve just proximal to the apex of the incision (Fig. 32–3). A 5-0 silk suture with a curved needle fused to the suture is transfixed through the substance of the nerve at a proximal point of the exposed nerve but just distal to the local anesthetic injection site. The suture is not tied. The free ends of the suture are then used as a retractor. The nerve is then transected just above the suture. Using the free ends of the suture between thumb and finger and avoiding traction (or applying only very light traction), the nerve is undercut with curved scissors, excluding most of the interstitial fat and vessels, for a distance of approximately 3 cm. The bent hook of a 10-mg weight is impaled into the distal end of the mobilized nerve specimen at a point 2.5 to 3 cm distal to the proximal transection. The nerve is then transected just distal to the weight. The nerve specimen is lifted using the retracting suture and is hung directly (in the operating room and without any delay) into 2.5% glutaraldehyde in buffer (see Histologic Methods below) for 5 minutes and then is transferred to 4.0% paraformaldehyde in buffer at 10° C and pH 7.4 (see Fig. 32–3). The next 2.5- to 3-cm segment of nerve (a direct continuation of the most proximal specimen) is similarly transfixed with a suture, undercut, weighted, and hung into 2.5% glutaraldehyde in buffer, the final solution isotonic with plasma and at 10° C and pH 7.4. The final 2.5- to 3-cm piece is immersed into just melting isopentane in a metal beaker suspended on liquid nitrogen. This is stored at ⫺80° C for later studies or is used for cryostat sections. The nerve itself is never tweaked or grasped with forceps. Its edges (the epineurium) may be grasped with fine-pointed forceps. Great care is also taken to avoid stretching the nerve. Removal of a single fascicle as compared to removal of the whole nerve is best performed using a surgical dissecting

FIGURE 32–3 Surgical technique of sural nerve biopsy (see text for details). (From Dyck, P. J., and Lofgren, E. P.: Method of fascicular biopsy of human peripheral nerve for electrophysiologic and histologic study. Mayo Clin. Proc. 41:778, 1966, with permission.)

microscope. The epineurium is slit along the length of the nerve trunk, exposing individual or groups of fascicles. A fascicle (or two or three fascicles), usually one located on the posteromedial aspect of the nerve, is selected for biopsy. A 7-0 silk suture is passed through the fascicle at the upper end of the incision and held as a loop retractor. The fascicle is transected just proximal to the suture. By dissecting the fascicle free of the epineurium and occasionally dividing small bridging fascicles, several fascicular lengths of nerve are obtained. Each of these is suspended in the fixative by the silk suture at one end, and a small weight is hooked into the other end (see Fig. 32–3). After meticulous hemostasis, Scarpa’s fascia is reapproximated, with care not to pass the suture around the remaining nerve. The skin is approximated with subcuticular stitches. A dry dressing is applied and held in place with a 10-cm elastic bandage. We encourage the patient to walk away following the procedure rather than use a wheelchair. He or she is instructed not to engage in sports in the ensuing 3 to 4 days. Analgesics generally are not necessary for postoperative discomfort, but the patient is forewarned that

Pathologic Alterations of Nerves

a surge or twinge of pain may be felt if the sural nerve is stretched, by walking or stooping forward. The dressings may be changed on the day after the procedure. The elastic bandage is worn for protection for 2 weeks. Biopsy of the Superficial Peroneal Nerve and Other Nerves. The superficial peroneal nerve emerges from between the peroneus longus and brevis and passes through the deep fascia in the distal third of the lateral leg, innervating the cutaneous distribution of the lower lateral leg and dorsum of the foot. Its size is comparable to that of the sural nerve. The surgical technique of biopsy of this and other nerves is similar to that described for the sural nerve.

Symptoms after Nerve Biopsy The symptoms from sural nerve biopsy are quite variable among healthy subjects and patients with neuropathy (see Chapter 30 in the second edition of this textbook). Patients with severe pain loss in the feet and legs generally experienced no adverse symptoms of nerve biopsy. Among patients biopsied for neuropathy, 60% had no symptoms 1 year after sural nerve biopsy, 30% had intermittent mild persisting symptoms that then disappeared, and 10% had greater degrees of pain or paresthesia. It is possible that some of the discomfort experienced by patients was related to the neuropathic process and not to the biopsy. One of the authors underwent fascicular biopsy of the sural nerve at the ankle, without injection of local anesthetic directly into the nerve.49 There was no discomfort until transection of the nerve fascicle. Immediately with nerve transection, a sharp, stinging, burning discomfort occurred in a region just below the lateral malleolus and spreading to the Achilles tendon. Although severe, the pain lasted only a few seconds. This pain could have been avoided by local anesthetic injection of the proximal nerve. In the first 3 or 4 postoperative days there was soreness at the site of the incision, especially when the tissues were stretched. On stretching the nerve, as in bending, sharp electric shock–like pain was experienced in the lateral heel. Altered sensibility and a mild raw, burning discomfort were experienced over and below the lateral malleolus and over the Achilles FIGURE 32–4 Anterior approach to spinal cord removal at postmortem examination. Dotted lines indicate planes of saw cut adjusted to shapes of different levels of vertebral column. A, Cervical. B, Thoracic. C, Lumbar. (From Okazaki, H., and Campbell, R. J.: Nervous system. In Ludwig, J. [ed.]: Current Methods of Autopsy Practice, 2nd ed. Philadelphia, W. B. Saunders, p. 95, 1979, with permission.)

739

tendon. This was noticeable especially on prolonged sitting or on taking the first steps after sitting. The pain vanished completely on walking. When the sural nerve was palpated at the level of the proximal transection site, no local discomfort was experienced. In the cutaneous distribution of innervation (below the lateral malleolus and over the lower Achilles tendon), a mild prickling, unpleasant burning sensation was produced. When the affected dermatome was touched, some sensation was elicited. Although the threshold for touch and pinprick was greatly heightened, stronger stimuli produced a diffuse, unpleasant, prickling burning in the affected region. These symptoms gradually faded away, returning intermittently with less intensity over several years. At 5 years the symptoms had disappeared, although touching the affected dermatome still elicited a slightly unpleasant and altered sensation. After approximately 10 to 15 years, sensation had returned to normal.

POSTMORTEM SPINAL CORD AND NERVE REMOVAL AND PROCESSING The spinal cord, roots, spinal ganglia, and a variable extent of segmental nerves can be removed post mortem by a posterior or anterior approach. The following description is Okazaki and Campbell’s121 modification of Kernohan’s method. The anterior approach is generally used because it is preferred by morticians, is more rapid, and allows removal of the spinal cord with roots, spinal ganglia, and plexus nerves in continuity. After evisceration, the block supporting the head is removed and placed under the middle of the thoracic spine to straighten it. A transverse cut is made through the first or second thoracic vertebra using an oscillating saw. With the oscillating saw, cuts are made along the sides of the vertebra, holding the blade at the angles shown in Figure 32–4. The cut is made just through the bone to avoid injury to roots or spinal cord. The angle changes from almost vertical to almost horizontal as one passes from the cervical through the thoracic to the lumbar spine. Okazaki suggests short cuts on both sides of the vertebra, beginning in the thoracic region

740

Pathology of the Peripheral Nervous System

before proceeding to more caudal levels. With the body properly positioned and with adequate cuts, the vertebral bodies will spring forward away from the spinal cord toward the prosector. The freed spine can be grasped with the noncutting hand as the cuts are extended caudally. As one proceeds to the lumbar region, the muscles must be cut away from the spine, making certain that one does not transect the emerging nerve roots. The lumbar spine is transversely cut at the L4-L5 interspace with a curved short knife. The L5 vertebral body is then cut away using the oscillating saw, chisel, and mallet. The nerve roots at the lumbosacral interspace can be cut across or can be dissected out after the bone is chipped away with the oscillating saw and rongeur. The exposed spinal cord and cauda equina, covered by the dura mater, can be lifted out of the spinal canal. If the bony floor of the spinal ganglia has been chipped away, the spinal ganglia, segmental nerves, and lumbar and sacral plexus can be taken as one. A suture may be used to transfix one of the nerve roots (e.g., L5, identified by its point of exit) to be able to correctly identify specific roots.

HISTOLOGIC METHODS For Nerve Biopsy Sural or other nerves taken at biopsy are suspended in fixative in the operating room (described above). On returning to the laboratory, excess fat is trimmed away. The 2.5-cm length for embedding into paraffin may be initially fixed in 2.5% glutaraldehyde for 5 to 10 minutes, then immersed in 4% paraformaldehyde in buffer. This tissue is fixed overnight, taking care that there is a large excess of fluid relative to tissue (ⱖ25:1). If the fixative solution becomes bloody on immersion of the nerve specimen, it is replaced by fresh fixative solution. After fixation and washing, 3- to 4-mm-long collars of tissue are cut transversely (to the longitudinal axis of fascicles on a sheet of dental wax) using a degreased razor blade. The collars are carefully embedded into paraffin so that transverse and longitudinal sections will result (Fig. 32–5). These paraffin sections are stained with hematoxylin and eosin, trichrome, Luxol fast blue, periodic acid–Schiff, methyl violet, Congo red, and an iron stain (e.g., Turnbull blue). Other sections are histochemically reacted for CD45 (leukocyte common antigen), for CD68 (macrophages), or for other markers. Other histochemical reactions for amyloid proteins, Schwann cell proteins (e.g., S100), epithelial membrane antigen, vascular wall markers (smooth muscle actin), and others may be obtained depending on need. The 2.5-cm length to be used for preparation of teased fibers and for embedding into epoxy is fixed in 2.5% glutaraldehyde in 0.025 mol/L cacodylate buffer at pH

7.38 and 10° C. The solution is isotonic to prevent uneven shrinkage of fascicles and axons, and we fix this tissue at 10° C—a compromise to avoid the depolymerization of microtubules (MTs), which occurs at 0° to 4° C, and to minimize autolysis, which is greater at room temperature than at low temperatures. In the operating room or shortly after return to the histology laboratory, the length of nerve fixed in buffered glutaraldehyde must be divided into two approximately equal lengths. For full-thickness biopsies this transection occurs 1 to 2 hours after surgery; for fascicular biopsy specimens it occurs 20 to 45 minutes after surgery. At the time of this transection, excess fat is removed and the specimen to be used for teased fibers is split into individual fascicles using a razor blade with a single cutting edge and on dental wax (see Fig. 32–5). The duration of aldehyde fixation for the sural nerve specimens to be embedded into epoxy varies from 4 to 24 hours depending on the size of the specimen. One collar is left undivided and another is divided into individual fascicles and embedded into epoxy for transverse sections. Another collar is embedded for longitudinal sections (see Fig. 32–5). The blocks are washed six times, for 5 minutes each, in isotonic cacodylate buffer; osmicated for 1.5 to 3 hours, depending on the tissue block size; and dehydrated, infiltrated, and embedded into epoxy. In preparing tissue blocks, one avoids splitting fascicles because the architecture of the nerve is better maintained with the perineurium left intact. Special molds are used to get true transverse and longitudinal sections (see Fig. 32–5). Semithin sections (1.0 ␮m) can be viewed with a phase contrast microscope without further staining or with a light microscope after staining with methylene blue (or toluidine blue) or paraphenylenediamine. Nerve fascicles are surrounded by perineurium, which, in concert with endothelial cells, maintains a special endoneurial environment. Dyck and colleagues found in a series of reported studies that immersion fixation with hyperosmolar glutaraldehyde fixation resulted in serious distortions not only of the transverse cross-sectional shape of fascicles but also of nerve fibers.50 Such fixation resulted in flattened, crescentic, and fluted shapes of the transverse profiles of fascicles and fibers—markedly different than the round profile of these structures in cryostat sections (Fig. 32–6).

For Postmortem Tissue Ideally, nerve tissue should be obtained as soon after death as possible. Very occasionally it may be possible to do this within 1 to 4 hours of death. Usually it is obtained within 6 to 12 hours of death. It is advisable to obtain separate permission to remove limb nerves. The medical record, especially the neurologic and electromyographic examination, should be reviewed in order to take tissue from appropriate anatomic sites. Care should be taken to sample known

741

Pathologic Alterations of Nerves

Epoxy embedding Whole nerve A1T1 Fascicles Proximal Distal

For transverse sections

For transverse sections A2T1, A2T2, ... A2Tn

Rinse Osmicate Dehydrate Infiltrate

Embed into epoxy For longitudinal sections

For longitudinal sections A3L1, A3L2, ... A3Ln Teased fibers see specific section

Rinse Osmicate Glycerinate Tease

See Fig. 32.6

Paraffin embedding Whole nerve Rinse Dehydrate Infiltrate

FIGURE 32–5 Steps used in processing nerve biopsy tissue for epoxy and paraffin embedding and for preparation of teased fibers as described in text.

Stains Methylene Blue Phenylenediamine Unstained - phase contrast Uranyl acetate and lead citrate for tem

Embed into paraffin

Stains Hematoxylin + Eosin Luxol Fast Blue - Periodic Acid Schiff Methyl Violet Congo Red Turnbull Blue Other Immunohistochemistry CD68 (KP-1) CD45 (LCA) Other

Freeze a block in chilled isopentane

anatomic segments and different proximal-to-distal levels of the peripheral nervous system and their central nervous system (CNS) extensions. It may be important to sample levels for which control values are present. For biochemical study, portions of peripheral nerve, of brain, and of parenchymatous organs should be obtained quickly after death and preserved in liquid nitrogen. Even with postmortem material, tissue may be obtained for teased fiber preparations. For this purpose, the nerve selected should be one that is affected by the disorder and one for which normative morphometric measurements are available.

The many histologic methods that are used for postmortem peripheral nerve tissue can be found in such books as the one by Romeis130 and the manual of the Armed Forces Institute of Pathology.108 A few histologic procedures that are especially helpful are described here. Further processing depends on what use will be made of the tissue. Spinal cord, spinal ganglia, and autonomic ganglia may be embedded by segment or by ganglion into celloidin so that serial sections can be cut and systematically sampled ones can be used for morphometric analysis (see Morphometry of Peripheral Nerves below). Celloidin is ideal for this purpose because heat is not employed and

742

Pathology of the Peripheral Nervous System

FIGURE 32–6 Transverse sections of human sural nerve fixed by immersion in buffered glutaraldehyde to show the reproducible alteration in fascicular and myelinated fiber shape when hyperosmolar fixation is used. The nerve on the left has been fixed in isosmolar and that on the right in hyperosmolar fixative. The fascicles and fibers assume flattened, crescentic, and other shapes when fixed in hyperosmolar solutions. (From Dyck, P. J., Low, P. A., Sparks, M. F., et al.: Effect of serum hyperosmolality on morphometry of healthy human sural nerve. J. Neuropathol. Exp. Neurol. 39:285, 1980, with permission.)

shrinkage appears to be minimal. Cresyl violet, which stains nuclei and cytoplasm well, is employed for such studies. Other portions may be embedded into paraffin, methacrylate, or epoxy. For nerve trunks one may wish to fix and process serial blocks differently. The first block may be fixed in aldehyde, embedded in paraffin or methacrylate, and stained so that it can be evaluated for interstitial pathologic abnormality. The next block is cut up into smaller pieces and is additionally fixed in osmium tetroxide and embedded in epoxy to be able to evaluate semithin and thin sections. The third block may then be used for study of teased fibers. Serial blocks may be evaluated in this way along the length of the nerve.

TEASED FIBER History The use of teased single or small strands of myelinated fibers (MFs) to study anatomic features and pathologic abnormalities dates back more than 100 years. Early investigators of peripheral nerve—Remak, Ranvier, Kölliker, Renaut, Gombault, Retzius, and Ramon y Cajal—illustrated both normal and abnormal features of MFs using teased fibers.90,128,130 Even though the preparation and analysis of teased fibers are time consuming, teased fiber analysis has been reintroduced, especially for evaluation of MF

743

Pathologic Alterations of Nerves

anatomy153 and neuropathology. Introduction and use of improved fixatives,131 simpler techniques of preparing teased fibers, a descriptive classification of pathologic abnormalities,48 and systematic sampling of fibers and statistical treatment of results48 have permitted wider and more diverse use and validation of results. Analysis of teased fibers is so helpful in characterizing fiber changes that we perform it on all nerve biopsies we assess.

Uses The procedure (1) permits evaluation of consecutive internodes or segments of the same myelinated nerve fiber over long distances; (2) may be used to establish the frequency of certain neuropathologic fiber abnormalities and hence detect neuropathic abnormality generally with more sensitivity than by use of fixed sections; (3) may serve to show whether there is ongoing fiber degeneration; (4) may indicate whether fiber branching or sprouting is occurring; and (5) allows the performance of ultrastructural studies at preselected sites of focal pathologic abnormality, if fibers have been embedded in epoxy.35,146 Obviously it cannot be used to study unmyelinated fibers.

Histologic Technique Sural nerve biopsy, described in a previous section, should be performed with care to avoid crush and stretch. The sural nerve specimen, with the proximal end cut at right angles and the distal end cut obliquely, with a suture attached to one end and a light stainless steel weight attached to the other end, is suspended in freshly prepared 2.5% glutaraldehyde fixative in 0.025 mol/L cacodylate buffer, pH 7.38, at a temperature of 10° C for 20 minutes (single fascicle) or for 60 to 120 minutes (full-thickness nerve). The duration of aldehyde fixation is critical. Insufficient aldehyde fixation will later be associated with breaks or separation at nodes of Ranvier and at Schmidt-Lanterman incisures. Excessive fixation may make it difficult to tease apart single fibers or small strands of fibers and may cause breaking and splitting of fibers. After aldehyde fixation, the full-thickness nerve specimen is divided into its component single fascicles using a razor blade on dental wax. Epineurial fat is cut away. The fascicles are washed (six times, each for 5 minutes, in isotonic cacodylate buffer) and osmicated for 1.5 to 3 hours. The fascicles are carried through 45%, 66%, and 100% glycerin each for 24 hours and at 45° C.

Teasing Technique Our (Dyck, Lais, and Engelstad) method of preparing teased fibers is illustrated in Figure 32–7. On lightly glycerinated glass slides and with curved, pointed forceps and using a dissecting microscope, epineurium and perineurium are stripped away from the fascicular bundle of nerve fibers

(endoneurium). Next, small strands of nerve fibers are torn from the endoneurium. From these strands, single fibers or a bundle of a small group of fibers are pulled away from the main strand, grasping the proximal ends of each strand with curved pointed forceps. For a right-handed person, the left forceps grasps the proximal end of the main strand and remains motionless while the right one, grasping the small strand, traces an inverted U-shaped pathway. The tip of the forceps is kept in constant contact with the glass slide to prevent the fiber from curling up around it. Teasing is always performed from the proximal end of the fascicle so that branches are not torn off. A clean slide is laid adjacent to the right of the slide on which the teasing has been done. The proximal end of each separated fiber strand is grasped with forceps and slid onto the clean slide through a minute drop of glycerin; again, the tip of the forceps is kept constantly touching the glass. As the middle of the new slide is reached, the points of the forceps are allowed to spread apart. The friction of the fiber strand on the glass will straighten the fiber and cause it to pull away from the tips of the forceps. Fifty teased fiber strands are placed side by side in the center of each of two glass slides, allowing for an analysis of 100 teased fibers in most cases. If the proper technique is used, the proximal ends of the fibers will always be oriented in the same way, the fibers will be straight, and there will be a minimum of glycerin with them. The slides are air dried overnight in a warm oven. With observation under the dissecting microscope, a drop of mounting medium is applied to the coverslip, so that the apex of this drop is at the center of the fiber. The coverslip is dropped onto the slide. As the drop spreads outward, it will carry most of the glycerin at its moving front, displacing it to the edges of the coverslip. The coverslip edges may be sealed with nail polish. Using these methods on suitably prepared tissue, an experienced, diligent person can prepare 150 to 400 teased fibers per day. An elegant display of teased fibers laid side by side is shown in Figure 32–8 for a control nerve. A nerve with most fibers undergoing early axonal degeneration is shown in Figure 32–9. Figure 32–10 is from a nerve with most fibers undergoing late degeneration. The nerve in Figure 32–11 shows many large tomacula, with myelin reduplication typical of hereditary neuropathy with a tendency to pressure palsy. It is important to note that these figures represent only a quarter or so of the length of the fibers teased but, because they are isolated and lie flat on glass, each fiber can be visualized at all points along its length.

Fixation and Teasing Artifacts It is important to recognize that various artifacts, as shown in Figure 32–12, can be produced by the method described here. Excessive stretching of the nerve during biopsy or of

744

Pathology of the Peripheral Nervous System

FIGURE 32–7 From left to right and from top to bottom, consecutive steps in fiber teasing: a fascicle of nerve, fixed in glutaraldehyde and osmium tetroxide, lying in a pool of glycerin on a glass slide; proximal ends are grasped and fascicles are pulled apart; epineurium and perineurium are stripped off; strands of fibers are pulled apart; from separated strands of nerve, a single teased fiber is slid onto an adjacent slide as described in the text; and teased fibers in place under coverslip.

the fiber during teasing, when the tissue is inadequately fixed, produces separation at nodes of Ranvier (see Fig. 32–12, frame 2) and at Schmidt-Lanterman incisures. In addition, the fiber may assume a varicose appearance that mistakenly may be assumed to be due to pathologic abnormality. Separation at nodes and at Schmidt-Lanterman incisures is particularly common when this method of teasing is used on tissue fixed with osmium tetroxide alone; with correct aldehyde fixation, the artifact is not prominent. Any morphologic change observed at the ends of teased fibers

should be ignored because these may represent damage by the forceps. With overfixation with glutaraldehyde, splitting of fibers may occur (see Fig. 32–12, frames 3 through 5).

Sampling Before beginning to tease the specimen, the technician must decide how many fibers per fascicle are to be assessed and how representative sampling is to be

Pathologic Alterations of Nerves

745

FIGURE 32–10 Low- (upper) and intermediate- (lower) power light microscopic views of teased fibers showing late axonal degeneration (condition E). All fibers on this slide are undergoing degeneration. FIGURE 32–8 Low- (upper) and intermediate- (lower) power light microscopic views of about one-quarter of the length of teased fibers laid side by side to illustrate that the method of teasing allows fibers to be visualized for long distances along their length by our technique of preparation. The teased fibers are from the sural nerve of a healthy subject using fixatives and approaches outlined in this chapter. All fibers would be graded as normal (condition A). (Modified from Dyck, P. J., Dyck, P. J. B., Giannini, C., et al.: Peripheral nerves. In Graham, D. I., and Lantos, P. L. [eds.]: Greenfield’s Neuropathology, 7th ed., Vol. 2. London, Arnold Publishing, p. 551, 2002.)

FIGURE 32–9 Low- (upper) and intermediate- (lower) power light microscopic views of teased fibers from the sural nerve of a patient with necrotizing vasculitis showing severe axonal degeneration (condition E). (Modified from Dyck, P. J., Dyck, P. J. B., Giannini, C., et al.: Peripheral nerves. In Graham, D. I., and Lantos, P. L. [eds.]: Greenfield’s Neuropathology, 7th ed., Vol. 2. London, Arnold Publishing, p. 551, 2002.)

achieved. In health, a fascicle of sural nerve may contain from 500 to 2500 MFs. The sensitivity and reliability of the estimate of pathologic abnormalities and the variability of the estimate will depend on the rate and distribution of abnormality and on the sampling procedure. The error will decrease with increasing numbers of fibers sampled. For clinical and many research applications, at least 100 fibers/nerve should be sampled. This number was chosen to get representative results, sufficient reliability for detection of neuropathic abnormality, and a reasonable expenditure of time and money. For certain research

FIGURE 32–11 Low- (upper) and intermediate- (lower) power light microscopic views of teased fibers showing typical tomacula (condition G) in closely applied teased fibers from a patient with hereditary neuropathy with tendency to pressure palsy. (Modified from Dyck, P. J., Dyck, P. J. B., Giannini, C., et al.: Peripheral nerves. In Graham, D. I., and Lantos, P. L. [eds.]: Greenfield’s Neuropathology, 7th ed., Vol. 2. London, Arnold Publishing, p. 551, 2002.)

746

Pathology of the Peripheral Nervous System

FIGURE 32–12 Teased fiber preparations (⫻380). 1, Five consecutive portions along the length of a teased fiber. Nodes of Ranvier are seen in first and fifth strips. Unstained Schwann cell nucleus is seen in the third strip. 2, Three consecutive portions along the length of a teased fiber, showing excessive separation at nodes of Ranvier (top and bottom strips), separation at Schmidt-Lanterman incisures, and undulating thickness of fiber, all caused by improper fixation (osmium tetroxide alone) for this method of teasing. 3, Two adjacent teased fibers with the upper showing patchy loss of myelin, an artifact encountered during teasing when excessive hardening has occurred with glutaraldehyde fixation. 4, A more severe artifact in which the lower right half of the fiber has been split and pulled away during teasing from tissue that is excessively hard. 5, Artifact in the middle of strip of fiber caused by grasping with forceps. These artifacts are inevitable in this method and should be ignored. This fiber also shows the effect of excessive stretching; the undulating appearance comes from excessive stretching before adequate fixation.

purposes, as many as 500 fibers/nerve might be prepared. It may be preferable to increase the number of nerves evaluated as compared to increasing the number of teased fibers evaluated per nerve (beyond 100). Approaches to sample teased fibers randomly have generally not been

used. We suggest systematic sampling by approaches described here. This is accomplished by initially dividing the fascicle into 10 thick strands of approximately equal thickness. Each of these thick strands is then subdivided into five intermediate-thickness strands. From the right-hand

747

Pathologic Alterations of Nerves

side of each of these 50 strands, two small strands (without selection) containing one, several, or no (empty) MFs is teased. For certain purposes, it may be desirable to assess the frequency of abnormalities in teased fibers taken from the edges of fascicles versus those from the center (e.g., in ischemic neuropathy). To do so, the outer one sixth of the endoneurial portion is stripped from the right and then the left side of the fascicle. The fascicle is then rotated 90 degrees on its long axis, and again the outer one sixth is stripped from the left and right sides. The four strands stripped from the edges are then divided as described in the preceding paragraph so that 100 fibers are systematically sampled. Next, the center of the fascicle is subdivided in order to grade 100 fibers. A direct comparison of the frequencies of pathologic abnormalities from the edges and centers of fascicles is then possible.

Evaluation of Teased Fibers Assuming that teased fibers have been prepared as outlined in the preceding sections, four types of evaluation are possible. In the first, the percent of teased endoneurial strands containing MFs or their breakdown products can be estimated. In control nerves MFs will be present in most if not all of the teased nerve strands. This evaluation therefore provides a rough estimate of the density of MFs. In the second, the frequency distribution of descriptive pathologic alterations of MFs is estimated. In the third, internode length (IL) and caliber are assessed. In the fourth, ultrastructural alterations are assessed at predetermined sites of teased fibers. The second, third, and fourth approach are discussed below.

FIGURE 32–13 Drawings of consecutive lengths, from top to bottom, of teased myelinated fibers. A, Fiber in condition A as described in text. B, Fiber also in condition A, but note that the internode lengths are short compared with fiber shown above. This may be a regenerated fiber; however, because of the considerable variability in mean internode lengths between fibers of a given diameter even in healthy nerve, it is not possible with certainty to designate this fiber as being regenerated.

Classification of Teased Fibers Dyck and co-workers introduced the following classification of descriptive teased fiber pathologic abnormalities38 to provide objective criteria of teased fiber changes observed based solely on light microscopic observation without need for sophisticated measurement and independent of views related to mechanism of fiber pathology. The pathologic abnormalities that were codified were ones that had been encountered in human and experimental neuropathy. Because these abnormalities had also been seen, albeit at low frequencies, in nerves of healthy subjects, it is incorrect to declare a nerve “abnormal” simply by the occurrence of these abnormalities. Whether the nerve is abnormal or not depends on whether rates of teased fiber abnormalities exceed what is expected (from study of control nerve) for the nerve and the age of the patient. Condition A: Normal Fiber. This is teased fiber of normal appearance, ignoring criteria of IL and of internode diameter (ID) (Fig. 32–13). Myelin is not more irregular than that encountered in most MF internodes of control nerves. This judgment assumes that the observer is aware that myelin irregularity varies with the preparation used, the species, the nerve assessed, and age. In the young, myelin is more regular than it is in the old. The average thickness of myelin of the internode with the thinnest myelin is 50% or more of that of the internode with the thickest myelin. No paranodal or internodal segmental demyelination is seen. Condition B: Myelin Wrinkling. This is teased fiber with excessive irregularity, wrinkling, and folding of myelin but with the other features of condition A (Fig. 32–14).

748

Pathology of the Peripheral Nervous System

FIGURE 32–14 Drawing of consecutive lengths, from top to bottom, of a teased myelinated fiber in condition B as described in text.

This judgment assumes that the observer is familiar with the degree of myelin regularity considering species, nerve, site, and age. Condition C: Segmental Demyelination. This is teased fiber with a region or regions of paranodal or internodal segmental demyelination with or without myelin ovoids or balls in the cytoplasm of the associated Schwann cells* (Fig. 32–15). Thickness of myelin of the internode with the thinnest myelin is 50% or more of that of the internode with the thickest myelin. Myelin of internodes may be regular or irregular. Condition D: Segmental Demyelination and Remyelination. These are teased fibers with a region or regions of paranodal or internodal segmental demyelination with or without myelin ovoids in the cytoplasm of the associated Schwann cells and with remyelination. Thickness of myelin of the internode with the thinnest myelin is less than 50% of that of the internode with the thickest myelin (remyelination). Myelin of internodes may be regular or irregular (Fig. 32–16). Condition E: Axonal Degeneration. This is a teased strand of nerve tissue with linear rows of myelin ovoids and balls at the same stage of degeneration along the length of the teased fiber (Fig. 32–17). Fibers can be segmented into large ovoids, as seen in early axonal degeneration, or into widely separated small myelin balls, as is seen in late axonal degeneration.

*In demyelination, as judged by the high-dry objective of the light microscope, no myelin can be recognized. From our previous evaluation of teased fibers under the electron microscope, approximately four lamellae of myelin can be recognized as a thin dark line under the light microscope except when there is excessive tissue around the fiber. By paranodal demyelination it is implied that the site of the node of Ranvier is recognized and that the nodal gap is increased beyond that seen in normal fibers. In internodal demyelination the entire former internode is demyelinated.

Condition F: Remyelination. This is teased fiber without a region or regions of segmental demyelination but with excessive variability of myelin thickness among internodes. Thickness of myelin of the internodes with the thinnest myelin is less than 50% of that of the internode with the thickest myelin (Fig. 32–18). Myelin of internodes may be regular or irregular. Condition G: Tomacula. This is teased fiber with excessive variability of myelin thickness within internodes as a result of myelin reduplication to form “globules” or “sausages” (Fig. 32–19). Although the myelin thickness in the regions of the globule42 is greater, it probably merely represents a process of infolding and reduplication of myelin rather than excessive production of myelin. Frequently, some regions of the fiber have excessively thick myelin, whereas other regions have excessively thin myelin. Condition H: Regeneration after Axonal Degeneration. This is teased fiber of normal appearance as described in condition A, but in which there are myelin ovoids or balls contiguous to two or more internodes (Fig. 32–20). Condition I: Proximal Segmental Demyelination and Distal Axonal Degeneration. This is teased fiber having several proximal internodes or parts of internodes with or without paranodal or internodal segmental demyelination and, distal to these, a linear row of myelin ovoids or balls (Fig. 32–21). This condition implies axonal dystrophy proximally and distal degeneration. Empty Nerve Strands. These are strands of teased fibers without myelin or myelin breakdown products and without another classifiable fiber condition. They are the by-products of fibers after they have undergone axonal degeneration. These strands are tallied but are not included in the total of classifiable fibers.

FIGURE 32–15 Drawing of consecutive lengths, from top to bottom, of a teased myelinated fiber in condition C as described in text. Note that the myelin thickness is not as variable as it is in condition D.

FIGURE 32–16 Drawings of consecutive lengths, from top to bottom, of a teased myelinated fiber in condition D as described in text. Note that the ovoids of the internode in which the myelin has degenerated are much smaller in diameter than the large ovoids that form initially in condition E. Note also regions of segmental demyelination and the considerable variability of myelin thickness resulting from the presence of “old” and “new” internodes.

A

FIGURE 32–17 Drawings of consecutive lengths, from top to bottom, of a teased myelinated fiber with condition E as described in text. A, Segmentation into large ovoids. B, Mixture of myelin ovoids and balls of intermediate size. C, Clusters of small myelin balls occurring at widely separated points.

B

C

750

Pathology of the Peripheral Nervous System

A

B

C

Clinical Usefulness of Teased Fiber Assessment Teased fiber assessment is a valuable component of the diagnostic sural nerve biopsy assessment. The evaluation may answer these questions: • Is neuropathy present? • What type of fiber degeneration is taking place?

FIGURE 32–18 Drawings of consecutive lengths, from top to bottom, of three teased myelinated fibers in condition F as described in text. The teased fiber in A shows a single intercalated internode, a not-uncommon finding in fibers from healthy nerve.

• Is there active axonal degeneration and what is its frequency? • Is there evidence of repair? • Is there evidence of sprouting? • Is there evidence of axonal atrophy and secondary segmental demyelination, axonal swellings, tomacula, or focal interstitial effects on fibers? • Is demyelination primary or secondary?

FIGURE 32–19 Drawing of consecutive lengths, from top to bottom, of a teased myelinated fiber in condition G as described in text.

Pathologic Alterations of Nerves

751

FIGURE 32–20 Drawing of consecutive lengths, from top to bottom, of a teased myelinated fiber in condition H as described in text. This condition implies regeneration of a myelinated fiber. This condition is not to be confused with that of the teased fiber that has a single osmiophilic droplet in Schwann cell cytoplasm outside of myelin, which is discussed subsequently as the Elzholz body (␮ granule of Reich).

Comparison of the Frequency of Teased Fiber Abnormalities Determining the percent frequency of abnormality (of 100 teased fibers of a sural nerve) allows comparing teased fibers in disease specimens versus those of control specimens. For the comparisons to be valid, nerves should be taken from the same level of nerve, prepared by the same histologic and teased fiber approaches, and

A

FIGURE 32–21 A, Drawing of consecutive lengths, from top to bottom, of a teased myelinated fiber in condition I as described in text. This type of fiber change is seen typically several days after and at the site of crush. B, The same fiber as shown above after repair has occurred—this would be graded as condition F.

B

graded without the observer knowing whether each came from the disease or control specimen (masked grading). This is readily done by random and coded assignment of slides to be graded with the observer masked as to the identity of the slide. Later, the mean percentage of pathologic abnormalities from disease and control specimens can be compared using standard statistical tests.

752

Pathology of the Peripheral Nervous System

Sources of Variability in Grading Pathologic Abnormalities The degree of myelin irregularity is hard to define and grade objectively because it varies with age, species, nerve, and level of nerve. Assessment of the frequency of condition B may therefore vary considerably among observers. Another source of variability may relate to assessment of nodal length. A wide nodal gap resulting from excessive stretching of improperly fixed nerve may be spuriously judged as segmental demyelination (condition C). A careful preevaluation of myelin regularity and nodal length variability in control nerves, fixed and prepared by similar techniques, helps minimize this problem. A further source of variability relates to distinguishing condition C from condition F and from condition D. A careful reading of the description of the criteria and use of an intermediate high-dry objective should reduce this source of variability. Another source of variability relates to discriminating a late condition B from an early condition E. A common source of variability may relate to lack of detection of late condition E. Small myelin ovoids distributed at great distances from each other are readily missed by inexperienced observers. Large onion bulbs with many Schwann cell processes can also be mistaken for empty nerve strands. Frequency of Graded Pathologic Abnormalities in Control Nerves Table 32–1 provides normal values based on evaluation of teased fibers from sural nerves of paid volunteers. These persons were neurologically examined to ensure that they did not have neurologic disease.

Internode Length and Diameter Measurement of IL and ID of teased fibers is of interest because these attributes may provide an insight into altered myelin internode development and previous demyelination and remyelination. To illustrate, in diphtheritic and lead neuropathy, segmental demyelination and remyelination result in the formation of shortened internodes having more variable lengths. This can be shown graphically when IL of teased fibers is plotted against their average diameter.61 This approach, although useful for graphic display, does not allow statistical analysis. To assess IL and ID more fully, one should consider the issue of sampling, of how to estimate the average IL or ID per teased fiber or per nerve, and of how to estimate the variability of IL or ID per teased fiber or per nerve. First, for the assessment to be meaningful, it is necessary to obtain a representative sample of teased fibers from nerve. This can be done by a systematic sampling of endoneurial strands from nerve as described above (see Sampling). Nonrepresentative teasing of MFs may affect both average IL and its variability. Because internodes of large-diameter fibers are longer than those of small-diameter fibers, the mean of mean IL of teased fibers may be greater than it should be if one selectively picks large fibers, or smaller than it should be if one selectively chooses small fibers. Values indicating a short mean of mean IL may be due to selection of too many small fibers during teasing, selective loss of large fibers, or previous remyelination after segmental demyelination. Conversely, longer than normal mean of mean IL values may be due to selection of too many large fibers during teasing or loss of small fibers.

Table 32–1. Graded Pathologic Conditions of Teased Fibers (%) from Sural Nerve of Healthy Volunteers A* Normal Appearance Mean Median SD Range Mean Median SD Range Mean Median SD Range

B Myelin Irregularity

C, D, F, G Segmental De- and Remyelination

15–⬍30 yr (n ⫽ 21; 9F and 12M; mean, 22 yr; median, 22 yr; range, 20–27 yr) 95.0 0.2 3.9 94.9 0 3.1 3.1 0.5 2.7 90.2–100.0 0–2.3 0–8.6 30–⬍45 yr (n ⫽ 4; 3F and 1M; mean, 35.5 yr; median, 34.0 yr; range, 32–47 yr) 95.0 0.5 4.0 96.5 0 2.9 3.6 1.0 4.4 89.8–98.1 0–2.0 0–10.2 45–⬍60 yr (n ⫽ 4; 1F and 3M; mean, 49.3 yr; median, 49.0 yr; range, 45–54 yr) 81.5 1.8 14.5 83.0 0.6 15.9 11.0 2.8 9.3 67.3–92.6 0–5.9 2.2–23.8

F ⫽ female; M ⫽ male. *Descriptions of conditions A to H are given in the text.

E, H Axonal Degeneration and Regeneration 1.0 0 1.8 0–7.6 0 0 0 0–0 1.9 1.1 2.3 0–5.3

753

Pathologic Alterations of Nerves

The mean of mean IL may be normalized or adjusted to reflect a normal size distribution, allowing direct comparison to healthy nerve. The approach minimizes the effect of selection of small myelinated fibers (SMFs) or large myelinated fibers (LMFs) fibers by selective teasing or by disease. The approach can be illustrated for sural nerve, in which there is a bimodal size distribution and 60% of MFs generally are less than 5.5 ␮m in diameter (SMFs) and 40% are greater than or equal to 5.5 ␮m in diameter (LMFs). Let us assume, in our first example, that too many large teased fibers were selected (60% LMFs and 40% SMFs) and that the mean of mean IL of LMFs is 700 ␮m, whereas that of SMFs is 300 ␮m, to give a mean of mean IL (unadjusted) of 540 ␮m. Assuming a normal 60:40 ratio of SMFs to LMFs, the adjusted mean of mean IL is 0.6 (300) ⫹ 0.4 (700) ⫽ 460 ␮m. For our second example, let us assume that, in a nerve from a patient with Friedreich’s ataxia, 90% of MFs are small and 10% large. Let us also assume that the mean of mean IL of small fibers is 300 ␮m, whereas the 10% of large fibers have a mean of mean IL of 700. The mean of mean IL (unadjusted) is 340 ␮m, but the adjusted mean of mean IL is also 460 ␮m (normal)—evidence that remaining fibers have approximately normal ILs, considering the caliber of fiber. The variability of IL may be tested by determining the coefficient of variation (CV) of IL per individual teased fiber. The mean of mean CV of IL of 100 teased fibers (CV of IL/nerve) is then determined. Because in health the individual CV is somewhat larger for small than for large fibers, one must consider whether an abnormally high CV of IL is due to a selective loss of large fibers, selective teasing of small fibers, or previous segmental demyelination and remyelination. An abnormally low CV of IL may be due to selective loss of small fibers or a selective teasing of large fibers. Systematic sampling should preclude selective teasing by fiber size, but how can one compare the mean of mean CV of IL when there is an altered size distribution of fibers in disease? A normalized CV of IL, using the CV of IL of various size categories from disease but calculating values assuming

a normal distribution, will allow direct comparison of values from disease and control. In Table 32–2 we give values for CV of IDs and ILs of MFs obtained from sural nerves of two controls and a patient with an inherited neuropathy. In the latter disease, the mean of mean IL is much shorter and the variability of IL and ID is greater than in controls. Assessing for Clustering of Demyelination The study of the distribution of demyelination among “old internodes” of teased fibers can provide an important insight as to the mechanism of demyelination. If demyelination occurs at random among old internodes, a diffusely mediated damage to Schwann cells is suggested. A metabolic derangement of Schwann cells or an intoxication of Schwann cells mediated through the endoneurial fluid bathing them might cause such diffuse damage. This appears to be the case in lead neuropathy.53 Alternatively, if demyelination occurs repeatedly on a selected few teased fibers but not on many teased fibers, secondary demyelination should be suspected. Such secondary demyelination is found in uremic neuropathy38 and Friedreich’s ataxia,43 and is attributed to axonal atrophy. The approach to identification of “original” or old internodes is described and illustrated in Figures 32–22 and 32–23.

Electron Microscopic Assessment of Teased Fibers This is an elegant but time consuming and expensive evaluation seldom practical for the assessment of clinical biopsy material. The technique permits ultrastructural assessment at defined positions along the length of teased fibers.42,146 Uncommon focal abnormalities can then be quickly located and appropriately sectioned. Another use is to systematically examine structural features along the length of fibers.42 This approach has been used to advantage to show the variability in axonal caliber and of myelin thickness along abnormal fibers central to neuroma146 and in fibers from sural nerve in

Table 32–2. Coefficients of Variation of Diameters and Lengths of Internodes of 100 Teased Fibers from Healthy Sural Nerves and from Sural Nerve of Patient with HMSN type I Coefficients of Variation (%) Diameters Nerve

Age (yr), Sex

68–69 69–69 65–69

10, M 19, M 22, M

Lengths

Category of Nerve Donor

Mean

Range

SD

Mean

Range

SD

Accident Congenital heart disease HN-CMT

10.5 9.0 19.3

3.0–23.5 1.6–23.2 7.0–52.2

4.72 4.24 8.97

9.7 9.4 35.2

2.8–32.8 1.7–32.2 12.2–68.1

4.68 5.80 11.18

HMSN type I ⫽ hereditary motor and sensory neuropathy type I; HN-CMT ⫽ hypertrophic neuropathy of Charcot-Marie-Tooth type; SD ⫽ standard deviation. From Dyck, P. J., Ellefson, R. D., Lais, A. C., et al.: Histologic and lipid studies of sural nerves in inherited hypertrophic neuropathy: preliminary report of a lipid abnormality in nerve and liver in Dejerine-Sottas disease. Mayo Clin. Proc. 45:286, 1970, with permission.

754

Pathology of the Peripheral Nervous System

MORPHOMETRY OF NEURON SOMAS

FIGURE 32–22 Schematic drawings of teased myelinated fibers with segmental demyelination (upper) and remyelination (lower). The method of scoring is by the “territory of an old internode” for the determination of randomness of segmental demyelination (see text and original article for derivation). (From Dyck, P. J., O’Brien, P. C., and Ohnishi, A.: Lead neuropathy. 2. Random distribution of segmental demyelination among “old internodes” of myelinated fibers. J. Neuropathol. Exp. Neurol. 36:570, 1977, with permission.)

hereditary motor and sensory neuropathy (HMSN) type I42 (Fig. 32–24) and Friedreich’s ataxia.53

Electron Microscopic Correlation with Teased Fiber Abnormality There is only limited information on this subject. This is given in previous editions of this textbook.

Estimates of the number and size of neuron somas in nuclei, nuclear columns, or ganglia as well as of fibers in tracts or nerves have been reported for more than 100 years. Early results were usually so variable, and are so different from results now obtained using improved approaches, that they must be considered unreliable. However, some early studies of the number and size distribution of nerve fibers were well done; for example, Erlanger and Gasser’s results relating components of the compound action potential to peaks of the fiber diameter histogram are noteworthy.58 Now, better fixatives and recognition that hyperosmolar fixatives affect estimates of number, size, and shape, as well as improved embedding media, thinner sections, and improved microscopic resolution and morphometric approaches, are available that permit performing reliable morphometric analysis of fibers of nerves as summarized in this chapter. The objective of morphometric analysis is to determine the number, size, shape, and spatial distribution of various neuron cell bodies and their axons at predetermined anatomic levels. Such quantitative data are used to characterize alterations with development, with aging, with various environmental and toxic exposures, with disease, and with therapy. If the number and size distribution of cell bodies and of proximal and distal axons of peripheral neurons are estimated for a given disease and compared to controls, the altered three-dimensional structural changes associated with disease can be inferred. Reference morphometric results are now available for defined levels of certain motor neuron columns, intermediolateral columns (ILCs), spinal ganglia, sympathetic ganglia, and various levels of different nerves.

Motor Neuron Columns

FIGURE 32–23 The same schematic drawings of teased fibers as shown in Figure 32–22 but as scored by method 2 (see original article for derivation). Because evaluation for randomness using this method will result in a clustered distribution, even when the involvement of Schwann cells is random, this method cannot be used. (From Dyck, P. J., O’Brien, P. C., and Ohnishi, A.: Lead neuropathy. 2. Random distribution of segmental demyelination among “old internodes” of myelinated fibers. J. Neuropathol. Exp. Neurol. 36:570, 1977, with permission.)

Kawamura and colleagues87 systematically evaluated the number and size distribution of motor neuron column somas of the third, fourth, and fifth lumbar segments of humans. These somas are clustered into medial and lateral groups (Fig. 32–25). The medial group probably innervates truncal muscles while the lateral group innervates limb muscles. The typical morphology of one of the columns of a lateral group of motor neurons is shown in Figure 32–26. Plotting the frequency of diameters of lumbar spinal cord motor neuron somas, three peaks were found: one at a small cyton diameter (CS), one at an intermediate diameter (CI), and one at a large diameter (CL) (Fig. 32–27). The peaks and troughs were approximately the same from one spinal cord to another spinal cord. Using troughs as division points, it is possible to estimate the number of somas (or cytons15) in CL, CI, and CS peaks for L3, L4, and L5 spinal cord segments. The values for 18 control spinal cords of human patients without known disease of peripheral nerve are given in Table 32–3.

Pathologic Alterations of Nerves

755

FIGURE 32–24 Transverse sections at the same magnification (⫻79,050) from two sites on the same teased fiber from a child with hypertrophic neuropathy of the Charcot-Marie-Tooth type. In insets, both at the same low magnification (⫻5088), rectangles show area reproduced at higher magnification. These pictures illustrate that good electron microscopic preservation is possible by the method of teasing described in the text. (From Dyck, P. J., and Lais, A. C.: Electron microscopy of teased nerve fibers: method permitting examination of repeating structures of same fiber. Brain Res. 23:418, 1980, with permission.)

The usefulness of this approach is illustrated by a study of motor neuron columns in amyotrophic lateral sclerosis86 and in inherited dysautonomia.40 Figure 32–28, taken from the paper on dysautonomia, shows the selective absence of CIs in familial dysautonomia.

Spinal Ganglia The number and size distribution of somas of S1 spinal ganglia embedded in paraffin were reported for a group of control persons.119 In a later study, Kawamura and Dyck

examined L5 ganglia using celloidin embedding, greater enlargement to increase the number of counting windows, and averaging to improve peak recognition.85 Figure 32–29 shows six histograms of patients 17 to 77 years of age. Numbers of cytons in each peak are given in Table 32–4. Three overlapping peaks are present. In contrast to motor neurons, in these ganglia CS is generally the highest, followed by CI and then by CL. Using the troughs between CS and CI and between CI and CL, the numbers of small, intermediate, and large somas were calculated, and these are given in Table 32–3.

756

Pathology of the Peripheral Nervous System

FIGURE 32–25 Schematic presentation of transverse sections of L3, L4, and L5 spinal cord of human showing motor neuron columns. (From Dyck, P. J., Low, P. A., and Stevens, J. C.: Diseases of peripheral nerves. In Baker, A. B., and Baker, L. H. [eds.]: Clinical Neurology. Philadelphia, Harper Medical, p. 1, 1981, with permission.)

To illustrate the abnormality that may occur in disease, Figure 32–30 shows the severe depletion of L5 spinal ganglion somas in dysautonomia.

Intermediolateral Column The thoracic sympathetic preganglionic outflow is important in the maintenance of postural normotension in humans. Low and co-workers determined the number and size distribution of ILC somas of the T6, T7, and T8 segments of 12 spinal cords of humans.104 The mean counts for T6, T7, and T8 segments were 5002, 5004, and 4654, respectively. No sex difference was shown. Most somas ranged from 8 to 23 ␮m. The major peak was at 12 to 13 ␮m and a smaller peak was seen at 16 ␮m. There was a progressive reduction of ILC somas with age, amounting to 8% per decade. This reduction in ILC cytons with age may be a morphologic basis for postural hypotension of the aged.

Sympathetic Trunk Ganglion Dyck and colleagues have compared the number and size of T7 sympathetic ganglia neurons for a control individual and for a patient with dysautonomia (Fig. 32–31).40

MORPHOMETRY OF PERIPHERAL NERVES The approaches described here can be used to determine the number, density, diameter distribution, shape, and spatial distribution of fibers of nerves or of spinal cord tracts. In addition, the approaches used permit assessment of the

FIGURE 32–26 Representative microscopic field of one of the columns of the lateral group of motor neurons of the L5 segment of the human spinal cord showing the favorable histologic preparation for morphometric analysis, which can be obtained using aldehyde fixation, celloidin embedding, and cresyl violet staining, as described in the text. The spinal cord was fixed in buffered paraformaldehyde and glutaraldehyde solution, embedded into celloidin, sectioned at 25 ␮m, and stained with cresyl violet. (From Kawamura, Y., O’Brien, P. C., Okazaki, H., and Dyck, P. J.: Lumbar motoneurons of man. II. The number and diameter distribution of large- and intermediate-diameter cytons in “motoneuron columns” of spinal cord of man. J. Neuropathol. Exp. Neurol. 36:861, 1977, with permission.)

diameter distribution of MFs, axon areas, myelin areas, axon perimeters, and myelin perimeters. Each measured attribute can be separately assessed for each fiber or per nerve and related to one another (e.g., axon area regressed on myelin thickness). Should it become possible to reliably identify functional classes of axons in transverse sections using markers, it will be possible to determine their size class in health and disease. The density, size distribution of MFs, axon-to-myelin thickness relationships, and index of circularity (IC) appear to change during development, regeneration, aging, and degeneration. In addition, these approaches may be used to characterize neuropathologic abnormalities.

757

Pathologic Alterations of Nerves

a fiber as obtained from serial section will be larger and smaller than the average diameter. This variance in healthy nerve was close enough to allow Erlanger and Gasser to show a relationship between conduction velocities and diameter.58 Lambert and Dyck found that in demyelinating neuropathies, however, histograms from a sampled section are not a true reflection of the average diameter of fibers. Such histograms show fiber profiles that are much too large and much too small—the histogram from a single section does not accurately reflect the average diameter distribution of fibers (see Chapter 39).

Histologic Processing

FIGURE 32–27 Diameter frequency distributions of cytons of motor neuron columns of L3 (top), L4 (middle), and L5 (bottom) segments of a representative spinal cord from a 17-year-old male. Three distinct diameter peaks are seen: large diameter (CL), intermediate diameter (CI), and small diameter (CS). It is reasonable to assume that the majority of CL cytons are alpha motor neurons, that the majority of CI cytons are gamma motor neurons, and that the CS cytons are interneurons. (From Kawamura, Y., O’Brien, P. C., Okazaki, H., and Dyck, P. J.: Lumbar motoneurons of man. II. The number and diameter distribution of large- and intermediate-diameter cytons in “motoneuron columns” of spinal cord of man. J. Neuropathol. Exp. Neurol. 36:861, 1977, with permission.)

Because fibers of healthy nerves are cylinders of approximately similar caliber for long distances, it is possible to infer that the range and diameter peaks of one transverse section are representative for other levels of nerve. Extensive work relating diameter peaks to the various peaks of the compound action potential in vitro were based on this premise. Estimates of numbers of MFs will be slightly too low because sections through nodes of Ranvier will not be recognized as MFs. With disease, variability of axonal caliber, segmental demyelination and remyelination, and axonal degeneration may develop, so that the number and diameter distribution from one section may not be an accurate reflection of the number or average size of fibers. To illustrate, the largest and smallest diameter of

Immersion fixation of nerves in hyperosmolar glutaraldehyde results in reproducible fascicular and fiber shape alterations from normal (see Fig. 32–6). Whereas nerve fascicles and MFs assume roughly spherical shapes in rapidly frozen and freeze-dried transverse sections, they assume flattened, boomerang, or crenated shapes, to a variable degree, with hyperosmolar fixation. The IC (⫽ ␲D/P, where P is perimeter and D is diameter, and which is 1 for a circle) is significantly lower for nerves fixed in hyperosmolar fixative than it is for those fixed in isosmolar fixative. The perineurium and the axolemma act as semipermeable membranes. These distortions can be avoided, at least to a large degree, when nerves are fixed in isosmolar aldehyde fixative (e.g., 2.5% glutaraldehyde in 0.025 mol/L cacodylate buffer). With perfusion or in situ fixation, hyperosmolar fixatives do not have such a large adverse effect. Low temperatures (0° to 4° C) are associated with severe depolymerization of MTs, and it is for this reason that we fix tissues at 10° C. We employ a specially designed cart having an insulated (during transport from the hospital to the laboratory) metal block housing the fixative bottles and a thermoelectric unit (with controller and power supply) that maintains temperature at 10° C when connected to line current.

Section Thickness and Axon-Myelin Relationship Thick, semithin, and thin sections can be used to provide accurate estimates of numbers of MFs, but only the last two may be used to evaluate size, shape, and myelin-axon relationship. Use of thick sections or cutting the fibers slightly obliquely will result in too large a profile area, spuriously thick myelin, and thin axons. Although thin (~80-nm) sections, as used for electron microscopy, permit the most accurate measurement of fiber diameter, shape, and myelin-axon relationship, semithin sections are commonly used because systematic sampling is better on these larger sections, because the process is less time consuming, and because quite accurate results can be obtained. Although semithin sections 1.5 ␮m thick

758

Pathology of the Peripheral Nervous System

Table 32–3. Number of Large (CL), Intermediate (CI), and Small (CS) Cytons in 18 Spinal Cords of Humans L3

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Mean SD Range

L4

L5

CL

CI

CS

CL

CI

CS

CL

CI

CS

5720 6720 5731 4795 4797 5925 4481 4736 5133 5637 5035 5726 4595 5526 4910 4193 4017 5652

2595 3138 2538 1820 2220 1797 1334 2112 1250 3348 2024 1764 1195 1482 1429 1348 1600 2883

1316 924 932 1733 1075 1533 635 584 1497 479 1451 593 1247 592 710 725 1310 913

4753 5647 4138 4920 5311 4541 5275 5096 4219 4496 4453 4768 4361 3797 3470 4753 4672 3720

1489 2017 1984 2601 2536 1131 1953 2152 1450 1720 1230 1150 1401 1424 1213 1912 2131 1355

636 909 940 1154 1228 959 370 576 446 223 533 690 968 520 923 729 694 937

5108 4575 5353 5979 5454 5799 5691 4062 5101 4721 5727 4556 4643 4836 4739 4629 4372 3887

1480 1562 2315 1818 1562 1706 1188 1115 1445 1143 2077 1890 1365 2371 1485 1662 1770 1153

1306 1282 1422 857 490 638 318 502 500 1636 544 989 1360 1014 1452 932 1619 1027

5185 691.0 4017–6720

1993 666.8 1195–3348

1014 392.5 479–1733

4577 575.2 3470–5647

1714 464.0 1131–2601

746 273.8 223–1228

4957 606.4 3387–5979

1617 377.8 1115–2315

994 424.3 318–1619

From Kawamura, Y., O’Brien, P. C., Okazaki, H., and Dyck, P. J.: Lumbar motoneurons of man. II. The number and diameter distribution of large- and intermediate-diameter cytons in “motoneuron colons” of spinal cord of man. J. Neuropathol. Exp. Neurol. 36:861, 1977, with permission.

are usually employed, with adequate osmication and use of paraphenylenediamine, it is possible to reduce thickness to 0.5 to 1.0 ␮m, thus improving error caused by section thickness.

Attribute of Shape of Sampled Sections of Internodes That Best Estimates Average Diameter Karnes and co-workers84 have addressed this issue. Using computer imaging, measurements of MF profiles were made at intervals over half-internodes of 20 fibers from each of four sural nerves of healthy rats. In these serial skip transverse sections (every 10th section) of the 80 MFs followed from nodes of Ranvier to Schwann cell nucleus or vice versa, the following measurements were made: 1. Long axis (DL): longest distance across fiber 2. Short axis (DS): greatest distance across the fiber at right angles to the long axis 3. Area (A): calculated by a modified Simpson’s rule for irregular contours 4. Perimeter (P): summation of distances between adjacent border points

5. Diameter, assuming calculated area to be that of a circle: Dc ⫽ 2√(A/␲) 6. Diameter computed as mean of long and short axes (arithmetic mean): DM ⫽ (DL ⫹ DS)/2 7. Diameter computed as square root of product of long and short axes of ellipse (geometric mean): DE ⫽ √(DL ⫻ DS) 8. Diameter, assuming calculated perimeter to be that of a circle An idealized cylinder was reconstructed for each fiber over the half-internode measured using the cross-sectional areas (A), the thickness of sections (T), and the number of sections between pictures (S). The diameter of this idealized cylinder (DV) could then be calculated from its volume (V), which was computed using the following equation: N

V ⫽ T ⫻ ⌺ 1/2 (An ⫹ An⫺1)(Sn ⫺ Sn⫺1) n⫽2

The diameters, computed in six different ways as described, were compared for bias, precision, and accuracy between sections and were compared with the diameter of an idealized cylinder reconstructed for each fiber from multiple actual cross sections. In this study, DV for a single internode of a given fiber was evaluated from multiple cross sections.

759

Pathologic Alterations of Nerves

To evaluate how well an estimation method performed for a given internode, we computed the mean, standard deviation (SD), and square root of the mean square error of the estimates. These statistics measure bias, precision (consistency), and accuracy, respectively. On the basis of all three criteria, the diameter of a circle of equivalent area (DC) is clearly the method of choice. It was not surprising that this method shows least bias because DV and DC are derived from the area of the cross sections as determined by different methods, but this fact would not necessarily favor DC for consistency or accuracy between sections. DM and DE, which are similar to methods described in the literature, compared reasonably well to DC. Karnes and colleagues also found that a crenated shape was characteristic of the paranodal region, a boomerang shape of the nuclear region, and a circular shape of the area between paranodal and nuclear regions (Fig. 32–32).84

Myelinated Fiber Composition: Photographic Method FIGURE 32–28 Diameter histograms of cell bodies of L5 motor neuron columns in representative control (top) and dysautonomic patient (bottom). In the control, three peaks are found, large (CL), intermediate (CI), and small (CS). In the patient, only two peaks are found, large and small. Because CI and intermediate axons (AI) of ventral root are thought to be from gamma motor neurons, this is evidence of a preferential absence of gamma motor neurons in dysautonomia. (From Dyck, P. J., Kawamura, Y., Low, P. A., et al.: The number and sizes of reconstructed peripheral autonomic, sensory, and motor neurons in a case of dysautonomia. J. Neuropathol. Exp. Neurol. 37:741, 1978, with permission.)

FIGURE 32–29 Each of the diameter histograms of the cell bodies of neurons in L5 spinal ganglion shows three peaks at approximately the same diameter location, an indication that the methodology produces reproducible results and that there are at least three populations by size in this ganglion. (From Kawamura, Y., and Dyck, P. J.: Evidence for three populations by size in L5 spinal ganglion in man. J. Neuropathol. Exp. Neurol. 37:269, 1978, with permission.)

Earlier editions of this textbook outlined photographic morphometric methods that might be used. For some investigators the use of microprocessor digitizers, now widely available, will replace the mechanical counters of an earlier era.

Myelinated Fiber Composition: Computer Imaging Method The advantages of computer imaging for determining the number, size, shape, and distribution of myelin fibers and their components are as follows: it is speedy, automatic

760

Pathology of the Peripheral Nervous System

Table 32–4. Number of CL, CI, and CS Cytons in L5 Spinal Ganglion of Six Men 1 2 3 4 5 6 MV SD %

Age/Sex

CS

CI

CL

Total

17/M 22/M 40/M 56/M 68/M 77/M

29,654 22,642 26,039 28,160 27,431 18,410

35,537 30,229 29,366 30,950 29,954 33,943

13,897 12,155 17,713 10,145 7345 7316

78,088 65,076 73,119 69,255 64,730 59,669

25,389 4164 37

31,675 2481 46

11,429 4033 17

68,323 6597

From Kawamura, Y., and Dyck, P. J.: Evidence for three populations by size in L5 spinal ganglion in man. J. Neuropathol. Exp. Neurol. 37:269, 1978, with permission.

bordering occurs in approximately 85% of MFs, machine criteria not subject to human judgment are used in bordering, exact perimeters and areas of fibers are provided, the need for photographic processing is eliminated, and maximum flexibility in altering programs and handling

FIGURE 32–31 Diameter histogram of T7 sympathetic trunk neurons from control (top) and from patient with familial dysautonomia (bottom). (From Dyck, P. J., Kawamura, Y., Low, P. A., et al.: The number and sizes of reconstructed peripheral autonomic, sensory, and motor neurons in a case of dysautonomia. J. Neuropathol. Exp. Neurol. 37:741, 1978, with permission.)

data is possible. The system is operator interactive so that each enumerated and bordered fiber visualized on the video screen can be verified under the microscope. Programs have been developed that identify MFs and border the inner edge of myelin (Fig. 32–33). This is averaged (eight times) to increase accuracy. For each identified MF the following measurements are made: length of perimeter (from bordering the inner edge of myelin), area enclosed by perimeter, and myelin thickness. From these measurements a variety of derived attributes are calculated. Because the data for each fiber are stored individually, they can be flexibly assessed later by multiple approaches. Subroutines have been prepared for drawing histograms and relating various variables by regression analysis. FIGURE 32–30 Diameter histogram of L5 spinal ganglion neuron cell bodies in control (top) and dysautonomic patient (bottom). Note that the scale of the ordinate for the patient is approximately one tenth that of the control. Neurons of all sizes are severely reduced in number, but small ones are preferentially affected. (From Dyck, P. J., Kawamura, Y., Low, P. A., et al.: The number and sizes of reconstructed peripheral autonomic, sensory, and motor neurons in a case of dysautonomia. J. Neuropathol. Exp. Neurol. 37:741, 1978, with permission.)

Hardware Components The hardware components of the imaging system for nerve morphometry (ISNM) that we have developed were described in earlier editions of this textbook. System Approaches The approaches and feasibility studies that led to the histologic preparation used and to the design and software features employed in ISNM were given in detail in earlier editions of this textbook.

Pathologic Alterations of Nerves

761

FIGURE 32–32 A drawing of a hemi-internode of a myelinated fiber showing that characteristic shapes are seen near the internode (I1), adjacent to the nucleus (I3), and between I1 and I3. (From Dyck, P. J., and Karnes, J. L.: Morphometry of neuron columns and fiber tracts in neurobiology and pathology using computer imaging. Trends Neurosci. 4:138, 1981, with permission.)

FIGURE 32–33 Using an operator-interactive computer imaging system for nerve morphometry, the inner boundaries of myelin are bordered as described in the text. Axon area and diameter, myelin area and thickness, and number and size distribution of myelinated fibers are automatically determined as described in the text. (From Zimmerman, I. R., Karnes, J. L., O’Brien, P. C., and Dyck, P. J.: Imaging systems for nerve and fiber tract morphometry: components, approaches, performance and results. J. Neuropathol. Exp. Neurol. 39:409, 1980, with permission.)

As in the photographic method, the transverse area of fascicles is determined on low-power enlargements using digitization. Then at high-dry magnification (~2000⫻), the endoneurial area is conceptually overlaid by frames corresponding to a square frame in the eyepiece. Each area of the endoneurium, frame by frame, and parts of frames, is brought under the ocular frame by a x and y traverse of the endoneurial area. The first frame for analysis is then chosen from a random number table, and thereafter frames are assessed at regular intervals (e.g., 1 in 4 or 1 in 6). The ratio of assessed to all frames chosen is predetermined so that 400 to 1000 MFs per nerve will be assessed. This number assures that the histogram will be smooth. For each frame to be assessed, hemocytometric technique is employed. Because the analysis is interactive and because every fiber included is bordered and otherwise marked (by a white square in the axon) on the video screen, every fiber within a frame can be detected and bordered with certainty, without spurious inclusion of other histologic profiles, since the operator can accept or reject every profile by histologic criteria on the video screen or, if he or she wishes, by direct examination of the same frame of the histologic section in the microscope. After recognition and bordering of the inner edge of myelin, computer programs estimate the perimeter and areas of the bordered profiles by rules described in detail in earlier editions of this textbook. MF size, shape, and IC are obtained by the following steps. Measured and Derived MF Measurements from Transverse Sections. With the ISNM, the areas and perimeter of all fascicles are measured using a low-power objective. Using a high-power objective (63⫻), the endoneurial area, the number of MFs per frame, and the axon area, axon perimeter, and myelin thickness for each MF

762

Pathology of the Peripheral Nervous System

even in health. Therefore, each level must be compared to its own control. This approach may be used to recognize dying-back alterations or a focal pathologic abnormality. Density, average size, or range of diameters may be compared among frames within fascicles, and among fascicles to recognize multifocal pathology. P. C. O’Brien (personal communication, 1980) has suggested two approaches to assess variability. In the first approach, the variability in the total series of frames consists of two components: variability among frames within fascicles and variability among fascicles. That is, ␴ 2 ⫽ ␴A2 ⫹ ␴␻2 . In these terms, a suitable index is ␴A / √(␴A2 ⫹ ␴␻2 ). The second approach is the same as that used in evaluating clustering based on a series of frames within fascicles. The basic measurement becomes Xi, or the number of fibers observed in the ith fascicle, where i ⫽ 1, . . . , n. However, since the area (Ai) sampled will vary, it will be necessary to take this into account. (The same would have been true in our previous index, if we had taken partial frames into account.) For notation, if we define

within the rectangular frame are measured. Derived attributes that are calculated include the area of myelin, the area of combined axon and myelin, and the diameter of a circle of equivalent area (diameter of MF). An IC (⫽ ␲D/P) is calculated as a measure of noncircularity. From the measured and derived attributes for each MF, which are stored, various analyses are possible. First, the density of MFs (number/mm2 or per nerve) is determined. Second, the mean, median, range, and SD of the nerves, axons, myelin, and MFs as areas or as diameters (of a circle of equivalent area) are listed. Similar calculations can be made for IC. Third, any of these measured or derived attributes can be displayed as a frequency distribution, preferably by square millimeter of fascicular area or by nerve. For MFs we employ an exponential plot with bars of equal relative width to use the same precision in locating peaks for fibers of small or large diameter. If well-developed peaks and troughs are visualized, the number of MFs per peak can be determined. Fourth, one variable can be regressed on another. Derived values in a typical nerve recently analyzed are as shown in Table 32–5. Assessing Variability in MF Density. The density and size distribution of MFs may be used to recognize focal or multifocal abnormality along the length of nerves and between or within fascicles at one level. To do so, the location of the sampled frames and the measured attributes of MFs by frame must be known. In evaluating for alterations along the length of nerves, it is important to recognize that a direct comparison between sampled levels generally is not possible because they may be different

Di ⫽ Xi /Ai n

D ⫽ i⫽1 ⌺ Di /n S2D ⫽ ⌺(Di ⫺ D)2/(n ⫺ 1) AH ⫽ n(⌺Ai⫺1) ⫺1 (harmonic mean range)

Table 32–5. Peroneal Nerve of Cat: Number and Size Distribution of Myelinated Fibers Nerve 140–79L LT1 Whole nerve Number of frames Number of measured fibers Number of unmeasured fibers Fascicular area (mm2) % of fascicle sampled Fascicular IC

Mean SD Median Number/mm2 SD Confidence interval Upper limit: 8.72114E Lower limit: 7.80565E

50 891 61 1.05 11.00 0.96 Diam (␮m)

IC

MT (␮m)

MA (␮m2)

Per (␮m)

8.50 3.94 8.73

0.88 0.03 0.89

1.23 0.44 1.34

32.83 24.18 31.04

29.88 13.24 30.41

8262 273 8721 7806

IC ⫽ index of circularity as defined in text; MA ⫽ myelin area; MT ⫽ myelin thickness; Per ⫽ axon perimeter.

Pathologic Alterations of Nerves

then the proposed index is ID ⫽

S2D AH D

Notice that, if A1 ⫽ . . . ⫽ An, this reduces to the wellknown index of dispersion for testing a common Poisson distribution. The derivation is as follows: E ⌺(Di ⫺ D)2 ⫽ ⌺ E(D2i ) ⫺ n E(D)2 ⫽ ␭ ⌺ Ai⫺1 ⫹ n␭2 ⫺ n ⫺1 E(⌺ Di)2 E ⌺(Di)2 ⫽ ⌺ E(D2i ) ⫹ ⌺ DiDj ij

⫽ ␭ ⌺ Ai⫺1 ⫹ n2␭2 Therefore, E ⌺(Di ⫺ D)2 ⫽ ␭ ⌺ Ai⫺1 ⫹ n␭2 ⫺ ⫽

1 ␭ ⌺ Ai⫺1 ⫺ n␭2 n

n⫺1 ␭ ⌺ Ai⫺1 n

⫽ ␭(n ⫺ 1)/AH Because E(D) ⫽ ␭ E[S2D AH] E(D)

⫽1

if the fibers are distributed according to the same random distribution on all n fascicles.

NORMAL NUMBERS OF FIBERS IN CONTROL HUMAN NERVES This section provides an abbreviated account of MF density (number/mm2 of fascicular area), number/nerve, diameter distribution (plotted either per nerve or per square millimeter of fascicular area), and ranges and peaks of human nerves taken at autopsy or at biopsy (sural nerve). Patients or subjects who provided nerves were without neurologic disease or disease predisposing to peripheral neuropathy. The method of fixation and embedding was as described for postmortem and biopsy tissue in the preceding sections. Epoxy transverse sections were cut at 0.75 ␮m and stained with paraphenylenediamine and evaluated with ISNM as described above. The diameter histograms plotted per square millimeter of endoneurial area for various

763

human nerves are shown in the illustrations in this section. In each figure, short single vertical lines pointing down from the baseline indicate the position of the first percentile and the 99th percentile positions. Double lines indicate the median diameter. The position of frequency distribution peaks is sometimes also shown as long vertical lines pointing down from the baseline. To the left of the histogram we show the age and sex and the number of MFs per nerve and per square millimeter; to the right we show the MF percent of fascicular area. For subsequent histograms of the same nerve in each figure, only the baselines of the histogram are shown to be able to include many histograms in one figure. The bottom line provides the values for the composite histograms. By assessing these normative morphometric data, some inferences can be formulated. The first is that, for a given level of a specific nerve, there is considerable variability both in density and in number of MFs/nerve, but the range, diameter histogram, median diameter, and diameter position of peaks is remarkably constant. The second is that, for an unbranching nerve, the number of MFs/nerve appears to remain constant, whereas the number/mm2 decreases from proximal to distal. The diameter histogram shows no evidence for fiber attenuation over long proximal-to-distal distances—the peaks being essentially at the same diameter position in proximal as in distal nerve. This evidence suggests that there are mechanisms for maintaining axon caliber over long distances. The third inference is that there are differences in fiber composition among roots and nerves reflecting differences in composition by fiber class. The fourth is that there are striking differences in diameter histograms among species. The fifth is that morphometric values are sufficiently consistent that, when they are evaluated in patients with neuropathy, important insights may emerge regarding the nerves affected, the class of neurons (axons) affected, the level within the neuron affected, and the components (axon or myelin) affected. A comparison of the MF diameter histograms of T7 and L5 ventral spinal roots (VSRs) (Figs. 32–34 and 32–35) illustrates that the size distribution pattern may be characteristic for segmental levels and different among levels. This variability is described here in more detail. Designating small axons as AS, intermediate axons as AI, and large axons as AL, the T7 VSR has a high peak for AS, a low peak for AI, and an intermediate-height peak for AL. These peaks correspond to preganglionic splanchnic fibers, gamma fibers, and alpha fibers, respectively. The height and diameter position of the peaks in L5 VSRs are strikingly different, there being only a few AS axons and a ratio of AI to AL axons similar to that found in T7. There is another difference between T7 and L5. The AI and AL peaks of T7 are at smaller diameter positions than they are for L5. These differences reflect the greater number of preganglionic axons in T7 and the smaller caliber of gamma and alpha axons for thoracic than for lumbar roots, perhaps reflecting the shorter length of thoracic neurons.

764

Pathology of the Peripheral Nervous System

FIGURE 32–34 Histograms and morphometric data on T7 human ventral spinal roots (VSR).

FIGURE 32–35 Histograms and morphometric data on L5 human ventral spinal roots (VSR). MF ⫽ myelinated fiber. (From Dyck, P. J., Karnes, J., Sparks, M., and Low, P. A.: The morphometric composition of myelinated fibers by nerve, level, and species related to nerve microenvironment and ischaemia. Electroencephalogr. Clin. Neurophysiol. Suppl. 36:39, 1982, with permission.)

Pathologic Alterations of Nerves

There is also a difference in the diameter position of MF diameter peaks between species. The VSR axons of the cat are considerably larger than those of humans (Fig. 32–36), the AL peak of the former being at approximately 14 to 15 ␮m and the latter at approximately 13 to 14 ␮m. Other evidence for diameter distribution differences among species may be found in the work of Boyd and Davey16 and in our previous communications.39 The striking difference in fiber composition between ventral and dorsal spinal roots of humans is illustrated in Figures 32–35 and 32–37. Diameter histograms of human L5 segmental nerve, midthigh sciatic nerve, tibial nerve, peroneal nerve, and proximal and distal sural nerve are presented in Figures 32–38, 32–39, 32–40, 32–41, 32–42, and 32–43, respectively. The second inference (the lack of tapering of nerve fibers) is illustrated by a comparison of histograms of the sural nerve at midcalf and at the ankle (see Figs. 32–42 and 32–43). The positions of diameter peaks (AS and AL) are not significantly different for the two, but the density is somewhat less at ankle level—the fascicular area tends to be large distally. The fifth inference is illustrated by providing two examples, the first related to the selective vulnerability of alpha motor neurons in motor neuron disease (Fig. 32–44) and the second the selective vulnerability (considering

FIGURE 32–36 Histograms and morphometric data on L7 ventral spinal roots (VSR) of cat. MF ⫽ myelinated fiber.

765

motor neurons) of gamma motor neurons in inherited dysautonomia (Fig. 32–45).

ARTIFACTS OF SURGICAL REMOVAL AND OF HISTOLOGIC PREPARATION Before nerve tissue can be visualized under the microscope, it must be surgically removed, fixed, dehydrated, embedded, cured, sectioned, and stained. Despite this severe manipulation, structures seen under the microscope bear a strong similarity to what is seen in life, in freshly teased tissue, or in cryostat sections. The fact that compound action potentials in vitro can be reconstructed from diameter histograms of fibers provides further evidence that the structure of fibers has been approximately retained during histologic processing. Likewise, subcellular structures such as nodes of Ranvier, Schmidt-Lanterman incisures, and organelles observed in living preparations are retained in embedded tissue. However, artifactual alterations caused by removal or processing of nerve also occur and must be recognized so that they are not interpreted as evidence of pathologic alteration. At nerve biopsy the major artifacts that may be induced are hemorrhage, excessive stretch, compression, osmotic

766

Pathology of the Peripheral Nervous System

FIGURE 32–37 Histograms and morphometric data on L5 human dorsal spinal roots (DSR). MF ⫽ myelinated fiber.

FIGURE 32–38 Histograms and morphometric data on L5 segmental human nerve. MF ⫽ myelinated fiber. (From Dyck, P. J., Karnes, J., Sparks, M., and Low, P. A.: The morphometric composition of myelinated fibers by nerve, level, and species related to nerve microenvironment and ischaemia. Electroencephalogr. Clin. Neurophysiol. Suppl. 36:39, 1982, with permission.)

Pathologic Alterations of Nerves

767

FIGURE 32–39 Histograms and morphometric data on human sciatic nerve. MF ⫽ myelinated fiber.

shrinkage, and introduction of foreign material. Some degree of hemorrhage is probably inevitable because sharp dissection is used. Typically it is found at the edges of the biopsy specimen. Fresh hemorrhage may also occur in disease and is recognized as such by its localization to

vessel alterations or to inflammatory reaction. Excessive traction, compression, and manipulation induce major tissue alteration; avoidance of these artifacts is always better than trying to distinguish them from pathologic alterations. Local anesthetic should be injected several

FIGURE 32–40 Histograms and morphometric data on human tibial nerve. MF ⫽ myelinated fiber.

768

Pathology of the Peripheral Nervous System

FIGURE 32–41 Histograms and morphometric data on human peroneal nerve. MF ⫽ myelinated fiber. (From Dyck, P. J., Karnes, J., Sparks, M., and Low, P. A.: The morphometric composition of myelinated fibers by nerve, level, and species related to nerve microenvironment and ischaemia. Electroencephalogr. Clin. Neurophysiol. Suppl. 36:39, 1982, with permission.)

FIGURE 32–42 Histograms and morphometric data on human proximal sural nerve. MF ⫽ myelinated fiber. (From Dyck, P. J., Karnes, J., Sparks, M., and Low, P. A.: The morphometric composition of myelinated fibers by nerve, level, and species related to nerve microenvironment and ischaemia. Electroencephalogr. Clin. Neurophysiol. Suppl. 36:39, 1982, with permission.)

Pathologic Alterations of Nerves

769

FIGURE 32–43 Histograms and morphometric data on human distal sural nerve. MF ⫽ myelinated fiber. (From Dyck, P. J., Karnes, J., Sparks, M., and Low, P. A.: The morphometric composition of myelinated fibers by nerve, level, and species related to nerve microenvironment and ischaemia. Electroencephalogr. Clin. Neurophysiol. Suppl. 36:39, 1982, with permission.)

centimeters above the upper end of the nerve to be transected so the injected fluid is not misinterpreted as edema. The suture should be passed through the nerve and the free ends used as loop retractors to elevate the nerve from its bed. The operator must avoid tying and thus crushing the nerve. The degree of tension applied on the suture used as a retractor should be just sufficient to straighten the nerve or fascicles while being undercut. Excessive stretch induces tears, bleeding, separation at nodes, and a beaded appearance of teased fibers. Immediately after removal of the nerve specimen, it should be put into the fixative solution. The issue of when artifactual crush of nerve fibers seen in histologic sections occurs is still not finally settled. Perhaps it can occur at several times (i.e., at surgery and during histologic fixation). A severe example of crush is shown in Figure 32–46. Crush can be distinguished from axonal degeneration by the following criteria: (1) large fibers are more affected than small fibers; (2) edge fibers are more affected than central fascicular

fibers; (3) in the former, myelin is splayed apart and sometimes disrupted, whereas in the latter and especially at later stages of degeneration, myelin is in discrete ovoids or solid balls; and (4) in the latter, there is an increase in Schwann cell cytoplasm and there may be autophagic vacuoles (some of which may contain sudanophilic lipids). This artifact is perhaps more readily recognized in teased fibers by the observation that the area of crush occurs at the same proximal-to-distal level in multiple examined fibers (Fig. 32–47). Crush artifact may also be misinterpreted as segmental demyelination in teased fibers. This possibility should be considered when (1) the demyelination does not affect the paranode and (2) myelin at the edge of the apparent demyelination appears thickened, distorted, frayed, and irregular; there are not associated smooth myelin ovoids and balls in the adjacent cytoplasm; and similar alterations occur at approximately the same region of other teased fibers (see Fig. 32–47). Other artifacts related to teasing were shown in Figure 32–12. In serial sections,

770

Pathology of the Peripheral Nervous System

FIGURE 32–44 Representative histograms of L5 motor neuron cell bodies (left) and ventral root myelinated axons (right) from control subject (top), Case 1 with amyotrophic lateral sclerosis (ALS) (middle), and Case 3 with ALS (bottom). The histogram of Case 1 is similar to that of most cases of ALS in that the CL and AL peaks are missing. The histogram of Case 3 is different in that the CL and AL peaks are higher and broader than normal. (From Kawamura, Y., Dyck, P. J., Shimono, M., et al.: Morphometric comparison of the vulnerability of peripheral motor and sensory neurons in amyotrophic lateral sclerosis. J. Neuropathol. Exp. Neurol. 40:667, 1981, with permission.)

the artifact may be shown to be restricted to a given proximal-to-distal level of nerve. Major artifacts come from freezing and from autolysis. In Figure 32–48, ice crystal formation has severely distorted the histologic structures. Autolysis, which may be due to prolonged periods before immersion into fixative or use of too weak a fixative solution, results in watery axons, distorted and ballooned mitochondria, dispersed chromatin, and other artifacts.

HISTOLOGIC STRUCTURES WITH AND WITHOUT PATHOLOGIC SIGNIFICANCE Renaut Corpuscles Histologic structures referred to as Renaut corpuscles were first described in the nerves of human, horse, dog, and ass by Renaut (Fig. 32–49).128,129 Subsequently they have been seen in nerves of persons with various disorders. They have been extensively studied and reported in the German and French literature.35,97 In the English

literature they were sometimes misinterpreted as infarcts.88 The corpuscles usually are found on the inside of the perineurium and protrude into the endoneurial space. Fibers course around them without being distorted by them. In longitudinal profile they taper at both ends and merge with the perineurium. In transverse sections they consist of concentric lamellae of filamentous strands. Oval, plump, pale nuclei—the cellules gondronnées of Renaut—are associated with filamentous strands. At the center of the corpuscle there usually are three or four plump nuclei similar to those noted in mucoid degeneration. The lamellae stain light pink with hematoxylin and eosin, pale green with trichrome, not at all with methyl violet, and strongly blue with Alcian blue at pH 2.5. The staining reaction suggests that they contain acid mucopolysaccharides. They are seen in the nerves of healthy persons, in greater frequency with increasing age, and, according to Asbury and co-workers,4 especially adjacent to limb joints. The consensus is that Renaut corpuscles are not specific for a particular disease and certainly are not infarcts of nerve.120 Occasionally, pacinian corpuscles are located in the epineurium (Fig. 32–50).

Pathologic Alterations of Nerves

FIGURE 32–45 Diameter histogram of myelinated axons of L5 ventral root of control (top) and of dysautonomic patient (bottom). In the control, two well-defined peaks are seen, large (AL) and intermediate (AI). In dysautonomia AI is well developed but AL is markedly reduced in size from normal values. (From Dyck, P. J., Kawamura, Y., Low, P. A., et al.: The number and sizes of reconstructed peripheral autonomic, sensory, and motor neurons in a case of dysautonomia. J. Neuropathol. Exp. Neurol. 37:741, 1978, with permission.)

Schwann Cell Inclusions Schwann cell inclusions occur primarily at polar regions of the nuclei of Schwann cells and in paranodal cytoplasm of Schwann cells. Found at these and other locations are mitochondria, glycogen granules, lipid droplets, lysosomal organelles, ␲ (pi) granules, ␮ (mu) granules of Reich (also called Elzholz bodies), tuffstone bodies, and less well characterized inclusions, such as needle-like inclusions and multilamellar bodies (e.g., zebra bodies) (Fig. 32–51). In paraffin or thick epoxy sections, the ␲ granules stain with thionine, toluidine blue, or methylene blue and appear as metachromatic granules. Usually they are rodlike or curvilinear in shape and approximately 1 ␮m long (Fig. 32–52). They are directed radially or circumferentially to the direction of the fiber. In electron micrographs, they undoubtedly correspond to Schwann cell cytoplasmic lamellar bodies. It may be that zebra bodies in lysosomal storage disease are closely related but similar organelles. They have a periodicity of approximately 4.8 nm, may be surrounded by a double membrane, and may occur singly or in clusters. Because they are more frequent and larger in

771

FIGURE 32–46 A nerve fascicle that has been severely crushed during surgical removal or before adequate fixation. The distinctive features are that edge fibers on one side of the fascicle and large fibers are selectively affected, there is no consequent Schwann cytoplasmic increase or reactivity, and the changes are restricted to the crushed region (the last not shown). (From Dyck, P. J., Dyck, P. J. B., Giannini, C., et al.: Peripheral nerves. In Graham, D. I., and Lantos, P. L. [eds.]: Greenfield’s Neuropathology, 7th ed., Vol. 2. London, Arnold Publishing, p. 551, 2002, with permission.)

lysosomal storage diseases, they are assumed to represent degradative products generated by lysosomal enzymes in healthy nerves. They have been described by Thomas and Slatford152 in nerve of the rabbit, by Evans and co-workers59 in nerves of patients with various neuromuscular disorders, by Gonatas and colleagues65 in nerves of a patient with a disorder resembling Refsum’s disease, and by Dyck and Lambert47 in a series of nerves from healthy persons and nerves from patients with hypothyroid neuropathy. Another Schwann cell degradative inclusion is the ␮ granule of Elzholz (Fig. 32–53). Because it stains with the Marchi method, it is assumed to be a degradative product of myelin. Rarely a lipocyte is found in the endoneurial area (Fig. 32–54).

772

Pathology of the Peripheral Nervous System

FIGURE 32–47 Teased fibers laid side by side from a nerve grasped during surgery. Notice that the point of crush is recognized in all sampled teased fibers. Although this is a severe example, finding fibers without myelin showing at approximately the same proximal-to-distal level of different teased fibers provided the clue that the nerve has been crushed. (Modified from Dyck, P. J., Dyck, P. J. B., Giannini, C., et al.: Peripheral nerves. In Graham, D. I., and Lantos, P. L. [eds.]: Greenfield’s Neuropathology, 7th ed., Vol. 2. London, Arnold Publishing, p. 551, 2002.)

FIBER ALTERATIONS Neuronal Degeneration Historically, anterior horn cell degenerations as occur in progressive spinal muscular atrophy or amyotrophic lateral sclerosis were separated from peripheral neuropathies by clinical, pathologic, and later electrophysiologic features. Simply stated, in the first group of disorders there was selective involvement of motor neurons and the changes were not length dependent. In peripheral neuropathy, other classes of neurons are involved and the process is

often length dependent. However, the distinction between motor neuron degeneration and peripheral neuropathy is not always apparent when it is appreciated that, in motor neuron degeneration, axonal spheroids may be located in peripheral nerves (e.g., ventral root segments), and that, in inherited motor neuropathies, signs may indicate some degree of length dependence. In some neuronal degeneration, the brunt of the disease process appears to be the perikaryon, in others the nucleus, and in still others the axon just distal to the soma. If severe enough, the entire neuron degenerates. Chapter 35 outlines the electrophysiologic features that separate anterior horn from peripheral nerve disorders. Chapter 31 outlines pathologic alterations and mechanisms of degeneration of somas. In acute poliomyelitis, inflammation is localized to the ventral gray columns of the spinal cord, especially damaging the perikaryon of lower motor neuron somas (Fig. 32–55). Figures 32–25 and 32–26 show lower motor neuron cell columns in health. In Fabry’s disease, an X-linked metabolic disorder caused by deficiency of a lysosomal enzyme (see Chapter 81), lipids accumulate especially in somas of primary afferent neurons but also in perineurial and endothelial cells (Fig. 32–56). Neuron somas may also be adversely affected by the action of sensitized mononuclear cells, as may be the case in inflammatory polyganglionopathy (Fig. 32–57).

Wallerian Degeneration FIGURE 32–48 A cross section of a nerve fixed in glutaraldehyde and osmium tetroxide, embedded in epoxy, and stained with methylene blue. This section shows major artifactual separation of endoneurial contents as a result of ice crystal formation. The nerve in fixative was inadvertently frozen during transport to our laboratory. See Color Plate

When axons are transected, distal axons undergo stereotyped sequential alterations resulting in disappearance of the axon and myelin, known as wallerian degeneration.107 In the rat, after nerve transection, axons continue to conduct for 36 to 48 hours, with only a minor reduction of conduction velocity of the fastest fibers. The muscle action

Pathologic Alterations of Nerves

773

FIGURE 32–49 Epoxy transverse sections of sural nerve stained with methylene blue showing prominent Renaut corpuscles—the spherical pale areas in the endoneurium. In the American medical literature, these were sometimes incorrectly identified as infarcts. See Color Plate

FIGURE 32–50 A pacinian corpuscle in epineurium of sural nerve paraffin section stained with hematoxylin and eosin. This is an infrequent finding. See Color Plate

potential also remains relatively normal for 36 hours and then drops to 0 over the next 6 to 8 hours. In transverse electron micrographs at 36 to 48 hours, some axons have a normal appearance, others show clumping of their cytoskeleton, and still others are devoid of axonal contents and show only degenerating organelles. By the third or fourth day, most recognizable axonal contents will have disappeared (Fig. 32–58). Schlaepfer and Hasler133 have suggested that axonal contents are extruded into the interstitial fluid, but immunohistologic localization of neurofilament (NF) and tubular protein identification remain to be done to confirm this observation. Using immunologic markers, Bignami and colleagues13 demonstrated that axonal debris reacting with NF antisera disappeared much earlier from sectioned sciatic nerve (10 days) than from sectioned optic nerve (up to 4 months). The sequential changes that fibers undergo in wallerian degeneration are studied to advantage in teased fibers. Beginning at 36 to 48 hours, MFs become segmented into linearly arranged, closely spaced ovoids and myelin balls. It appears that separations occur at nodes of Ranvier and at Schmidt-Lanterman incisures. Over the next 2 to 4 weeks, the ovoids become progressively smaller and are increasingly separated from each other by longer regions of the nerve tube without ovoids. Before their final disappearance, ovoids lose their osmiophilia and are widely separated from each other. In young adult rats (~300 to 500 g), segmentation of myelin is evident in essentially all fibers at 3 to 6 days,

774

Pathology of the Peripheral Nervous System

FIGURE 32–51 A, Within the cytoplasm of the Schwann cell of a myelinated fiber are clusters of ␲ granules (arrow), also called Schwann cell cytoplasmic lamellar bodies or zebra bodies (⫻8480). B, Higher magnification of ␲ granule cluster. C, Electron micrograph of mast cell from nerve (⫻11,280). The metachromatic staining properties of this cell may be confused with the metachromasia of metachromatic leukodystrophy disease granules or with deposits of amyloid. D, Portion of longitudinal section through a Schmidt-Lanterman incisure, showing deposit of glycogen (arrow) (⫻39,480). E, Perineurial cell with glycogen (arrow) (⫻20,800).

discrete separation into ovoids and balls is typical in 1 to 3 weeks, following which the ovoids and balls are smaller and separated by long distances, and thereafter only small and widely spaced osmiophilic droplets are seen. It is important to recognize that the linear orientation of ovoids and late breakdown products in nerve tubes are maintained for

weeks or months. This linear array, consisting of the basement membrane of the former nerve fiber, Schwann cells, and macrophages, appears to be important for regeneration to originally innervated target tissue. Under the electron microscope, the ovoids initially have the characteristic periodicity of myelin, but later

Pathologic Alterations of Nerves

775

FIGURE 32–52 The rod-shaped ␲ granules of Reich (discussed in text) are shown in cytoplasm adjacent to compact myelin in both A and C. The boxed areas in A and C are shown at high magnification in B and D, respectively. The lamellar periodicity of granules is approximately 4 to 5 nm. (From Dyck, P. J., and Lambert, E. H.: Polyneuropathy associated with hypothyroidism. J. Neuropathol. Exp. Neurol. 29:631, 1970, with permission.)

these degenerate into amorphous lipid, which loses its osmiophilia and becomes sudanophilic. The events of myelin degradation begin in autophagic vacuoles of Schwann cells. Later, myelin digestion appears to take place within macrophages in the nerve tubes referred to above. Macrophages of hematogenous origin play important roles in dealing with debris. Beginning at approximately 6 days, there is a great influx of these

cells into the endoneurium, and these enter the nerve tubes and aid in fiber degeneration. Simultaneously, there is a great increase also in the number of Schwann cell nuclei. The total increase in nuclei is variable, depending on the fiber composition of the nerve, but it may be as much as six fold or greater. The volume of the endoneurial area may increase considerably during this time.

FIGURE 32–53 A ␮ granule of Reich (Elzholz body). A myelin ovoid in the Schwann cell cytoplasm is indicated by a single arrow. A similar compact lamellar structure is shown by the double arrows (as discussed in text). AC ⫽ axis cylinder. (From Dyck, P. J., Johnson, W. J., Lambert, E. H., and O’Brien, P. C.: Segmental demyelination secondary to axonal degeneration in uremic neuropathy. Mayo Clin. Proc. 46:400, 1971, with permission.)

776

Pathology of the Peripheral Nervous System

FIGURE 32–54 A lipocyte in the endoneurium of sural nerve (transverse section, epoxy embedding, methylene blue stain) is a most unusual finding. See Color Plate

The sequence of wallerian degeneration of unmyelinated fibers has not been studied as intensively as for MFs. As for MFs, so also for the unmyelinated fibers: the earliest evidence of fiber alteration is the dissolution and clumping of the NFs and MTs of axons (Fig. 32–59). The axons of transected distal nerve unmyelinated fiber stumps at 2 to 3 days may be recognized by their position in a trough external to the Schwann cell cytoplasmic membrane and by their having an intact axolemma but axoplasm that appears watery, with cytoskeletal elements decreased or clumped. At this stage some unmyelinated fibers in a group of those associated with one Schwann cell may show degenerative changes, and others may still be normal. At later stages the trough of former sites of unmyelinated fibers may be recognized. Troughs may be filled with lipid deposits (e.g., in Tangier disease) or by collagen (collagen packets). At later times, remnants of myelin may persist for 30 to 60 days in bands of Büngner (Fig. 32–60). Clusters or stacks of Schwann cell cytoplasmic processes surrounded by a common basement membrane may indicate the previous site of a MF or of a cluster of unmyelinated fibers. It was previously thought that wallerian degeneration is highly stereotyped and inevitable both by the nature of the histologic reactions and by temporal events. Studies of certain mutant mouse strains indicate that this is not always the case, the process being much altered in certain mutations. Perhaps this observation suggests that extrinsic factors, not yet fully understood, play key roles in the development and progression of wallerian degeneration. Wallerian degeneration occurs in many human diseases. The causes are physical transection, compression, tear,

FIGURE 32–55 1, Round-cell infiltration of the ganglion cells of the lateral group of the gray matter, in the sacral region of the cord, from a polio case two and a half days after the onset. 2, Ganglion cells from the same region in the same case almost destroyed by phagocytic neuroglia cells (neurophages). 3, Infiltration of the lymphatic space of a central vessel from the lumbar cord of the same case. 4, Infiltration of a vascular space and of the surrounding tissue in the anterior horn of the lumbar cord in the same case. (Reprinted with permission from Wickman, I.: Acute poliomyelitis. J. Nerv. Ment. Dis. 16:1, 1913, with permission.) See Color Plate

radiation, burn, or other physical injuries. External nerve compression may be caused by tumor enlargement beneath tight fascial sheaths or in foramina. Another cause is ischemic transection. Wallerian degeneration may be caused by external compression of a fiber by focal interstitial pathologic alterations, as from a nodule of amyloid. We have postulated that it may be caused by a loop of myelin indenting and transecting the axon (internal strangulation; see below).37 Wallerian degeneration may be initiated at sites of active segmental demyelination.6 According to this view, the axons could be damaged as a nonspecific consequence of a specific cell-mediated immune reaction occurring in the vicinity—that is, a consequence of the so-called bystander effect.150,163 Toxic factors secreted by activated lymphocytes or other metabolic events have been postulated.109,132

Pathologic Alterations of Nerves

777

FIGURE 32–56 Photomicrographs from patient with Fabry’s disease, a lysosomal disorder resulting in storage of glycosphingolipid. A, One cell body appears to be filled by the stored material. B–D, Electron micrographs showing the typical lamellar structure of the stored material. (From Ohnishi, A., and Dyck, P. J.: Preferential loss of small sensory neurons in Fabry’s disease: histological and morphometric evaluation of cutaneous nerves, spinal ganglia, and posterior columns. Arch. Neurol. 31:120, 1974, with permission.)

Demyelinative Internal Strangulation The idea that secondary demyelination might cause axonal degeneration came from studies of the permanent axotomy model (see Permanent Axotomy Studies below). In this model axonal atrophy leads to secondary demyelination and axonal degeneration. Our studies suggest that, as the Schwann cells retract from the node, axial myelin buckling occurs. The infolded myelin loops may, at some sites, be large enough to compress and transect the axon. We have repeatedly demonstrated such myelin infoldings in various demyelinating conditions: after application of a pneumatic cuff,32 in the permanent axotomy model, and in various metabolic and inherited diseases. That the axon can be compromised by these myelin infoldings is suggested

by the finding that the infolded loop may contain fullthickness myelin, a size sufficient to severely compromise axonal contents. Axonal Degeneration in Severe Demyelination In acquired demyelinating neuropathies it is common to also find axonal degeneration. Could this axonal degeneration be the result of severe demyelination? Teased fibers that proximally show segmental demyelination and distally show axonal degeneration (teased fiber condition I) suggest that the two processes may be linked. Others have explained the occurrence of both processes by a bystander effect. Inflammation with the release of cytokines may damage Schwann cells (myelin) and also axons.

778

Pathology of the Peripheral Nervous System

FIGURE 32–57 Paraffin section with hematoxylin and eosin staining of spinal ganglion from a patient with inflammatory polyganglionopathy as described in the text. (Reprinted with permission from Grant, I. A.: Treatment of nonmalignant sensory ganglionopathies. In Noseworthy, J. H. [ed.]: Neurological Therapeutics: Principles and Practice. Vol. 2, London, Martin Dunitz, p. 2070, 2003, with permission.) See Color Plate

Axonal Dystrophy Definition Generally, the term axonal dystrophy is used synonymously with axonal swellings or axonal spheroids, which are multifocal accumulations of cytoskeletal elements. Here we extend the definition to include focal accumulations of other organelles or other formed structures. History Seitelberger found argyrophilic axonal swellings in the brains of children who died from a disorder that he named neuroaxonal dystrophy.136 Subsequently, axonal swellings were also described in various human disorders (giant axonal neuropathy, amyotrophic lateral sclerosis, hexacarbon intoxications, and various types of HMSN) and in experimental neuropathies (those caused by acrylamide, ␤,␤⬘-iminodipropionitrile [IDPN], n-hexane, methyl butyl ketone, and carbon disulfide). With the introduction of the electron microscope, these enlargements were shown to be focal accumulations of cytoskeletal components, particularly of NFs. Impairment of transport of NF components was implicated in these accumulations. Description of Typical Axonal Spheroids Axonal spheroids may be found in proximal nerves (e.g., in amyotrophic lateral sclerosis, neuroaxonal dystrophy, and IDPN and dimethyl hexanedione neuropathies) and in distal nerves (2,5-hexanedione–induced neuropathy, giant axonal neuropathy, and a variety of HMSN).

Giant axonal neuropathy is a disorder of children, probably recessively inherited, with large axonal spheroids in both peripherally and centrally directed primary afferent neuron axons (Fig. 32–61).4,28,29 Abnormal intermediate filament inclusions are also found in Schwann cells, endothelial cells, and fibroblasts. In toxic neuropathies the distribution of axonal spheroids may be related to the dosage and potency of the toxins. Typically, spheroids enlarge the axon, forming ellipsoidal accumulations of NFs. Neurofilaments are absolutely increased in number, closely packed, and often misdirected. In transverse sections of spheroids, several patterns are recognized: accumulation of NFs only; segregated regions of NFs only and of MTs only; and segregated regions of NFs, MTs, and accumulated organelles and vesicular profiles (dense lamellar, vesicular, and paracrystalline structures). Although the axon is generally enlarged, sometimes greatly, at other times only modest enlargement is found. The overlying myelin, especially near the equator of the spheroid, may be absent or thinned. At the poles of spheroids or in small spheroids, myelin may be of normal thickness. Distal to an axonal spheroid, the axon caliber is smaller than normal (atrophied) or may undergo degeneration. Usually the term axonal spheroid is used for the pathologic alteration of fibers in the conditions listed above, for example, giant axonal neuropathy and neuropathies caused by many industrial toxins. Accumulation of axonal organelles different from those described above may occur in compression and ischemic injury. Korthals and Wisneiwski96 and Nukada and colleagues114,115 have described in detail the three-dimensional alterations for ischemic lesions. If serial sections of MFs are followed from a nonischemic region to an ischemic region, as studied by Nukada and colleagues,115 the following changes from proximal to distal are encountered: (1) no abnormality of the axon; (2) slight enlargement and increased density of the axon (with or without light cores) and increased darkness on electron microscopy as a result of accumulated mitochondria, dense bodies, and vesicular profiles; (3) severe enlargement of the axon with accumulated organelles; (4) attenuated dark axons (also with accumulated organelles); and (5) disappearance of the axons or return to a normally appearing axon (see Fig. 32–58 below). Light cores are relatively preserved axoplasm surrounded by accumulated organelles. At the lateral margin of ischemic cores, the changes are less severe. Changes 1 to 4 may be seen, but loss of axons is less severe. Myelin thickness and size may be unaffected with changes 1 or 2 above. With change 3 above, myelin is usually thinned or absent. With changes 4 and 5 above, the myelin profile may be spherical or collapsed but the myelin may be of normal thickness or thin. The changes described above were observed within days or a week or two of the ischemic event. When fibers in ischemic cases undergo degeneration, events of axonal degeneration and repair occur and

Pathologic Alterations of Nerves

779

FIGURE 32–58 Transverse electron micrographs of peroneal nerve of mouse at 36 (A), 48 (B), 72 (C), and 96 (D) hours after sciatic nerve transection (⫻8000). At 36 hours, the appearance of myelinated fibers (MFs) is not greatly disturbed. Note the relative preservation of unmyelinated fibers (UFs). There is dissolution of neurofilaments and microtubules in axons. At 48 hours, granular clumping of axonal contents of MFs and degenerative changes in UFs are evident. By 72 hours, nerve microtubules of MFs show myelin ovoids without axons and many UFs have disappeared. By 96 hours, these changes are more advanced.

regenerating axons replace the degenerating ones. At a later date only small regenerating fibers occupy the ischemic cores. A different focal enlargement of axons is represented by polyglucosan bodies (PGBs) (Fig. 32–62). In this condition the transverse area of axoplasm probably remains normal

but total axonal area is increased because of the presence of PGBs. The PGBs are focal accumulations of abnormal glycogen polymers and are found in nerve fibers in both health and disease. Assessing the number of myelin lamellae and myelin spiral length, Yoshikawa and co-workers165 found that the number of myelin lamellae overlying PGBs

780

Pathology of the Peripheral Nervous System

FIGURE 32–59 Wallerian degeneration in distal segment of cervical sympathetic trunk 34 hours after crush. Many degenerating fibers have a watery appearance; others have material that is clumped and densely opaque. (From Dyck, J., and Hopkins, A. P.: Electron microscopic observations on degeneration and regeneration of unmyelinated fibres. Brain 95:223, 1972, with permission.)

bodies was normal but the myelin spiral length was greatly increased as compared to internode segments beyond the bodies. Because movement of myelin components along membranes is unlikely to account for this focal increase of the myelin sheet, local synthesis of myelin components is postulated. Accumulation of a variety of organelles (glycogen, myelin figures, endoplasmic reticulum, degenerated mitochondria, and lysosomes) may be encountered in both experimental and human neuropathies. Although these may not cause focal enlargements (swellings), they may represent dystrophic changes.

Axonal Atrophy, Myelin Remodeling (Secondary Demyelination), and Degeneration This subject was reviewed as a separate chapter in the second edition of this textbook.51 Although early pathologists talked about axonal atrophy as a mechanism by which fibers disappeared, the histologic events had not been described. The cellular events involved were revealed by a study of sural nerves of patients with uremic neuropathy and Friedreich’s ataxia and by a study of the permanent axotomy model. The

Pathologic Alterations of Nerves

781

A

FIGURE 32–60 Electron micrographs of bands of Büngner. A, Six days after crush. B, Fifteen days after crush (AC ⫽ axis cylinder). One, or possibly two, axon sprouts are seen in B. (From Dyck, P. J.: Ultrastructural alterations in myelinated fibers. In Desmedt J. E. [ed.]: New Developments in Electromyography and Clinical Neurophysiology. Basel, S. Karger, p. 192, 1973, with permission.)

B

sequential steps involved are postulated to be as shown in Figure 32–63. Studies of Uremic Neuropathy and Friedreich’s Ataxia The mechanism underlying secondary segmental demyelination was suggested from results of study of the sural nerve from the midcalf and ankle levels from patients with uremic neuropathy.38 The findings were as follows: 1. The MF density was much lower at the ankle than at the midcalf level, indicating greater fiber loss distally. 2. The peaks of fiber diameter histograms were at smaller diameter categories at the ankle than at the midcalf level, suggesting disappearance of the largest axons from distal nerves or distal fiber atrophy. 3. Teased fibers from the midcalf level showed myelin wrinkling and paranodal demyelination predominantly,

whereas teased fibers from the ankle level showed axonal degeneration predominantly, suggesting that the lesser pathologic involvement caused segmental demyelination and the greater involvement caused axonal degeneration. 4. Teased fibers frequently showed multiple regions of demyelination or remyelination of the same fiber, indicating that myelin remodeling was clustered on certain fibers. 5. The slopes of regression lines relating axonal area to myelin thickness were lower for distal than for proximal nerve and both were lower than in normal nerve, an indication that axons were attenuated relative to the amount of myelin. 6. Serial electron micrographs of fibers with multiple segments of demyelination and/or remyelination had small axons relative to the thickness of myelin of old internodes.

FIGURE 32–61 Selected electron micrographs from a young Iranian boy with giant axonal neuropathy. Upper left, Transverse section through a typical spheroid. The axon is without myelin and is surrounded by concentric layers of Schwann cell processes. Upper right, A longitudinal view of a spheroid in a myelinated fiber. Lower left, An axon showing a central zone of neurofilaments surrounded by a circumferential zone containing microtubules and accumulated organelles. Lower right, Electron micrograph of transverse section through an axonal spheroid, showing close spacing of neurofilaments and absence of side arms—considered by the authors to be evidence of an abnormality of neurofilament cross-linking in this disorder. (From Donaghy, M., King, R. H. M., Thomas, P. K., and Workman, J. M.: Abnormalities of the axonal cytoskeleton in giant axonal neuropathy. J. Neurocytol. 17:197, 1988, with permission.)

Pathologic Alterations of Nerves

783

FIGURE 32–62 Transverse section through a myelinated fiber of a patient with chronic inflammatory demyelinating polyneuropathy who also had frequent polyglucosan bodies (PGBs) in axons, as shown here in the lower left. The PGB had caused focal axonal enlargements that had not changed the number of overlying lamellae but was associated with increased myelin spiral length (MSL). The increased MSL was thought to be due to local synthesis of myelin components rather than movement of lipid or protein components along the membrane. BM ⫽ basement membrane; IM ⫽ inner membrane; OM ⫽ outer membrane. (From Yoshikawa, H., Dyck, P. J., Poduslo, J. F., and Giannini, C.: Polyglucosan body axonal enlargement increases myelin spiral length but not lamellar number. J. Neurol. Sci. 98:107, 1990, with permission.)

The idea of a secondary form of segmental demyelination was also suggested by a study of nerves from patients with early Friedreich’s ataxia. In biopsy samples of sural nerve from such patients, Dyck and Lais43 found that 1. Diameter histograms of MFs of patients with early involvement showed striking abnormalities, with the peak for small fibers being higher and to the left of the peak of normal nerves. The height of the largediameter fiber peak was smaller than in controls, suggesting that some large fibers had become smaller and some small fibers had become even smaller. 2. Most teased fibers had a normal appearance, but others had multiple regions of demyelination and/or remyelination, indicating that the events of demyelination and remyelination were not randomly distributed among Schwann cell territories but were clustered on certain abnormal (atrophying) fibers (Fig. 32–64). 3. Electron micrographs from fibers with demyelination and remyelination showed axonal atrophy.

4. Axonal degeneration occurred occasionally but appeared to be a final event in axonal atrophy. From these studies from human neuropathies, a number of conclusions can be drawn: 1. The study of fiber degeneration in uremic neuropathy and Friedreich’s ataxia supports the idea that there is a type of segmental demyelination and remyelination that is associated with axonal atrophy. Dyck and colleagues called this type of demyelination and remyelination “secondary.”38,43 This secondary segmental demyelination and remyelination should not be confused with nonspecific loss of myelin staining resulting from loss or decreased numbers of fibers. 2. Secondary demyelination is characterized by the following features: a. Initially, there is wrinkling of myelin of old internodes. b. Then, paranodal demyelination occurs.

784

Pathology of the Peripheral Nervous System

FIGURE 32–63 The proposed cellular events of chronic neuronal (axonal) atrophy, myelin remodeling, and degeneration from a study of uremic neuropathy and Friedreich’s ataxia. The permanent axotomy model has provided evidence for steps 1, 2, 3, 4, and 5. Several cycles of demyelination and remyelination (indicated by steps 5 and 7) may occur with development of hypertrophic neuropathy. Depending on the balance of degenerative and regenerative events, fibers might recover from various stages of fiber alteration, dwindle away, or undergo acute degeneration. These cellular events are common in a variety of metabolic and inherited diseases affecting peripheral and, conceivably, central nervous system neurons. MF ⫽ myelinated fiber. (Modified from Dyck, P. J., Lais, A. C., Karnes, J. L., et al.: Permanent axotomy, a model of axonal atrophy and secondary segmental demyelination and remyelination. Ann. Neurol. 9:575, 1981.)

c. With a and b, it may be possible to demonstrate reduced caliber of the axon relative to number of myelin lamellae. d. Secondary demyelination may be recognized in some diseases and in teased fibers by certain fibers having multiple demyelinated and remyelinated old internodes, whereas adjacent fibers are without these changes. In other words, the demyelination is clustered on some fibers and not on others. 3. Axonal atrophy, myelin remodeling with or without hypertrophic neuropathy, and axonal degeneration or loss constitute a common pattern of fiber abnormality in human inherited, metabolic, and other neuropathies. 4. Myelin alterations occur, including radial and axial infolding and myelin cleavage, discussed in detail later (see Permanent Axotomy Studies below). 5. Finally, there is adaxonal sequestration and removal of cellular breakdown products (Fig. 32–65). The sequence of postulated cellular events underlying axonal atrophy is shown in diagrammatic form in Figure 32–63. This sequence might be thought of as a common pathway for injury of various types. The scheme suggests that chronic neuronal (axonal) injury may lead ultimately to axonal degeneration, atrophic disappearance, or recovery. In an adult, a previously atrophic fiber with multiple segments of demyelination might be recognized after

recovery by having multiple short (remyelinated) segments intermixed with longer normal internodes. After recovery in a child whose limbs were still growing, the previously remyelinated internodes would elongate and could not be recognized as remyelinated because they are no longer short. Retrograde Effects of Nerve Transection Gutmann and Sanders71 evaluated the number and sizes of MFs in the proximal peroneal nerve of rabbit at various times up to 300 days after crush, resuturing, and nerve grafting of the distal nerve. They found no decrease in the number of MFs but a shift in caliber toward smaller sizes by 120 days, which was no longer present after 200 days. To explain the reduced NCV of nerves proximal to such distal peripheral nerve lesions as carpal tunnel syndrome, Cragg and Thomas24 compared the NCV and the size of the five largest MFs in each of four sections from the peroneal nerves of a rabbit proximal to the following lesions: crush, ligature, transection and reconnection, and transection and distal nerve avulsion. In the first three experiments, NCV fell to 90% of normal by 25 to 30 days, and to 80% by 50 to 100 days, and then rose to normal by 200 days. In sharp contrast, in transection and distal avulsion NCV fell to between 60% and 70% of normal by 200 days and did not

Pathologic Alterations of Nerves

785

FIGURE 32–64 Consecutive lengths along one teased myelinated fiber from the sural nerve of a patient with Friedreich’s ataxia. The atrophy of this fiber is evident from the multiple regions of demyelination and remyelination of the same fiber. Many contiguous large-diameter fibers had no demyelination, an indication that the demyelination is not occurring in a random fashion between “old” territories of Schwann cells but is clustered on some axons. The demyelination is therefore secondary to an axonal influence as discussed in this chapter (see text) and in Chapter 37. (From Dyck, P. J., Lambert, E. H., and Nichols, P. C.: Quantitative measurement of sensation related to compound action potential and number and sizes of myelinated and unmyelinated fibers of sural nerves in health, Friedreich’s ataxia, hereditary sensory neuropathy, and tabes dorsalis. In Cobb W. A. [ed.]: Handbook of Electroencephalography and Clinical Neurophysiology, Vol. 9. Amsterdam, Elsevier, p. 83, 1971, with permission.)

improve thereafter. Later, Aitken and Thomas2 concluded that atrophy affected both myelin and axons but that the axons were slightly more atrophied. Assuming that no segmental demyelination occurred, these findings suggest that fiber atrophy is transitory when the nerve is allowed to regrow to its target tissue, but it is more severe when reconnection is inhibited. After recording from the spinal roots of cats whose leg nerves had been cut and prevented from regrowing, Hoffer and co-workers79 reported a decline of NCV, which stabilized for ventral root fibers but was progressive for posterior root fibers.

Lubinska106 observed that, if only part of an internode was damaged by crush, the remainder (containing the Schwann cell nucleus) might survive in a truncated condition. This type of damage may be related to direct injury to the Schwann cell. Lubinska also noted that “in a certain proportion of fibers . . . the reaction to the lesion spreads back in the proximal direction and may involve several internodes.” The proximal end of the last “preserved internode” became demyelinated, and the space between the axon and neurilemma was filled with hypertrophied Schwann cell cytoplasm and many nuclei. The frequency of such demyelination was highest in the second to the last

786

Pathology of the Peripheral Nervous System

FIGURE 32–65 A, Small, myelinated fiber showing irregular configuration of axolemma with densely packed microtubules and neurofilaments in its axoplasm. (Original magnification, ⫻18,000.) B, Large myelinated fiber with abnormal round and oblong membrane-bound profiles containing dense bodies, mitochondria, and filaments between axolemma and compact myelin. (Original magnification, ⫻8500.) (From Ohnishi, A., Peterson, C. M., and Dyck, P. J.: Axonal degeneration in sodium cyanate-induced neuropathy. Arch. Neurol. 33:530, 1975, with permission. Copyright 1975, American Medical Association.)

internode and less in the third to the last internode. Lubinska observed wrinkling, folding, and invagination of the myelin of the terminal internodes. After paranodal demyelination, the demyelinated segment was replaced by short “intercalated” remyelinated segments. Lubinska likened these intercalated internodes to those found after recovery from periaxonal neuritis produced by lead, alcohol, and mechanical trauma63,64 and by other factors.8 She observed that the segmental demyelination occurred within days of crush and was quickly followed by repair. The mechanisms for the myelin remodeling were not explored further. The retrograde teased fiber effect of peroneal nerve transection and distal nerve avulsion was studied in the rat by Spencer and Thomas.146 Their interest focused on portions of teased fibers with myelin swellings or “bubbles.” The nerve above and below the bubble was reported to have a normal appearance but was composed of short internodes of smaller diameter and with thinner myelin than normal. Transverse sections through the myelin

swellings revealed that an axon ran through the swelling, partly surrounded by thin myelin and partly without myelin. The myelin had been split at the interperiod line to form large vacuoles containing myelin debris and macrophages. The authors reported that the bubbles formed on remyelinated portions of a fiber but were uncertain about the cause of these bubbles, although they noted similarities to alterations occurring in experimental tin intoxication and experimental allergic encephalitis. In a later report, Spencer144 stated that there was fiber attenuation and suggested that fiber degeneration starts distally and retrogresses to the cell body. An attempt to quantitate fiber attenuation by regressing internodal length on fiber diameter, however, failed to demonstrate a significant difference in the line for amputated and control nerves. The study of the retrograde effects of permanent axotomy in cats confirmed the correctness of the concept that axonal atrophy leads to secondary demyelination and axonal degeneration.

Pathologic Alterations of Nerves

Permanent Axotomy Studies The seventh segmental nerve and sciatic nerve stumps were studied in groups of cats 3 weeks and 4, 12, and 24 months after hind limb amputation. The sequential changes that MFs undergo were readily demonstrated in teased fibers (Figs. 32–66 and 32–67). At 3 weeks there was a statistically significant increased percentage of stump fibers showing myelin wrinkling. By 4 months a significantly increased percentage of amputated teased fibers showed wrinkling even in the proximal nerve. At this time also the percentage of teased fibers showing demyelination and remyelination was increased in the distal segment of the amputated side compared to the nonamputated side. By 12 months the frequency of demyelinated and remyelinated fibers had reached statistical significance. By 24 months significant percentages of teased fibers were undergoing axonal degeneration in both proximal and, to a greater degree, distal segments of the stump nerves.

FIGURE 32–66 Pathologic abnormalities of teased fibers derived from amputated proximal nerve stumps and comparable levels of nonamputated nerve. A, Four consecutive lengths of a teased fiber from the nonamputated side without abnormality (type A). B, Two consecutive lengths of fiber showing a slight to moderate degree of myelin wrinkling sometimes seen at 3 weeks but more characteristically seen 4 months after amputation (type B). B⬘, Three consecutive lengths of fiber showing severe myelin irregularity characteristic of the change seen at 4 and 12 months (type B). C, Severe myelin wrinkling and partial nodal lengthening at 12 months (type C). D, Three consecutive lengths of fiber showing myelin wrinkling, demyelination, and remyelination at 12 months (type D). F, Four consecutive lengths of fiber showing myelin wrinkling and segmental remyelination, located approximately at the former nodes of Ranvier, at 12 months (type F). (From Dyck, P. J., Lais, A. C., Karnes, J. L., et al.: Permanent axotomy, a model of axonal atrophy and secondary segmental demyelination and remyelination. Ann. Neurol. 9:575, 1981, with permission.)

787

Insight regarding the change in size of MFs and their components is gained from inspection of histograms of MF diameter, myelin areas, and axon areas of the L7 segmental nerve and the sciatic nerve (Fig. 32–68). The most severe alteration appears to be the decrease in axonal area. Assuming the primary event to be axonal atrophy, regression lines relating axonal area to myelin thickness in amputated as compared to nonamputated sciatic nerves are revealing (Fig. 32–69). Axons are smaller, relative to myelin thickness, in axotomized nerves than they are in control nerves. To test whether decreased NF production might underlie the atrophy after permanent axotomy, the number of NFs was regressed on axonal area, number of myelin lamellae, or myelin spiral length of proximal and distal segments of control and axotomized sciatic nerves. Regression lines relating numbers of NFs to axonal area were not significantly different between control and

788

Pathology of the Peripheral Nervous System

These results led to the following conclusions: 1. Permanent axotomy by hind limb amputation provides an informative model of chronic axonal atrophy and of myelin changes. 2. Permanent axotomy confirms conclusions, based on study of nerves of patients with uremic neuropathy and Friedreich’s ataxia, that MFs might undergo degeneration by the following sequential steps: Axonal atrophy : myelin wrinkling : segmental demyelination and remyelination : further atrophy : axonal degeneration.

FIGURE 32–67 Average percentages of teased fiber pathologic abnormalities at various intervals of time after amputation (A) compared with contralateral control (C). A, No abnormality. B, Myelin wrinkling. C, D, and F, Demyelination and remyelination as described in the text. E and H, Axonal degeneration and axonal regeneration. After axotomy, fibers undergo myelin wrinkling, then at a later time demyelination and remyelination, and finally axonal degeneration as described in the text. The changes begin in distal nerve and are more severe in distal (sciatic) than in proximal (segmental) nerve.

amputated nerves. This finding indicates that the density of NFs is unchanged even in atrophic fibers. Regression lines relating numbers of NFs to myelin spiral length were significantly lower for amputated than for nonamputated nerves at both proximal and distal levels. This result suggests that NF number is decreased in atrophic fibers. An explanation for axonal degeneration is needed. It may be that, when the axon becomes very attenuated, it no longer can sustain a distal axon and degenerates. Alternatively, an infolded loop of myelin can cause demyelinative internal strangulation (Fig. 32–70).

3. The events of axonal atrophy, myelin remodeling, and axonal degeneration begin and appear to be more severe in the distal nerve segment but later also develop in the proximal nerve segment. 4. The myelin changes can be explained by an attenuation of the volume of the axon in the presence of a mostly unchanged myelin membrane sheet, which therefore becomes too large for the axon. With a further decrease in axonal caliber, the wall collapses radially, so that flattened and crenated shapes develop in a transverse profile. With retraction of Schwann cells from the node, further axial wrinkling and infolding develop. Clefts, sometimes beginning at Schmidt-Lanterman incisures, are thus formed in compact myelin. 5. Two mechanisms are suggested by which axonal atrophy may lead to axonal degeneration. In the first, the decrease in NF content of the axon reaches a critical level below which axon integrity cannot be maintained. In the second, myelin infoldings compress the axon, resulting in internal axonal strangulation (described earlier). 6. Decreased numbers of NFs could explain the axonal attenuation induced by permanent transection. Decreased synthesis of NF components after nerve transection has been reported by Hoffman and Lasek.80

SEGMENTAL DEMYELINATION AND REMYELINATION Myelin Remodeling in Normal Nerves During development, Schwann cells come to occupy short territories along axons. Each of these Schwann cells forms an internode of myelin. The place at which adjacent internodes abut is the node of Ranvier. With lengthening of the limb during growth, internodes also lengthen. In normal nerve, the Schwann cells of internodes appear to remain relatively unchanged in form over many years—sometimes for a lifetime. This conclusion seems to follow from the observation that the mean ILs of MFs of nerves remain approximately the same for

Pathologic Alterations of Nerves

FIGURE 32–68 Composite histograms of myelinated fiber diameters (left), myelin areas (middle), and axon areas (right) of nonamputated (control) and amputated segmental (upper five rows) and sciatic (lower two rows) nerves of cats to show the altered size distribution that follows axotomy, as described in the text.

FIGURE 32–69 Common regression lines and regression line intercepts for individual nerves of ln axon area to myelin thickness (left) and of index of circularity (IC) to myelin thickness (right) for nonamputated (⫹ symbols and continuous lines) and amputated (open circles and broken lines) sciatic nerve at 24 months after amputation. Both axonal area and IC are lower for axotomy nerves than for controls affecting fibers with all myelin thicknesses.

789

790

Pathology of the Peripheral Nervous System

FIGURE 32–70 Left, Partial view of a myelinated fiber showing the compression that an infolded loop of myelin may produce on the axon. This may be sufficient, in some cases, to cause axonal interruption and distal regeneration. Right, A partial myelin cleft that appears to begin at a Schmidt-Lanterman incisure (arrow).

most of adult life and probably decrease in old age and the fact that, in health, myelin membrane protein and lipids are incorporated at a very low rate. Actually, with age an increasing percentage of paranodal regions may break down. New short intercalated internodes are formed in these demyelinated paranodes. The term segmental demyelination, as used in this chapter, refers to degeneration of the myelin of a paranode (paranodal segmental demyelination) or of an internode (internodal segmental demyelination). As described earlier, absence of paranodal or internodal myelin may be developmental, physiologic (at a low frequency even in healthy persons), primary (metabolic, immune, compression, and other causes), or secondary (to axonal enlargement or atrophy). It is noteworthy that, for most demyelinating neuropathies, remyelination occurs concurrently with demyelination. This suggests that the factors that induce demyelination are presumably not the same ones that would interfere with remyelination. In an experimental study, after nerve crush, regeneration and remyelination occurred as in control animals even when the experimental group had circulating antibodies to galactocerebrosides and while experimental allergic neuritis was being induced.149

Segmental demyelination was first described by Gombault.63,64 He emphasized the segmental nature of the process, integrity of the axon, and smaller ovoids and lipid droplets than in wallerian degeneration. Since the time of Gombault, several varieties of primary segmental demyelination (macrophage induced, pressure induced, antibody induced, and caused by faulty composition of myelin) have been recognized. In addition, a secondary type of segmental demyelination resulting from axonal enlargement or atrophy is now recognized. In human neuropathies, this secondary type of segmental demyelination is common. When confronted with nerves showing segmental demyelination, it therefore now becomes necessary to know whether it is primary (extrinsic or intrinsic events affecting Schwann cells or myelin) or secondary (to axonal enlargement or atrophy). Gombault’s criteria would not discriminate between primary and secondary demyelination.

Differences from Secondary Demyelination Primary and secondary demyelination both show: (1) paranodal or internodal demyelination, (2) remyelination, (3) small myelin breakdown products, (4) intact axons, and (5) normal ultrastructural features of axons.

Pathologic Alterations of Nerves

Three major features distinguish secondary from primary demyelination. In secondary demyelination: (1) the axons of old internodes are (or were) smaller or larger than they should be considering myelin thickness; (2) demyelination and remyelination occur especially on fibers with severe axonal attenuation; and (3) among old internodes, demyelination and remyelination are clustered and not random. A too-small axon may be recognized by infolded loops of myelin between myelin and axon, severely attenuated axons not caused by hyperosmolar fixation, and less-steep regression lines relating axonal area to number of myelin lamellae or myelin spiral length. It should be appreciated that, with myelin remodeling, many fibers may be clothed with newly developed myelin, making interpretation of regression lines difficult. For evaluation of nerve biopsy specimens, it may be too time consuming or difficult to assess for randomness of demyelination among old internode territories. A simpler approach may suffice for clinical purposes. Secondary demyelination should be considered when (1) there is no obvious pathologic alteration that is primary (e.g., inflammation or infiltration); (2) there is unequivocal fiber loss; (3) demyelinated axons (larger than unmyelinated fibers) are not seen; and (4) some teased fibers have several or many old internodes with demyelination and remyelination, whereas many other fibers are normal (clustered demyelination). Irrespective of the type of demyelination, axons do not remain bare except at nodes of Ranvier and quickly become surrounded by Schwann cell cytoplasm and its cytoplasmic membrane. Repeated demyelination and remyelination, whether primary or secondary, result in the formation of onion bulbs and increased collagen synthesis. It has been possible to create onion bulbs experimentally by intoxication with lead99 or phenanthracene162 and repeated application of a pneumatic cuff.32 Dyck postulated the following events: (1) partial or complete demyelination of an internode, (2) mitosis of a Schwann cell in response to demyelination, (3) capture of a demyelinated internode by one Schwann cell, (4) circumferential orientation (influenced by the persistence of an earlier basement membrane) of the displaced Schwann cell and its cytoplasm, and (5) successive outward displacement of layers of basement membrane and Schwann cells.32

DEBRIS REMOVAL FROM THE ENDONEURIUM As outlined in Chapters 3, 29, and 30, nerve fascicles are maintained in a protected microenvironment by the structure and properties of the perineurium, the capillary endothelium, and the anatomic structure at the proximal and distal ends of peripheral nerves. No lymphatic channels are found within the endoneurial space.

791

Our studies, using the dye patent blue (National Aniline Division of Allied Chemical Corporation) to map lymphatic channels, do not provide evidence of intraneural lymphatic channels and indicate that dye removal is delayed when the agent is injected intraneurally (unpublished results). When dye is injected into the epineurial tissue, it is generally removed in several hours and the lymphatic channels by which it is carried to regional lymph nodes can be traced. When microinjections of 2 ␮L of the dye are made endoneurially into sciatic nerves of 300-g rats, the dye spreads almost instantaneously up and down the nerve for a distance of 4 to 6 cm, stains the entire endoneurial contents, and remains unchanged for many hours. After about 6 hours the staining becomes lighter, but pale staining may persist for 2 to 3 days. No lymphatic channels leaving the nerve were identified in several dozen experiments of this kind. In addition to demonstrating the lack of obvious lymphatic channels, these studies indicate that, after injection of minute amounts of fluid into the endoneurial pool, the fluid spreads over a very large distance. Removal of this small volume of dye is greatly delayed compared with removal of dye via lymphatic channels from the epineurial space. In far-advanced lead neuropathy resulting from feeding 4% inorganic lead carbonate in the diet to 200-g rats, macrophages often line up in rows on the inside of the perineurium and (to a lesser extent) around capillaries (Fig. 32–71). These macrophages contain lipid debris. In addition, macrophages may be found between all layers of perineurium and even just inside the last lamella. The perineurial cells may also contain extensive lipid droplets. These findings raise the question of whether macrophages may unload lipid debris to perineurial cells or simply pass through the perineurium and enter the lymphatic system. Subperineurial palisading of vacuolated cells, identified as fibroblasts, was first described by Schoene and co-workers134 in cases of hereditary sensory neuropathy. Various Schwann cell inclusion bodies may be involved in degradative events. The ␲ granules of Reich are elongated, comma- or rod-shaped structures that are found at paranodal regions and at polar ends of Schwann cells.27,47,59,134,152,154 These granules stain metachromatically with basic aniline dyes. With the electron microscope they are seen to be membrane bound and have a periodicity of 4 to 5 nm. Their frequency and size are reported to increase with age and possibly in various disease states.135 We distinguish the ␲ granules of Reich127 from the ␮ granules of Elzholz,57 although the ␲ granules of Reich are commonly situated at paranodal sites in Schwann cell cytoplasm just outside compact myelin. They stain metachromatically with aniline dyes and appear as short rods or curvilinear structures surrounded by a double membrane. Elzholz bodies are osmiophilic and lie adjacent (on the outside or inside of compact myelin). They have a lamellar structure whose

792

Pathology of the Peripheral Nervous System

FIGURE 32–71 Transverse semithin sections of peroneal nerve from animal intoxicated with 4% lead carbonate for 6 months. A, Low-power view showing the subperineurial margination of macrophages. The arrow indicates a region of the perineurium that is very attenuated. B, A macrophage is situated in the perineural defect. Serial sections confirmed the presence of this defect. C, Multinucleated giant cell, presumably a fused macrophage cell. D–G, Other segments of the subperineurial and perineurial regions, showing lipid deposits in perineurial cells, perineurial separation, macrophages between perineurial leaflets, and subperineurial edema.

periodicity is about one third less than that of myelin. As shown in Figures 32–52 and 32–53, such inclusions may also be situated between the axolemma and the myelin sheath. Spencer and Thomas147 postulated the existence of a mechanism for the sequestration and phagocytosis of axoplasmic organelles by Schwann cells in hexacarbon-

induced neuropathies. They suggested that, adjacent to an organelle to be removed, the adaxonal cytoplasm forms a thin ridge that indents the axon. Next, this sheet of adaxonal cytoplasm surrounds the axon, infolds on itself, and sequesters groups of axoplasmic organelles to form an interdigitating profile when viewed in cross section. A possible example of this phenomenon is shown in Figure 32–65.

793

Pathologic Alterations of Nerves

NERVE REGENERATION Soon after focal nerve injury, regenerative events occur. The transected axons become sealed off and are transformed into growth cones that elongate along former Schwann tubes (see Fig. 32–60). Here, they lie among Schwann cell cytoplasmic processes and macrophages in former or old Schwann cell basement membrane (SCBM) tubes. The cluster of axonal sprouts, Schwann cells, and macrophages enclosed in an SCBM tube are the counterpart of what light microscopists called bands of Büngner. These bands of Büngner may be subdivided into those with neurites (myelinated, amyelinated, or unmyelinated) and those without neurites. The key to their identification would be the presence of former SCBM or remnants of SCBM. With time the old SCBM disappear, and regenerating nerve fiber clusters are recognized by their close apposition to one another. Schwann cells multiply inside SCBMs, maintaining their columnar orientation. During early phases of regeneration, proximal segments of regenerating axons become ensheathed by Schwann cells and myelination begins. Usually more than one regenerating neurite is found in one SCBM. Sometimes in abortive regeneration, many (as many as 15 to 20 or more) neurites will be found in a cluster—“regenerative nerve fiber clusters.” We assume that at a later time the number of regenerating axons may be decreased from the earlier value, suggesting that axons that do not innervate target tissue disappear. The completeness of nerve regeneration after nerve degeneration caused by a focal injury depends on the species, age, site of injury, type and degree of complexity of the lesion, type of repair, and interval between the time of the injury and the time of the repair. Our knowledge about adequacy of regeneration has come especially from study of the distal nerve stump at various times after nerve crush, transection and suture, cable grafting, and permanent transection. The striking difference between regeneration after crush (restoration to nearly normal number and size distribution) and that after transection and suture (poor restoration of number and size distribution) is probably related to events and mechanisms at the site of injury, because the distal stump mechanism should not be different between the two. With nerve transection the blood supply is also interrupted, the endoneurial environment is exposed to the epineurial environment, and the integrity of the perineurium and SCBM tubes is breached. By contrast, in crush, vessel continuity is maintained and the perineurium and SCBM tubes remain uninterrupted. Nerve regeneration is always worse after transection than it is after compression, in torn and in macerated lesions than in clean transections, in proximal than in distal lesions, in nerves sutured under tension than without tension, and probably in the old than in the young. To illustrate, in severe brachial plexus lesions virtually

no recovery below the elbow can be expected, whereas virtually complete recovery is usual from a severe compression injury whether proximal or distal. Because degenerative and regenerative events in the distal stump should be the same whether the nerve is cut or crushed, whether there is a gap or not, and so on, the difference must be related mainly to events at the site of injury. Thus factors that influence regenerative outcome probably include (1) the number of neurites growing into the distal stump, (2) whether they are directed to an appropriate or inappropriate target (which they cannot innervate, and therefore they fail to develop or undergo atrophy and degenerate), and (3) whether they are so delayed in their outgrowth to distal limbs that SCBM tubes become sufficiently disorganized that neurites frequently do not regrow to appropriate targets. A theoretical possibility is that, when regeneration is delayed, the target tissue itself can no longer be reinnervated because it has degenerated. In a study of Schwann cell tubes at various times after permanent nerve transection in mouse, changes in the tubes have been described that may help explain poor regeneration (Figs. 32–72 through 32–75). Initially during fiber degeneration, the SCBM has dilated and collapsed regions. Increasingly the SCBM is no longer in close approximation with Schwann cell cytoplasmic basement membrane, thus losing a nutritive association. Even within a few weeks the SCBM begins to fragment. Later the fragmented pieces become dispersed in the endoneurial space. It seems likely that these changes in the normal scaffold for regrowth are sufficient alteration that neurites may not successfully regenerate back to original or appropriate targets. There is suggestive evidence that a species of laminin that is found on the inner surface of SCBM might be important in guidance of regenerating axons. It has been suggested that SCBMs from patients with diabetic neuropathy are less corrugated and collapsed than are those in controls and may persist longer before degenerating.91

Microfascicular Regeneration After nerve transection microfascicles, each containing only a few fibers, may regrow outside or within the main nerve trunk. Korthals and co-workers94 have described an extreme degree of microfascicular degeneration after ischemic nerve injury (Fig. 32–76) in cat.

INTERSTITIAL PATHOLOGY OF NERVES Pathologic alterations of peripheral nerves may be classified into parenchymatous—alterations of the parenchyma (neurons [axons] or Schwann cells [myelin])—or interstitial— alterations of the interstitium. Under interstitial alterations we include inflammatory, infiltrative, mechanical, or entrapment injury and vascular and ischemic injury.

794

Pathology of the Peripheral Nervous System

FIGURE 32–72 A low-power electron micrograph of peroneal nerve of mouse to illustrate the Schwann cell basement membrane (SCBM), which is closely applied to the external surfaces of Schwann cells that invest individual myelinated fibers and unmyelinated fiber bundles. Inset, A tracing of the SCBM at low power. Because the SCBM is continuous from one internode to the next, one must understand this structure to be a sleeve that extends from one end of a peripheral nerve fiber to its other end. Containing as it does a chain of many Schwann cells along its length, after degeneration the sleeve serves to maintain the longitudinal orientation of Schwann cells and to act as a conduit for neurite regrowth. If the nerve is crushed, leaving the SCBM intact, there is a better chance for the neurite to grow back to its original target. If the nerve is cut, a neurite may grow down to an inappropriate target and fail to innervate it; as a result, the regenerating fiber does not develop or undergoes atrophy. (From Giannini, C., and Dyck, P. J.: The fate of Schwann cell basement membranes in permanent transected nerves. J. Neuropathol. Exp. Neurol. 49:550, 1990, with permission.)

Edema and Hemorrhage Although edema of the endoneurial space is characteristic of certain neuropathies, there is little evidence that it, or the associated increased endoneurial fluid pressure, directly causes fiber damage (see Chapters 29 and 30). An exception may be the sudden accumulation of endoneurial fluid resulting from accidental needle injection into nerve45 or from hemorrhage into nerve.105 In most chronic neuropathies with associated edema, there is sufficient compliance that damaging levels of pressure do not develop. Interstitial edema may be severe in experimental lead and galactose neuropathies. Because comparable levels of increased endoneurial pressure are associated with frequent segmental demyelination in lead neuropathy and infrequent axonal degeneration in galactose neuropathy, mechanisms other than edema must be postulated for the fiber degeneration. Interstitial edema is also seen in various types of inflammatory demyelinating polyradiculo-

neuropathies, in certain monoclonal protein-associated and paraneoplastic neuropathies, in myxedema, in acromegaly, and in certain inherited neuropathies (e.g., Déjérine-Sottas disease). In tissue sections, edema must be distinguished from artifactual tissue separation resulting from histologic processing. In well-fixed and processed nerve specimens with intact perineurium, edema fluid accumulates around vessels and beneath the perineurium and may diffusely separate fibers. The total transverse fascicular area of edematous nerves is greater than that of control nerves. Generally, edema fluid also becomes stained with protein stains. Hemorrhage into nerve may be artifactual, as occurs at the time of surgery, or may be due to disease. Hemorrhage at the edges of the biopsy specimen, without inflammatory reaction, is probably related to transection of vessels during biopsy. Disease bleeding into nerve may occur because

Pathologic Alterations of Nerves

FIGURE 32–73 In acute wallerian degeneration, the Schwann cell basement membrane is focally dilated (A) and dilated and collapsed (B). Myelin debris lying free in the endoneurium is seen in A. (From Giannini, C., and Dyck, P. J.: The fate of Schwann cell basement membranes in permanent transected nerves. J. Neuropathol. Exp. Neurol. 49:550, 1990, with permission.)

FIGURE 32–74 Electron micrograph of a collapsed Schwann cell basement membrane (SCBM), showing redundant basement membrane and Schwann cell processes. This is from a permanent axotomy mouse model. The processes in the SCBM are Schwann cell processes without neurites. (From Giannini, C., and Dyck, P. J.: The fate of Schwann cell basement membranes in permanent transected nerves. J. Neuropathol. Exp. Neurol. 49:550, 1990, with permission.)

795

796

Pathology of the Peripheral Nervous System

FIGURE 32–75 At progressively longer times after permanent axotomy (A to D), the Schwann cell basement membranes (SCBMs) fragment and are dispersed. Some of the basement membrane disappears over 3 months. Although the SCBM sleeve was found to remain continuous over a period of 3 months, by this time there were many discontinuities along its length and there was dispersion. Some neurites may now follow original SCBMs to original targets because of this fragmentation and dispersion. (From Giannini, C., and Dyck, P. J.: The fate of Schwann cell basement membranes in permanent transected nerves. J. Neuropathol. Exp. Neurol. 49:550, 1990, with permission.)

797

Pathologic Alterations of Nerves

FIGURE 32–76 Microfasciculation following nerve ischemic lesion at low power (A) and higher power (B). (From Korthais, J. K., Gieron, M. A., and Wisniewski, H. M.: Nerve regeneration patterns after acute ischemic injury. Neurology 39:932, 1989, with permission.)

of trauma, necrotizing angiopathy, hemorrhagic diathesis, and septic emboli. In necrotizing angiopathy the point of hemorrhage seldom is identified in the sections evaluated (Fig. 32–77). Ischemic damage of capillaries or venules may be sufficient to cause diapedesis or frank hemorrhage. Hemosiderin in macrophages provides evidence that the hemorrhage occurred antemortem or prebiopsy. Finding hemosiderin in macrophages in nerve usually implies diseases of vessels and is an important finding, and is common in necrotizing vasculitis (see Fig. 32–77).

Inflammation of microvessels with neural separation and bleeding occurs in necrotizing vasculitis of microvessels. This may occur in Sjögren’s neuropathy, nonsystemic vasculitic neuropathy, and diabetic or nondiabetic lumbosacral radiculoplexus neuropathy (Figs. 32–78 and 32–79).

Compression The subject of compression is reviewed in Chapters 55, 56, and 57. The terms acute compression and entrapment are

798

Pathology of the Peripheral Nervous System

FIGURE 32–77 Alterations in nerve trunks typical of periarteritis nodosa. A, Gross hemorrhage into nerve trunk. B, Occluded small artery surrounded by hemorrhage. C, Occlusion of lumen and segmental absence (necrosis) of media of small artery. D, Complete absence of media and perivascular infiltrate on mononuclear cells of small artery. E, Mononuclear infiltrate around small vessels. F, Hemosiderin macrophages at the site of a severely damaged artery.

not synonymous. The term compression is reserved for a monophasic application of force exerted to the nerve from the outside so that it is compressed against a rigid underlying structure. Acute compression may also occur repeatedly. Entrapment refers to the injury that ensues when a nerve passes through an opening that is too small considering the caliber of the nerve. Acute or repeated compression, stretch, and other factors may be operative. If the nerve is tethered

because of previous injury, stretch, repeated injury, scarring, or other mechanisms may be involved. The characteristic clinical features of nerve injury arising from application of a tourniquet to an arm are (1) muscle weakness overshadows other neuropathic manifestations, (2) recovery may begin in days or weeks, and (3) eventual recovery is expected. Nerve crush is a more severe form of nerve compression. In both compression and crush, the nerve

Pathologic Alterations of Nerves

799

FIGURE 32–78 Serial skip paraffin sections of a microvessel above (upper), at (middle), and below (lower) a region of microvasculitis in the sural nerve of a patient with nondiabetic lumbosacral radiculoplexus neuropathy. The sections on the left are stained with hematoxylin and eosin, the sections in the middle are reacted to anti–human smooth muscle actin (Dako), and the sections on the right are reacted to leukocytes (leukocyte common antigen [LCA]). The smooth muscle of the tunica media in the region of microvasculitis (middle) is separated by mononuclear cells, fragmented, and decreased in amount. The changes are those of a focal microvasculitis. (Reprinted with permission from Dyck, P. J. B., Engelstad, J., Norell, J., and Dyck, P. J.: Microvasculitis in non-diabetic lumbosacral radiculoplexus neuropathy [LSRPN]: similarity to the diabetic variety [DLSRPN]. J. Neuropathol. Exp. Neurol. 59:525, 2000, with permission.) See Color Plate

fascicles and the nerve basement membrane tubes remain more or less in continuity. By contrast, in nerve transection the nerve fascicles and the nerve fiber tubes are severed and, on resuturing, cut ends are misaligned. In nerve crush virtually all fibers undergo axonal degeneration, but regeneration is much better than after nerve transection and reconnection or after cable grafting. Nerve regeneration after crush injury restores fiber number and size distribution almost to normal. It is thought that very little misdirection of fibers occurs. By contrast, nerve regeneration after nerve transection usually does not restore fiber number or size distribution to normal. In nerve transection, not only fibers but also the gross nerve,

blood supply, fascicles, and Schwann tubes are transected. Even with surgical reconnection there is failure to establish blood supply immediately, fascicles are not adequately aligned, a gap develops between the proximal and distal stump, and SCBM tubes are not correctly apposed to their distal stumps. The effect is that many fibers do not regrow to their original targets or to appropriate targets and so do not innervate tissues and may degenerate. Injury caused by nerve compression is less severe than that caused by nerve crush. The spectrum of alterations may vary from reversible physiologic block to paranodal or internodal demyelination to frank degeneration.

800

Pathology of the Peripheral Nervous System

FIGURE 32–79 Transverse semithin epoxy sections of sural nerve from patients with nondiabetic lumbosacral radiculoplexus neuropathy stained with methylene blue showing multifocal fiber degeneration and loss (top) and reactive repair injury neuroma (bottom), which we attribute to ischemic injury and repair. Top, The upper left fascicle shows an admixture of normal and degenerating (arrowhead) fibers. The lower left fascicle shows an almost normal density of fibers and an occasional degenerating fiber. In contrast, the lower and upper right fascicles are essentially devoid of intact fibers and only rare degenerating profiles remain. Bottom, Transverse section of part of a fascicle showing nerve regeneration. There is a crescent of microfasciculi (injury neuroma, short thick arrow) between the leaflets of thickened perineurium (edges demonstrated by long, thin arrows). The arrowhead identifies a cluster of regenerated fibers. Most of the fibers in the large fascicle are small myelinated fibers and are probably regenerating. (Reprinted with permission from Dyck, P. J. B., Engelstad, J., Norell, J., and Dyck, P. J.: Microvasculitis in non-diabetic lumbosacral radiculoplexus neuropathy [LSRPN]: similarity to the diabetic variety [DLSRPN]. J. Neuropathol. Exp. Neurol. 59:525, 2000, with permission.)

The mechanism of nerve injury resulting from compression was formerly thought to involve ischemia. DennyBrown and Brenner26 compressed nerves with a spring clip and found that severity of injury was related to duration of compression. Direct compression was not thought to be the essential factor because excised segments of nerve in a

chamber could be severely compressed without their function being lost.70 However, this argument is no longer considered persuasive because shear forces are not operative when nerves are compressed in a chamber. By contrast, in pneumatic cuff compression major shear forces develop at the edge of the cuff. In a series of studies, Ochoa and co-workers116,117 emphasized the role of mechanical injury. They suggested that axons had been telescoped into adjacent internodes with actual displacement of nodes of Ranvier. The loops of myelin that overrode nodes of Ranvier later degenerated and resulted in demyelination. Further evidence that ischemia was not the responsible mechanism was that compression injury was not worsened by a preceding period (of an hour) of leg ischemia.161 Studies by Dyck and colleagues provided definitive further evidence that the major functional and structural alterations of nerve compression were due to mechanical and not ischemic injury.44 As reviewed in other parts of this chapter, histologic alterations attributable to ischemia take from 12 to 24 hours to be expressed. Any structural alterations that are already fully expressed within a few minutes of tourniquet application, therefore, cannot be attributed to ischemic injury.114 In our experiments, histologic alterations were stopped within a minute or so of cuff application by perfusion fixation and by flooding the nerve with fixative. The stereotypical alterations of compression injury were fully expressed within a few minutes; therefore they cannot be due to ischemia. Under the cuff the following sequential changes were observed: (1) the endoneurial fluid was expressed from under the region of the cuff, resulting in close apposition of fibers and of cellular elements (Fig. 32–80); (2) fluid was also expressed from the axon with compaction of cytoarchitectural elements; (3) with greater compression, cytoskeletal elements were dislocated and expressed; and (4) internodes were lengthened and shear injury of myelin occurred. At the edges of the cuff, additional changes occurred: (1) increase of endoneurial area from expression of endoneurial fluid from under the cuff (Fig. 32–81); (2) increase of axonal fluid causing distention of axons, especially at nodes of Ranvier; (3) widening and dilatation of nodes of Ranvier; and (4) cleaved myelin loops appearing to overlap nodes of Ranvier. These results provide strong evidence that stress force causes these pathologic changes. If cuff compression is maintained for long times, ischemic injury may contribute to the pathologic changes described. We noted further that the histologic alterations resulting from compression are quite different in time course, anatomic distribution, and histologic changes from those caused by ischemia.

Ischemia A schematic view of the blood supply to nerves is shown in Figure 32–82. The human sciatic nerve, at its upper end, receives an arterial branch from the inferior gluteal

Pathologic Alterations of Nerves

FIGURE 32–80 Electron micrographs of transverse section of peroneal nerve of rat showing the compaction of cytoskeleton and organelles of axons that occurred under the cuff when the peroneal nerve was compressed at 300 mm Hg for 2 minutes (top) and for 2 hours (bottom). Inset, Low-power view (⫻4800) of a transverse section of a myelinated fiber showing the rectangular area at a higher magnification (top) (⫻38,300). The ultrastructural preservation of microtubules and of cristae of mitochondria provides evidence that fixation is adequate (bottom). In the transverse section (⫻7700) of the fiber, myelin is split and crumpled and axoplasm is not seen. The pressure that induced these changes and their sites and mechanisms of the changes are described in the text. (From Dyck, P. J., Lais, A. C., Giannini, C., and Engelstad, J. K.: Structural alterations of nerve during cuff compression. Proc. Natl. Acad. Sci. U. S. A. 87:9828, 1990, with permission.)

artery. In angiograms at autopsy, this arterial branch may be visualized for 10 to 15 cm down the sciatic nerves. The distal tibial and peroneal nerves receive their blood supply from popliteal branches. Between these levels, the distal sciatic and proximal tibial and peroneal nerves receive a variable arterial supply from branches of the deep femoral artery. The epineurial arteries branch, and

801

their arteriolar and precapillary branches penetrate the perineurium to perfuse the endoneurium. Arteries and veins have three layers of cells: tunica intima, tunica media, and tunica adventitia. Elastic arteries in nerve are infrequent and of large caliber. Their tunica media is made up of smooth muscle and elastic fibers. Muscular arteries vary from 200 ␮m to several millimeters in diameter. Most, especially the large and intermediate vessels, have an internal elastic lamina. A clear boundary does not distinguish large arterioles from small muscle arteries, although in some older textbooks 100 ␮m was set as the boundary. The smallest arterioles (precapillaries) have only endothelial cells, one or two layers of smooth muscle cells, and pericyte cells. Capillaries have only a few abutting endothelial cells and associated pericytes. Vessels may also be classified by function. The large elastic and muscular arteries might be thought of as conduits (conducting and distributing vessels) from the heart to the tissue. Arterioles provide regulation of blood supply to tissue and resistance to modulate blood pressure (regulating and resistance vessels). Capillaries are exchange vessels. Muscular venules and veins are capacitance and returning vessels. In nerve essentially only the epineurial and perineurial compartments have arterioles and small arteries 50 to 400 ␮m in diameter. That is why the characteristic pathologic abnormalities of necrotizing angiopathy must be sought in the epineurium and why whole-nerve biopsy must be performed to be able to diagnose necrotizing vasculitis of these vessels. Typical changes of small artery and arteriole necrotizing vasculitis, as occur in polyarteritis nodosa, are seen in Figures 32–77 and 32–83 (also see Chapters 106 and 107). The epineurium tends also to have the larger collecting veins. The capillaries of epineurium do not have tight junctions between endothelial cells, and tracers such as horseradish peroxidase pass freely between cells. The arterioles penetrating the perineurium are small, usually less than 50 ␮m. The vessels in the endoneurial space include only an occasional small arteriole (25 to 75 ␮m in diameter). The endoneurial vessels are precapillaries, capillaries, venules, and thin-walled veins. The endothelial cells of capillaries abut against each other and form tight junctions not allowing horseradish peroxidase to pass between them, providing a morphologic basis for the blood-nerve barrier. Three vascular processes typically cause ischemia of nerve. The first is large-vessel occlusive disease or emboli. Ligation of a single large limb vessel does not cause ischemic injury—several (actually many) must be ligated to induce ischemic injury. The second process that may cause nerve ischemia is necrotizing vasculitis. In fully expressed disease, many epineurial arterioles 50 to 400 ␮m in diameter become occluded at multiple sites, which occurs in such disorders as polyarteritis nodosa, rheumatoid vasculitis, Churg-Strauss syndrome, and Wegener’s granulomatosis (see Chapter 106). The third process is microvascular disease leading to nerve ischemia.

802

Pathology of the Peripheral Nervous System

FIGURE 32–81 Three frames of teased fibers showing normal node from above the cuff (first) and obscured node from under the cuff (second) (upper), and lengthened node from the edge of the cuff (lower). (From Dyck, P. J., Lais, A. C., Giannini, C., and Engelstad, J. K.: Structural alterations of nerve during cuff compression. Proc. Natl. Acad. Sci. U. S. A. 87:9828, 1990, with permission.)

There is good autopsy and nerve biopsy evidence that peripheral vascular disease (large artery occlusive disease) may result in ischemic nerve damage, but the nerve damage is in the general region and roughly proportionate with distal limb gangrene. There is little evidence that peripheral vascular disease is the basis of or a major contributor to diffuse polyneuropathy (e.g., old age or diabetic sensorimotor polyneuropathy). Results of experimental ligation of the large arteries to lower limbs76,95,96 confirm and extend these views: (1) very extensive ligation of large arteries are needed to cause ischemic nerve damage to nerves of lower limbs because collateral circulation is extensive and

(2) the lower sciatic and upper tibial and peroneal nerves and central fascicular regions of widely separated fascicles are sites of vulnerability. Study of patients with necrotizing vasculitis and multiple mononeuropathy has provided important information about the spatial distribution of nerve ischemic damage (Fig. 32–84). In necrotizing vasculitis, neurologic deficits may develop acutely first in the territory of one nerve, then in another, and then in others. This pattern of involvement is called “mononeuritis multiplex” or, a preferable term, multiple mononeuropathy. The three-dimensional relationships between occluded arteries and regions of ischemic nerve damage have been studied.34

Pathologic Alterations of Nerves

803

FIGURE 32–82 Diagrammatic representation of the blood supply to nerve as discussed in the text. (From Dyck, P. J.: Peripheral neuropathy. Postgrad. Med. 41:279, 1967, with permission.)

In several patients, nerve sections from serial tissue blocks of median, ulnar, radial, sciatic, tibial, and peroneal nerves were fixed in buffered formaldehyde; then, removing the entire sciatic nerve and its tibial and proximal branches, serial blocks were cut from along its length from proximal to distal. The first block was embedded into paraffin and the sections stained (hematoxylin and eosin, trichrome, and other stains), and the next block was additionally fixed in Flemming’s fixative (Weigert stain) for fiber alterations. These preparations were repeated such that consecutive blocks along the length of nerve were alternatively studied for vessel alterations and for fiber alterations. Segmental occlusions of small epineurial arteries or arterioles were found to be diffusely distributed along the length of the large nerve trunks. By contrast, fiber degeneration began at mid-thigh levels only in central fascicular or wedge-shaped regions of widely distributed fascicles. In affected fascicles subperineurial fibers tended to be unaffected. At more distal levels the discrete regions of degeneration could no longer be identified, but degenerating fibers were scattered diffusely throughout most fascicles. These postmortem studies demonstrated that (1) small artery or arteriole occlusion occurring at multiple sites along the length of nerve, resulting in nerve fiber damage beginning in a patchy distribution at mid-upper arm and mid-thigh levels; (2) individual occlusion of arteries or arterioles could not be related to regions of ischemic damage, so multiple sites of multiple vessels must be occluded before multifocal fiber loss occurs and ischemic injury occurs at watershed zones of poor perfusion; and (3) axons seemed more vulnerable to ischemia than did Schwann cells.

50μ FIGURE 32–83 Oblique section of sural nerve epineurial arteriole showing florid necrotizing vasculitis. Top, The trichrome stain of a paraffin section shows the bright red fibrinoid degeneration, neural necrosis, and inflammation. Bottom, An adjacent paraffin section has been reacted for CD68 (KP-1), revealing macrophages in vessel walls and in the surrounding inflammatory infiltrates. The changes are diagnostic of necrotizing vasculitis of small arteries and arterioles. See Color Plate

These studies describe the pathologic alterations of the large nerve vessel vasculitides as occur in rheumatoid arthritis, polyarteritis nodosum, Wegener’s granulomatosis, and systemic lupus erythematosus. In contrast, there are small blood vessel vasculitides that include nonsystemic vasculitis of nerve, Sjögren’s syndrome, and the radiculoplexus neuropathies. The radiculoplexus neuropathies are asymmetrical disorders causing pain and weakness in the upper (cervical), trunk (thoracic), and lower (lumbosacral) regions. These are monophasic illnesses that happen to diabetic and nondiabetic patients. The pathologic lesion in these conditions is ischemic injury of nerve (perineurial degeneration and thickening, neovascularization, multifocal fiber loss and degeneration, and microfasciculation) (see Fig. 32–79).55,56 This ischemic injury appears to be due to a microvasculitis of small nerve vessels. Because of these

804

Pathology of the Peripheral Nervous System

FIGURE 32–84 Upper row, Low-power tracings showing the density of myelinated fibers and the fascicular pattern of fiber loss from the median nerve of a patient with rheumatoid arthritis and necrotizing angiopathic neuropathy. These were taken at 4 cm (left), 8 cm (middle) and 12 cm (right) from the axilla. The density of intact myelinated fibers in fascicles was graded as follows: A ⫽ normal numbers; B ⫽ equivocally decreased; C ⫽ moderately decreased; and D ⫽ greatly decreased. Lower row, Epon sections of fascicles as seen in the rectangles in the 8-cm tracing. Note the central fascicular loss of myelinated fibers, surrounded by a borderline zone of degenerating fibers, which in turn is surrounded by fibers of normal density. (From Dyck, P. J., Conn, D. L., and Okazaki, H.: Necrotizing angiopathic neuropathy: three-dimensional morphology of fiber degeneration related to sites of occluded vessels. Mayo Clin. Proc. 47:461, 1972, with permission.)

vessels’ small size, fibrinoid degeneration of the wall is rare. Typically, there is fragmentation and disruption of the vessel wall by mononuclear inflammatory cells that is often very focal (see Fig. 32–78). These studies take on added significance because diabetic third-nerve palsy was found to be due to central fascicular damage, surmised to result from ischemic damage.5,30,158 Which cells in nerve are especially vulnerable to ischemia? Studies by Korthals and colleagues,94 Dyck and co-workers,39 and Nukada and Dyck114 suggested that axons are more vulnerable to ischemia than Schwann cells or myelin. Korthals and colleagues94 found that axons entering an ischemic area became dilated and filled with organelles. They suggested that ischemia resulted in interference with rapid axonal flow, resulting in the accumulation of organelles.

Nukada and Dyck114 studied these changes in more detail (Fig. 32–85). Work by Johnson and co-workers82 and Benstead and colleagues11 found that endothelial cells of capillaries and perineurial cells also were vulnerable to ischemic injury. Can precapillary and capillary disease cause nerve fiber degeneration? The evidence from experimental studies is unequivocal—it can. In human neuropathy the evidence is incomplete and less certain. Precapillary and capillary occlusion can be produced by intra-arterial injection of arachidonic acid122 and by injection of 15-␮m polystyrene microspheres.114 Such studies have demonstrated that central fascicular nerve damage can be produced by capillary occlusion. When these substances are injected into vessels of supply to the sciatic nerve, the vulnerable region is the

Pathologic Alterations of Nerves

805

FIGURE 32–85 Upper, Skip serial electron micrographs of individual myelinated fibers entering an ischemic core of sciatic nerve from an area without pathologic abnormality (proximally) to well-developed ischemic changes distally. Both fibers show the following sequence of events: normal-appearing fibers (1), dark axons with light core (2), enlarged dark axon (3 and 4), attenuated dark axons (5), and missing axon (5 and 6). Lower, The results from serial transverse sections were used to draw a representative fiber entering an ischemic core. On the left, the fiber is normal; in the middle, the axon is enlarged and dark owing to accumulated organelles; and on the right, it has become attenuated and has disappeared. AX ⫽ axon; BM ⫽ basement membrane; MY ⫽ myelin; SC ⫽ Schwann cell. (Modified from Nukada, H., and Dyck, P. J.: Acute ischemia causes axonal stasis, swelling, attenuation, and secondary demyelination. Ann. Neurol. 22:311, 1987.)

central fascicular region of proximal tibial and peroneal nerves. Axons at proximal borders of ischemic areas may be swollen and filled with various organelles.

Intraneural Injection of a Foreign Substance Rarely, the sciatic nerve of humans is directly injected with a medication, sometimes resulting in a severe protracted neuropathic deficit. Injecting penicillin in oil directly into the sciatic nerve may result in a granulomatous reaction. In war, nerves may become lacerated and involved in purulent abscesses. The perineurium provides a barrier to the entry of purulent material into the fascicular contents.

Microinjection of minute amounts of lysolecithin73 and other monochain phospholipids45,102 into a fascicle of nerve causes focal demyelination (Fig. 32–86). Demyelination can be passively transferred by intraneural injection of experimental allergic neuritis serum.132 Passive transfer of demyelination resulting from microinjection of Guillain-Barré serum has been reported,60 but other studies are not convincing.102 The results of these types of studies have to be interpreted with great caution, however, because the technique of injection itself (the size of the needle, the movement of the needle or animal, and the volume and rate of fluid injected) may produce fiber degeneration.45

806

Pathology of the Peripheral Nervous System

FIGURE 32–86 Transverse sections of peroneal nerve of rat to show control (A) and the confluent segmental demyelination from microinjection of 5 nmol/5 ␮L of lysolecithin (B). (From Dyck, P. J., Lais, A. C., Hansen, S. M., et al.: Technique assessment of demyelination from endoneurial injection. Exp. Neurol. 77:359, 1982, with permission.)

Inflammatory Demyelinating Disorders Inflammatory demyelinating polyradiculoneuropathies (IDPs) are usually symmetrical polyradiculoneuropathies caused by immune mechanisms resulting in multifocal demyelination and fiber degeneration, which results in a paretic or paralytic disorder affecting proximal and distal limbs or even bulbar or truncal muscles (especially in acute IDP) and sensation; nerve conduction abnormalities characteristic of segmental demyelination; a cytoalbuminologic dissociation of the cerebrospinal fluid; and mononuclear cell exudates, edema, and segmental demyelination in nerves (Figs. 32–87 through 32–91). The IDP disorders may be subdivided into acute IDP (AIDP; see Chapter 99) and chronic IDP (CIDP; see Chapter 100). Others and we have described other inflammatory demyelinating varieties.6,33 Whether inflammatory demyelinating processes underlie plexus neuropathies, diabetic radiculoplexus neuropathy and multiple mononeuropathy, and prednisone-responsive inherited motor and sensory neuropathy has not been demonstrated, but this remains a possibility. The hallmarks of inflammatory demyelinating disorders are edema, particularly around capillaries and beneath the perineurium; endoneurial perivascular collections of mononuclear cells; macrophage contact with Schwann cells; and macrophage delamination of myelin. With chronicity, onion bulbs are formed. In AIDP, early pathologic reports emphasized the vulnerability of ventral spinal roots,75 but Asbury and co-workers6 found the involvement to be more diffuse and both proximal and distal nerve trunks to be affected. Biopsy of sural nerve may not show

FIGURE 32–87 A cytoplasmic process of a macrophage has inserted itself into the cytoplasm of a Schwann cell. Macrophages make contact with Schwann cells and are though to participate in active demyelination in experimental allergic neuritis. (From Wisniewski, H., Prineas, J., and Raine, C. S.: An ultrastructural study of experimental demyelination and remyelination. I. Acute experimental allergic encephalomyelitis in the peripheral nervous system. Lab. Invest. 21:105, 1969, with permission.)

50μ

25μ

50μ

50μ FIGURE 32–88 Serial transverse paraffin sections from a fascicular sciatic nerve biopsy in a patient with focal chronic inflammatory demyelinating polyradiculoneuropathy. The patient presented with a sciatic mononeuropathy, and on magnetic resonance imaging the sciatic nerve was enlarged and had increased T2 signal. Top, A transverse paraffin section of the biopsied nerve (hematoxylin and eosin staining) shows large epineurial perivascular inflammatory cell infiltrates. In addition, there is a small perivascular endoneurial inflammatory cell collection (arrowhead). Middle, Section immunoreacted for CD45 (leukocyte common antigen) that confirms that the mononuclear cell infiltrates in the epineurium and endoneurium (arrowhead) are lymphocytes. Bottom, Section immunoreacted with CD68 (KP-1) shows that macrophages are present but less common than lymphocytes. Teased fibers from this biopsy showed increased rates of segmental demyelination. These findings suggest that focal demyelinating inflammatory lesions occur and that some mononeuropathies are chronic inflammatory demyelinating mononeuropathies. See Color Plate

FIGURE 32–89 Paraffin and epoxy transverse sections from the greater auricular nerve from a patient with multifocal hypertrophic chronic inflammatory demyelinating polyradiculoneuropathy. She had apparent “tumors” within the brachial plexus and greater auricular nerves and had previously been incorrectly diagnosed as having neurofibromatosis. Top, Transverse paraffin section (hematoxylin and eosin staining) showing an epineurial perivascular mononuclear inflammatory cell infiltrate (right) and diffuse onion bulbs (left). Middle, Dense onion bulbs affecting all fibers with only thin myelin remaining (epoxy, methylene blue stain). Bottom, Large subperineurial mononuclear inflammatory cell infiltrate adjacent to frequent onion bulbs (epoxy, methylene blue stain). See Color Plate

prominent inflammatory change even when there is considerable evidence of segmental demyelination. On the assumption that sensitized hematogenous mononuclear cells invade nerve, Asbury and colleagues6 proposed that these initially caused segmental demyelination and, if more severe, axonal degeneration. Whatever the mechanism, axonal degeneration of distal nerves, such as the radial cutaneous nerve of forearm and sural nerve, may be severe in fulminant cases of AIDP. There is increasing evidence that humoral mechanisms are involved and perhaps are primary and initiate the T-cell

808

Pathology of the Peripheral Nervous System

30μ

FIGURE 32–90 Teased nerve fibers from the sural nerve of a patient newly diagnosed with chronic inflammatory demyelinating polyradiculoneuropathy (CIDP). These are separately teased myelinated fibers laid side by side demonstrating multiple areas of segmental demyelination typical of active CIDP.

or macrophage response. The efficacy of plasma exchange in AIDP and CIDP is generally attributed to removal of antibodies, but other mechanisms may be involved.

30μ

Multifocal CIDP An inflammatory demyelinating multiple mononeuropathy has been described by Lewis and Summer101 (see Fig. 32–89). Dyck and Dyck have also identified mononeuropathies (often sciatic) that are due to inflammatory demyelinating lesions (see Fig. 32–88).54

Chronic Immune Sensory Polyradiculopathy Sinnreich and co-workers142 have described a sensory syndrome with ataxia with raised cerebrospinal fluid protein, normal sensory action potentials and abnormalities of somatosensory nerve potential, abnormality of dorsal spinal root, and MRI enlargement and enhancement. Figures 32–92 and 32–93 show the typical MRI and histologic features.

Multifocal Motor Neuropathy Multifocal motor neuropathy is discussed elsewhere in the text (see Chapter 102).

Leprosy The pathology of leprosy is discussed elsewhere in the text (see Chapter 91).

Sarcoidosis Sarcoidosis (see Chapter 108) (see Figs. 32–1 and 32–2), a noncaseating granuloma of unknown cause, involves cranial

6μ

FIGURE 32–91 Paraffin and epoxy transverse sections of a nerve root from a patient with a long history of relapsing/remitting chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) that initially responded to immunosuppression but later was less responsive, and he became wheelchair dependent. Magnetic resonance imaging of his lumbosacral spine showed markedly enlarged nerve roots, one of which was biopsied to exclude other causes of polyradiculopathy. Top, A paraffin section immunoreacted with CD68 (KP-1). The picture is taken at high contrast to demonstrate the frequent background onion bulbs and frequent macrophages, which are associated with them. Middle, An epoxy transverse section stained with methylene blue that shows frequent large onion bulbs surrounding fibers without myelin or with thin myelin interspersed with myelinated fibers without onion bulbs and normal myelin thickness. This “mixed pattern” of onion bulb formation is typical of acquired demyelinating neuropathies (e.g., CIDP). Bottom, An electron micrograph showing three large onion bulbs. The onion bulb in the middle has a thinly myelinated profile in its center. The onion bulb on the right is denervated (no longer has an axon) and has collagen at its center. This patient was treated with aggressive immunotherapy with a combination of intravenous immunoglobulin and plasma exchange, and he had marked improvement. See Color Plate

Pathologic Alterations of Nerves

809

FIGURE 32–92 Transverse T2-weighted (left) and sagittal postgadolinium (right) magnetic resonance images of the cauda equina from a patient with chronic immune sensory polyradiculopathy. The nerve roots are enlarged and show enhancement (arrows). One of these rootlets was later biopsied (see Fig. 32–93).

nerves, meninges, and roots and infrequently causes a polyneuropathy. In transverse sections of the nerve, the granuloma may form crescentic lesions involving the perineurium and extend into the epineurium and endoneurium. The lesions tend to be focal, affecting short lengths of certain fascicles. At these sites severe damage of certain fascicles develops by unknown mechanisms. Four possible mechanisms for fiber damage might be considered: (1) cellular proliferation causing direct compression and distortion of nerve fibers; (2) cellular proliferation leading to vascular occlusion and ischemic damage; (3) release of cellular toxins, with cytokines damaging fibers; and (4) other immunologic mechanisms. The sarcoid nodules contain epithelioid cells, giant cells, lymphocytes, plasma cells, and macrophages, and caseation does not occur. The epithelioid cells may be arranged in rows or clusters surrounded by lymphocytes, plasma cells, and macrophages. Typically they form bands of inflammation within, or just outside of, the perineurium and extend into the epineurium or endoneurium. Giant cells are not always evident. Usually there is considerable scarring associated with the inflammatory process. Lepromatous leprosy must be distinguished from sarcoidosis by a different histologic appearance and by the finding of the leprous bacilli in globoids using a petroleumprotected acid-fast stain (Fite stain) (see Chapter 91).

Amyloidosis The subject of amyloidosis is discussed in Chapter 83. Amyloid is polymerized in interstitial fluid, particularly in connective tissue sheaths. In nerve it is found in epineurium,

perineurium, and endoneurium. It may be seen as linear streaks along connective tissue sheaths, as cuffs around blood vessels, or as nodules. In sections it has a ground-glass, feltlike, acellular appearance (Figs. 32–94 through 32–96). Several stains may be used to reveal it. In sections stained with hematoxylin and eosin, it is pink; with methyl violet it stains bright pink; with alkaline Congo red stains and under a polarizing filter, it is apple green and is birefringent; and with thioflavin T, it fluoresces (see Fig. 32–94). Based on the variety of precursor proteins that can be identified immunohistochemically, at least 10 varieties of amyloidosis occur. Four varieties affect nerve: light chain (LC) causes primary amyloidosis, mutations of transthyretin (TTR) cause familial amyloid polyneuropathy, mutations of gelsolin cause Icelandic amyloidosis, and apolipoprotein A-I can also cause neuropathy. Typically, in TTR amyloidosis, monoclonal LCs are not found in the serum. Many allelic varieties of TTR amyloidosis have now been described encompassing varieties previously called PortugueseJapanese, Appalachian, German-Swiss, and other varieties. In each of the alleles, a simple amino acid substitution of the DNA that codes for TTR is responsible. Increasingly, these known mutations can be detected rapidly with molecular genetic techniques. Primary amyloidosis (LC amyloidosis) is usually associated with monoclonal proteins in the serum. Immunohistochemical characterization of amyloid in nerve now permits a laboratory diagnosis of inherited versus primary amyloidosis. The mechanism of nerve fiber injury in amyloidosis remains unsettled. Earlier authors invoked ischemia.89 The frequent occurrence of carpal tunnel syndrome suggests that entrapment plays a role in some patients. In addition,

810

Pathology of the Peripheral Nervous System

nerve fibers may be compressed by an amyloid nodule immediately adjacent to the nerve46 (see Fig. 32–94). In a patient with dominantly inherited amyloidosis, amyloid nodules had indented MFs, causing marked dilatation immediately above and below the nodule.46 At other sites, absent myelin or remyelinated segments were seen adjacent to such nodular indentations—suggestive evidence that nodules can result in demyelination. It is evident that compression is implicated in the pathologic alterations of fibers; what is less clear is how much of an impact this has on the symptomatology of amyloid neuropathy. Ischemia may play a role in nerve damage, because both epineurial and endoneurial vessels may be heavily surrounded and infiltrated with amyloid even to the point of occlusion. Central fascicular fiber loss, as found in ischemic damage, has not, however, been described. Deposition of amyloid in spinal and autonomic ganglia may cause degeneration of primary afferent neurons by unknown mechanisms. Small-diameter neurons may be preferentially affected (Figs. 32–97 and 32–98).

Tumor

10μ

2μ

FIGURE 32–93 A paraffin preparation reacted to a macrophage antigen (CD68) (top), a methylene blue–stained epoxy transverse section (middle), and an electron micrograph (bottom) from a patient with chronic immune sensory polyradiculopathy. Top, Scattered macrophages associated with onion bulbs and myelinated fibers (arrows). Middle, A normal density of myelinated fibers but a altered size distribution with fewer large myelinated fibers and more small myelinated fibers than normal. The myelin of many fibers is abnormally thin, frequent onion bulbs are present, and some fibers are demyelinated (arrowheads). Bottom, Frequent onion bulb formations associated with thinly myelinated profiles. These findings provide evidence of chronic demyelination and abortive repair, and the loss of large myelinated fibers explains the ataxia. See Color Plate

This subject is reviewed in Chapter 116. Peripheral nerves tend to be spared unless they are compressed by tumor beneath tight fascial sheaths, by infiltrated meninges, or by narrowed foramina at points of exit of cranial and spinal nerves. The lymphoproliferative and plasma proliferative tumors tend to cause radicular and retroperitoneal involvement of nerves. The factors that lead to fiber degeneration may be related to direct compression of nerves. Ischemia seems likely to be a factor. Monoclonal proteins may be produced in lymphoproliferative and plasma proliferative disorders, in amyloidosis, and in conditions unrelated to tumor (monoclonal gammopathy of unknown significance). Might these monoclonal proteins have a deleterious effect on components of Schwann cells (such as myelin) or on axons? There is evidence that monoclonal proteins and LCs are found in the nerve microenvironment, but their effect there is still speculative. The passive transfer of an abnormality of nerve conduction to mice using monoclonal protein serum is suggestive of a direct deleterious effect.12 In some cases plasmapheresis appears to have a beneficial effect on the course of neuropathy, possibly providing further evidence that these proteins are involved in the neuropathic effect. The mechanism of nerve fiber damage in such primary tumors as neurofibromatosis is not well understood, but compression probably is a factor (see Chapter 116).

PERINEURIAL PATHOLOGY The important barrier functions of perineurium have been discussed in Chapters 29 and 30.

Pathologic Alterations of Nerves

811

FIGURE 32–94 A–F, Selected teased fibers from the sural nerve of a woman with the dominantly inherited Andrade type of amyloidosis causing early bladder and bowel incontinence, postural hypotension, and a syringomyelia-like dissociated sensation loss in lumbosacral dermatomes due to a selective severe loss of unmyelinated and small afferent myelinated fibers (MFs). The position of amyloid nodules (arrows) adjacent to pathologic MF alterations is shown in selected frames. C, The nodule is shown to have markedly indented an MF axon with dilatation of the fiber above and below. E and F, Remyelinated segments are found adjacent to several nodules. G, In the sural nerve of a patient with primary amyloidosis, the marked thickening of a vessel wall is demonstrated. H, The endoneurial position of a large nodule in a paraffin section. (Modified from Dyck, P. J., and Lambert, E. H.: Dissociated sensation in amyloidosis: compound action potential, quantitative histologic and teased fiber, and electron microscopic studies of sural nerve biopsies. Arch. Neurol. 20:490, 1969. Copyright 1969, American Medical Association.)

As mentioned in the earlier sections on interstitial pathologic alterations, the perineurium may be infiltrated by inflammatory cells, as in perineuritis, and may be involved in sarcoidosis, in necrotizing angiopathy, and in tumor. In severe ischemic damage, it may become necrotic. Also in ischemic injuries it will become thickened and, when its integrity is interrupted, abortive regeneration (microfasciculation) can occur within or beyond its original boundary (see Fig. 32–79). As described later, the perineurium may contain lipid debris in severe lead neuropathy. Whether this represents disturbed metabolic events of the perineurial cells themselves or a degradative or transport function of debris across

the perineurium is not known. Because similar lipid breakdown products are found in macrophages lined up along the inside of the perineurium, it may be that macrophages unload debris to the perineurium either for degradation or for transport out of the endoneurial space. The perineurium may be thickened in various inherited neuropathies. In neurofibromatosis (Fig. 32–99), the number of perineurial leaflets may be greatly increased and they may be attenuated and disorganized. The perineurium is markedly abnormal in a curious condition of digit gigantism caused by macrodystrophia lipomatosa or macrodystrophia hamartoma. These nerve tumors are associated with overgrowth of

812

Pathology of the Peripheral Nervous System

FIGURE 32–96 The typical crisscrossing, feltlike pattern of fibril in amyloidosis.

fiber alterations. That this separation is somewhat arbitrary may be illustrated by the example of lead neuropathy. In lead neuropathy, widespread segmental demyelination without inflammation is typical. Study of the experimental model has shown that endoneurial fluid is under increased pressure and is rich in lead, which exposes Schwann cells to lead, which results in segmental demyelination. Thus changes in microvessels and endoneurial fluid are implicated in pathogenesis.

Neuronal (Axonal) Pathologic Alterations FIGURE 32–95 Photomicrographs of amyloid from sural nerve of a patient with inherited amyloidosis. A, Congophilia of amyloid. B, Transthyretin positivity using immunohistochemistry. C and D, Lack of kappa (left) and lambda (right) light chain reactivity. The transthyretin reactivity supports the clinical diagnosis of inherited amyloidosis. (Courtesy of K. Ii, M.D.)

part of the hand and fibrous and lipomatous alterations. The digital nerves are made of multiple small fascicles (microfasciculation), each surrounded by an excessive number of normal-appearing perineurial lamellae.151

PARENCHYMATOUS PATHOLOGIC ALTERATIONS The separation of the pathologic alterations of nerves into interstitial and parenchymatous is somewhat arbitrary. In parenchymatous alterations, there is no obvious circulatory, inflammatory, infiltrative, or interstitial cause for the

For most neuropathies with neuronal or axonal degenerations, the three-dimensional structural features on light and electron microscopy are known only incompletely. Generally the available information is descriptive only. The causes and mechanisms of fiber degeneration are poorly understood. For these reasons the classification provided is a descriptive one. The following characteristics are taken into account: 1. 2. 3. 4. 5. 6. 7. 8. 9.

The segmental levels of neurons affected The size or functional classes of neurons affected The level(s) within neurons affected The temporal order and profile of classes of neurons affected The pathologic alterations of the soma and axon The pathologic alterations of the Schwann cells and myelin Pathologic reactions to axon or Schwann cell alterations The primary sites of neuronal vulnerability Specific mechanisms

Selectivity by Segmental Level Different segmental levels may be affected in peripheral neuron degeneration. To illustrate, in adult-onset Tangier

Pathologic Alterations of Nerves

813

FIGURE 32–97 A, Cluster of unmyelinated fibers (UFs) from healthy human sural nerve to emphasize species, age, and fiber-to-fiber variability. Characteristically, one to two UFs lie in separate Schwann cell cytoplasmic processes. Mesaxons may be short or long. Schwann cell cytoplasmic (SCC) stacks and pseudopod-like structures are not uncommon and occur more frequently with age. B and C, The fiber variability of mitochondria (M) and endoplasmic reticulum (ER) size and the density of both microtubules (MT, NT) and neurofilaments (NF/F) are illustrated. (From Dyck, P. J., and Lambert, E. H.: Dissociated sensation in amyloidosis: compound action potential, quantitative histologic and teased fiber, and electron microscopic studies of sural nerve biopsies. Arch. Neurol. 20:490, 1969, with permission. Copyright 1969, American Medical Association.)

disease, a genetic disorder with a deficiency of high-density lipoprotein apolipoprotein and tissue deposition of cholesteryl esters, cranial, cervical, and brachial neurons are severely affected at a time when lumbosacral neurons are only mildly or not affected.36,72,92,98 This distribution, however, is not found in the childhood form of the disorder even though it involves an identical molecular abnormality. The reasons for this are not known. For many parenchymatous peripheral neuropathies, the brunt falls on lumbosacral neurons, sparing those of higher segmental levels. In still others, most segmental levels are affected.

Selectivity of Functional Classes of Neurons Different functional classes of neurons may be involved in neuromuscular diseases. This selectivity is sometimes quite striking in inherited neuropathy and may be the basis of classification. Selective corticospinal or corticobulbar involvement is typical of spastic paraplegia, motor neuron involvement is characteristic of progressive muscular atrophy (hereditary motor neuropathy); sensory and autonomic neurons are involved in spinocerebellar degeneration and hereditary sensory and autonomic neuropathies (HSANs); and motor, sensory, and autonomic neurons are involved in HMSNs. Increasingly,

814

Pathology of the Peripheral Nervous System

FIGURE 32–98 Electron micrograph of a denervated myelinated fiber cluster from the sural nerve. No unmyelinated fibers are seen; only Schwann cell cytoplasm (SCC) and collagen fibrils (CF) of the epineurium are present. This figure should be compared with Figure 32–97, an unmyelinated fiber cluster from healthy sural nerve. (From Dyck, P. J., and Lambert, E. H.: Dissociated sensation in amyloidosis: compound action potential, quantitative histologic and teased fiber, and electron microscopic studies of sural nerve biopsies. Arch. Neurol. 20:490, 1969, with permission. Copyright 1969, American Medical Association.)

each of these major groups of neural disorders is divided into defined molecular genetic disorders. Thus there are now at least 10 unique disorders that might be classified as spinocerebellar degeneration. Similarly, HMSN (Charcot-Marie-Tooth disease) is known to be associated with perhaps 10 gene mutations. Selectivity by Levels within Neurons (Axons) Different levels within neurons may be selectively vulnerable. Krucke,97 in a historic account of the neuropathology of peripheral nerve disorders, emphasized the importance of selective vulnerability of different neuronal levels for pathologic change. Sometimes the major brunt of the pathologic alteration is on the soma (neuronopathies), in the root (radiculopathies), or in different proximal-to-distal levels of the axon. The postulated pharmacologic sites of action of toxins within neurons are discussed in Chapters 18 and 113. In some disorders the sites of metabolic action and initial structural change are localized to the perikaryon. Examples of this are certain viral infections, lysosomal storage diseases, and toxins. There may be examples of selective involvement of the centrally directed axons of peripheral sensory neurons. In other intoxications, different levels (from proximal to distal) of the peripheral axon may be affected.20–23,68,69,138 In murine dystrophy a focal lesion

involving a developmental abnormality of myelination appears to be confined to the roots.17 In other disorders the distal aspects of central and peripheral axons undergo pathologic changes.125,126,145 In some diseases abnormal metabolic products may accumulate in the soma (e.g., glycosphingolipid in Fabry’s disease; see Fig. 32–56); in other diseases, they may accumulate in the axon (e.g., glycogen in myxedema neuropathy47 and experimental diabetes mellitus124); and in still other diseases, they may accumulate in Schwann cell cytoplasm (e.g., metachromatic granules in metachromatic leukodystrophy [MLD]; Fig. 32–100). Neurofilaments with mitochondria dense bodies and other vesicular profiles may accumulate in spherical swellings (spheroids) in axons. This is a prominent feature in proximal axons of lesions caused by IDPN22–25,68,138; in distal aspects of both central and peripheral axons; in lesions caused by acrylamide125; in neuropathies produced by such hexacarbons as n-hexane, methyl-n-butyl ketone, and 2,5-hexanedione145,146; and in distal nerve of some patients with inherited neuropathy.75 Temporal Selectivity of Neurons (Axons) The temporal sequence of involvement of neurons may vary in different disorders. In Friedreich’s ataxia only a minority of large primary afferent neurons appear to be

Pathologic Alterations of Nerves

FIGURE 32–99 Semithin transverse sections stained with methylene blue at various locations of cutaneous neurofibroma.

FIGURE 32–100 Transverse semithin sections of sural nerve of a child with metachromatic leukodystrophy (MLD). The arrows indicate regions in which there is a collection of MLD granules, which stain metachromatically brown with acid cresyl violet. (From Dyck, P. J., Gutrecht, J. A., Bastron, J. A., et al.: Histologic and teased fiber measurements of sural nerve in disorders of lower motor and primary sensory neurons. Mayo Clin. Proc. 43:81, 1968, with permission.)

815

816

Pathology of the Peripheral Nervous System

undergoing degeneration at any one time. Large-diameter axons undergo axonal atrophy, myelin remodeling, and eventually degeneration. In other disorders, such as HMSN type I (a dominantly inherited hypertrophic neuropathy), most of the axons appear to be abnormal, and this abnormality is now known to be secondary to a Schwann cell abnormality. Anterograde and Retrograde Effects The morphologic differences that have been illustrated in the preceding paragraphs imply an interplay of different pathophysiologic mechanisms. It is clear that the neurons affected, the site of vulnerability within neurons, and the tissue reaction for most human neuropathies are poorly understood. The problem is made more difficult by the fact that derangement of the function of the soma may lead to dysfunction and degeneration of central and peripheral axons, and vice versa. The permanent axotomy model appears to show that decreased metabolic activity of the soma can lead to axonal atrophy, myelin remodeling, and axonal degeneration most severe distally and less severe centripetally. It is not known whether perikaryal metabolic failure in disease may selectively cause axonal atrophy, myelin remodeling, and degeneration of distal peripheral axons, of central axons, or of both central and peripheral axons, although the permanent axotomy model implies that this may be the case. It is also difficult to establish sites of vulnerability because axonal damage causes retrograde effects in the soma, which then have to be distinguished from possible primary effects.

Classification of Neuronal (Axonal) Pathologic Alterations The diversity of pathologic changes, the overlap of primary and secondary events and effects, the incompleteness of knowledge of the three-dimensional pathologic and morphometric changes with disease, and the lack of knowledge regarding the underlying molecular mechanisms make classification difficult. The tentative organization provided here emphasizes descriptive differences: 1. Neuronal agenesis or maldevelopment 2. Acute neuronal degeneration 3. Chronic neuronal degeneration ⫾ axonal spheroids ⫾ axonal atrophy, myelin remodeling, and axonal degeneration 4. Wallerian degeneration 5. Acute axonal degeneration 6. Axonal spheroids ⫾ axonal atrophy ⫾ myelin remodeling ⫾ axonal degeneration 7. Axonal atrophy ⫾ myelin remodeling ⫾ axonal degeneration

Agenesis and Maldevelopment Congenital absence or failure of development in utero may account for various cranial, motor, and sensory and autonomic neuropathies. A unilateral third cranial nerve involvement is reported. Because some patients with this disorder have a jaw-winking syndrome (the Marcus Gunn phenomenon, in which the ptotic eyelid elevates with jaw opening), we assume that faulty reinnervation after nerve damage, conceivably in utero, may be involved. Other cranial nerves are reported to be congenitally absent. Congenital absence of one or both third cranial nerves is called Möbius’ syndrome. Failure of development may underlie some cases of congenital progressive muscular atrophies. Failure of development may be the basis of inherited dysautonomia (HSAN type III) and other types of hereditary sensory and autonomic neuropathies (HSAN types II, IV, and V) (see Chapter 78). In HSAN types II through V, several observations suggest that a congenital failure of small neurons to develop is responsible: (1) the deficit is already present at birth, (2) the deficit appears to remain static for many years, (3) specific size classes of neurons are absent, and (4) obvious degeneration of fibers is not apparent. An abnormality of the neural crest involving neuron multiplication, differentiation, migration, or other events may be involved. Because remaining fibers do not have a normal appearance, one needs to consider an ongoing dystrophic process. Congenital maldevelopment may be involved in such disorders as inherited neurofibromatosis (see Fig. 32–99) and multiple endocrine neoplasia type 2B (Fig. 32–101). Faulty development may be involved in redundant lumbosacral roots, gigantism of digits and other structures, and certain neural tumors. Acute Neuronal Degeneration Primary pathologic derangement of the soma may lead to neuronal degeneration of lower motor, primary sensory, and autonomic neurons. In certain viral and lysosomal storage diseases, the soma appears to be the site of primary pathologic abnormality. The axon undergoes acute axonal degeneration having the morphologic features of wallerian degeneration. The two prototypes of neuronal degeneration resulting from virus infection of the soma and axonal degeneration are poliomyelitis (in motor neurons) and herpes zoster (in sensory neurons). The poliomyelitis virus probably enters the body through oral intake. It gains entrance into the neuron through the action of surface receptors. Why the disorder tends to be asymmetrical and remains limited to certain segmental levels is poorly understood.132,133 Postmortem examination in acute poliomyelitis reveals congestion and

Pathologic Alterations of Nerves

817

FIGURE 32–101 Postmortem findings in a typical case of multiple endocrine neoplasia, type 2B. A, Drawing of the cervical cord showing aberrant hyperplastic neural tissue (containing Schwann cells and myelinated axons) posterior to the cord and sending pegs into posterior columns. B, Photomicrograph of posterior column (bottom) and of aberrant neural tissue (top). C, A semithin section of the plaque showing its Schwann cell myelinated axon content. (From Dyck, P. J., Carney, J. A., Sizemore, G. W., et al.: Multiple endocrine neoplasia, type 2: phenotype recognition, neurological features and their pathological basis. Ann. Neurol. 6:302, 1979, with permission.)

infiltration of the meninges, ventral gray horn, and ventral aspect of the posterior horn with lymphocytes, polymorphonuclear cells, and macrophages. In chronic lesions ventral gray horns and ventral roots may be atrophied because of loss of somas. Because the poliomyelitis virus does not produce exotoxins or endotoxins, its cytolytic effect must be through other mechanisms. Replication of virus to the point where virions exceed the neuron’s ability to meet metabolic requirements is said to be unlikely. Therefore, specific synthesized proteins, liberation of autolysosomes, and direct toxicity of viral proteins have

been postulated. Whatever the mechanism, the effect is to cause destruction of the cell body with acute axonal degeneration. These matters are discussed in more detail elsewhere.93 Herpes zoster injury of sensory neurons also appears to be due to primary damage of the soma. The virus for varicella and herpes zoster is the same. Whereas varicella occurs in children, herpes zoster generally occurs in the old. In exposed susceptible children, a viremia leads to the typical skin lesions of varicella. The virus replicates in skin and is then transported centripetally to sensory cranial or

818

Pathology of the Peripheral Nervous System

spinal ganglia, where it may lie dormant for years without cytopathologic effect. In old age, possibly because of decreasing immunity, viral replication in sensory ganglia may be activated. Other activating factors may include an associated lymphoma or other malignancy, irradiation, trauma, or entrapment. Because pain in the dermatome precedes cutaneous vesiculation, it is thought that the virus begins to replicate in the sensory ganglia before it is transported to the skin. Replication in epidermal cells is associated with vesicle development. In herpes zoster, the gasserian ganglia are most frequently affected, followed by thoracic spinal ganglia and then other segmental ganglia. In a low percentage of cases, motor neurons or a generalized encephalomyelitis, or both, may develop. The center of the ganglion in an acute case may be necrotic and hemorrhagic (possibly as a result of ischemia) and the margins may be heavily infiltrated by lymphocytes and to a lesser degree by polymorphonuclear cells and plasma cells. As in poliomyelitis, neuron cell bodies undergo cytolysis with acute axonal degeneration. Certain lysosomal storage diseases appear to cause acute neuronal degeneration by interference with metabolic events at the level of the soma. Such lysosomal storage diseases as Fabry’s disease are of special interest because the glycosphingolipid is preferentially stored in the soma of spinal ganglion neurons, implying that this is the primary site of neuronal damage. In other storage diseases, other cells may be preferentially affected, for example, the Schwann cell in MLD. The lysosome, a cytoplasmic organelle containing various hydrolases, proteases, and other degradative enzymes, plays an important role in removing insoluble cell debris during cell metabolism. In lysosomal storage disease there is generally a severe reduction of one lysosomal enzyme resulting from a structural gene mutation, so that removal mechanisms are interfered with, leading to accumulation of specific precursors that cannot be degraded. In lysosomal storage diseases affecting sphingolipids, multimembranous bodies accumulate in the cytoplasm of neuron cell bodies to the extent that the cytoplasm seems filled with the abnormal material. Fabry’s disease, resulting from a gene mutation on the X chromosome in males, presents with lancinating and other types of pain; angiokeratoma, especially of the lower abdomen and flanks; and premature vascular occlusion and renal disease. That the pain might be due to a peripheral neuropathy was suggested by the finding of pathologic abnormalities in biopsy specimens of sural nerve.14,93 Deposition of lipid granules in spinal ganglion perikarya in this condition was described by Steward and Hitchock148 (see Fig. 32–56). Considering peripheral neurons, a pathologic and morphometric study of two cases at death confirmed that the brunt of the process fell on small-diameter spinal ganglia neurons; this study also provided insight into the sites of primary storage within neurons.118 The main locus of pathologic abnormality was

the soma, which was lipid laden to the point that the endoplasmic reticulum appeared to be decreased in amount. The typical lysosomal deposits also occurred in perineurium and endothelial cells but not in axons or Schwann cells. Ultrastructural abnormalities of axons—accumulation of NFs, dense bodies, and mitochondria—were occasionally seen. Considering the observed three-dimensional alteration of primary afferent neurons, several conclusions may be drawn: (1) considerable deposition of glycolipid occurs in the soma of primary sensory neurons; (2) there is selective vulnerability of neurons whose axons are small; (3) the entire neuron degenerates; (4) because glycolipid deposition is abundant in the soma and not in the axon and because axon death and cell body death occur at approximately the same time, it is inferred that the primary site of neuronal damage is the perikaryon; and (5) whether axonal atrophy with myelin remodeling precedes fiber degeneration is not known. Chronic Neuronal Degeneration ⫾ Axonal Spheroids ⫾ Axonal Atrophy, Myelin Remodeling, and Axonal Degeneration Descriptively, this pathologic process is characterized by (1) loss of neuronal somas, (2) presence of axonal spheroids, (3) axonal atrophy, (4) myelin remodeling, and (5) axonal degeneration. In some disorders, the proximally directed axon of the primary sensory neuron may be most affected; in others, the distally directed axon is most affected; and in still others, both centrally and distally directed axons are affected. An adequate three-dimensional description of the pathologic alterations, especially by electron microscopy, is not available for any neuropathy. Whether the primary abnormality is in the soma, in proximal axons, or in both soma and axon is not known. The pathologic alteration in motor neuron disease (amyotrophic lateral sclerosis) is a prototype of this type of neuronal degeneration. The nature of the neuronal degeneration in motor neuron disease is still poorly understood, possibly because spinal cord tissue from patients with early disease seldom becomes available for study, fixation of tissue (both disease and control) obtained at postmortem examination is generally poor, and it is not possible to identify motor axons in mixed nerves. The neuronal populations affected in motor neuron disease have been studied. Pyramidal neurons of the cortex, motor nuclei of the brainstem, and motor neuron columns of the ventral gray horns of the spinal cord are especially vulnerable. Helpful clinical information for differential diagnosis is the fact that the involvement tends to be asymmetrical. The suggestion has been made that the process begins to appear at one segmental level and then spreads outward from it. Although motor neurons are selectively vulnerable, there is evidence from detection threshold

Pathologic Alterations of Nerves

measurement of sensation, morphometric and teased fiber studies of sural nerve, and morphometric study of spinal ganglia that large primary afferent neurons are also affected but to a lesser degree. How does the neuron degenerate? Using diameter histograms of neuron cell bodies and of ventral root myelinated axons, a selective decrease of the large-diameter classes can be demonstrated. The three-dimensional morphologic alterations during degeneration are only incompletely known. Does the neuron degenerate acutely? Does such degeneration begin with distal axonal degeneration or with axonal atrophy? Axonal spheroids, containing NFs and other organelles, are found in proximal axons. Are these spheroids early primary events or can they in some way be related to neuronal failure? For sural nerve the predominant pathologic abnormality is acute axonal degeneration with degeneration of the fiber in linear rows of myelin ovoids and balls. The reported morphologic alterations of motor neuron cell bodies are atrophy, chromatolysis, and eosinophilic cytoplasmic inclusion bodies approximately 1 to 2 ␮m in diameter.78 The possible significance of the argentophilic spheroids19,164 in proximal axons is discussed in Chapters 53 and 114. Wallerian Degeneration The experimental features of wallerian degeneration were described in an earlier section of this chapter. In humans, acute transection, compression, laceration, and stretch of nerve may cause wallerian degeneration. Ischemia or compression by accumulated material in nerve may also cause wallerian degeneration. Acute Axonal Degeneration In many neuropathies distal fibers undergo the changes typical of wallerian degeneration; examples are arsenic and thallium intoxication, diabetes, uremia, alcohol, malnutrition, and vitamin deficiency. Because the mechanism is not understood and acute transection of fibers by a disease process has not been demonstrated, it may be preferable to refer to this process as acute axonal degeneration. Axonal Spheroids ⫾ Distal Axonal Atrophy ⫾ Myelin Remodeling ⫾ Axonal Degeneration IDPN neuropathy is the model best exemplifying this reaction. With IDPN intoxication, axonal spheroids form, causing focal enlargements of axons. The focal enlargements contain closely packed NFs. Altered transport of NF components is thought to be involved in the local accumulation. Sometimes an enlarged spheroid is surrounded by a thinned layer of myelin or the axon becomes demyelinated. The mechanism by which myelin becomes thin is not fully understood. Slippage of myelin lamellae on one another seems unlikely.

819

Frank cleavage and retraction of some of the myelin is another possibility. A third possibility is degeneration and incomplete regeneration of myelin. Distal to the spheroids the axon is atrophied and demyelinated and remyelinated. If the process is severe, fiber degeneration may occur. This type of pathologic alteration may be found in certain intoxications, such as IDPN, n-hexane, and acrylamide; inherited neuropathies (giant axonal neuropathy, a variant of HMSN type II, infantile neuroaxonal dystrophy, and other neuropathies); and possibly acquired neuropathy. Axonal Atrophy ⫾ Myelin Remodeling ⫾ Axonal Degeneration This fiber alteration is like that just described, but spheroids are not encountered. Whether this is due to incomplete studies or true absence of spheroids has not yet been determined. Uremic neuropathy (see Chapter 87) and Friedreich’s ataxia appear to have the fiber changes described here.

Schwann Cell Pathologic Alterations Developmental Abnormality Failure to develop myelin is found in muscular dystrophy of mouse. Axons may remain amyelinated. There appears to be a failure of Schwann cell segregation, ensheathment, and myelination of axons. Basal lamina formation is thought to be abnormal.18 Schwann cells of Trembler mice (in which there is hypomyelination of axons) express the hypomyelinated state when transplanted into nerves of immunologically suppressed normal mice.1 Normal Schwann cells transplanted into Trembler mice express the normal myelinated state. This provides clear evidence that the major abnormality of myelination is related to the Schwann cell. Developmental Abnormality in HMSN III The nerves of patients with Déjérine-Sottas disease (HMSN type III; see Chapter 69) have many of the histologic features seen in Trembler mice. The MFs have regions with excessively thin myelin and other regions without myelin. The findings suggest a hypomyelinated state possibly caused by a metabolic abnormality of Schwann cells. Xenograft studies have been difficult to interpret, but there may be evidence favoring faulty myelination. Storage Diseases In certain lysosomal storage diseases abnormal product, not degraded because of an enzyme deficiency, accumulates in the Schwann cell. Two examples are MLD (Fig. 32–102) and Krabbe’s disease (see Chapter 79). In MLD, metachromatic granules have a typical ladder-and-rung appearance in

820

Pathology of the Peripheral Nervous System

FIGURE 32–102 Typical ultrastructural features of MLD granules. A, Granules at polar ends of Schwann cells. B, MLD granules and ovoids adjacent to compact myelin. C, High-power view of ladder-and-rung periodicity typical of this organelle. D, Low-power view of MLD granules. E, Segmentally amyelinated or demyelinated region of myelinated fiber of sural nerve. (From Dyck, P. J.: Ultrastructural alterations in myelinated fibers. In Desmedt, J. E. [ed.]: New Developments in Electromyography and Clinical Neurophysiology. Basel, S. Karger, p. 192, 1973, with permission.)

electron micrographs (see Fig. 32–102). In MLD, myelin appears to be thinner than it should be, segmental demyelination is widespread, and there appears to be excessive myelin ovoid formation in paranodal and internodal sites.159 Toxins The role of diphtheria toxin in producing a primary demyelinative neuropathy is discussed in Chapter 95. Lysolecithin and various other monochain phospholipids are potent demyelinating agents when injected into the endoneurial space; these compounds have as yet unknown actions on Schwann cells or myelin.73,102

Immune Demyelination Antigalactocerebroside and experimental allergic neuritis serum, when injected endoneurially, produced demyelinative changes. Serum from patients with Guillain-Barré syndrome has also been reported to cause demyelination,60 but other investigators were unable to confirm the observation.45 Lead Intoxication as a Model of Primary Segmental Demyelination A series of observations suggest that lead neuropathy induces primary demyelination63: (1) irrespective of whether a low dose or a high dose of lead is given,

Pathologic Alterations of Nerves

segmental demyelination is induced; (2) ultrastructural abnormalities of axons have not been observed; (3) Schwann cells contain intranuclear inclusions, hydropic degeneration, and myelin degeneration; (4) the numbers of somas of lumbar spinal ganglia are not decreased; (5) the frequency of segmental demyelination and remyelination is not different in proximal and distal nerves; (6) demyelination is randomly distributed among old internodes; and (7) the putative mechanisms described in the next section favor involvement of Schwann cells. Mechanisms of Schwann Cell Damage. Nerve fibers are bathed by abundant endoneurial fluid whose constituents are maintained within narrow limits by blood-nerve, perineurial, and fluid gradient (proximally at the spinal ganglia and distally at target tissue) barriers. The possibility that endoneurial fluid is increased in lead neuropathy was suggested by Lampert and Schochet99 and demonstrated by Ohnishi and Dyck118 (Fig. 32–103). The latter authors found that the transverse fascicular area was greatly increased in lead intoxication, the density of MFs (MF/mm2) was decreased, but the number of MFs per nerve was normal. Later, Windebank and co-workers162 reported that the water content of the nerve was greatly increased. These observations suggested that edema represented a toxic breakdown in the blood-nerve barrier with subsequent entry of water and lead into the endoneurial microenvironment. A second hypothesis was that, with entry of fluid into the endoneurial space, hydrostatic pressure was increased

FIGURE 32–103 Epoxy embedded transverse sections of the peroneal nerve at mid-thigh level from a control animal (left) and from another after 6 months on a diet containing 4% lead carbonate (right). Note the markedly increased endoneurial area of the nerve from the leadfed animal. (⫻112.) (From Ohnishi, A., Schilling, K., Brimijohn, W. S., et al.: Lead neuropathy. 1. Morphometry, nerve conduction, and choline acetyltransferase transport: new finding of endoneurial edema associated with segmental demyelination. J. Neuropathol. Exp. Neurol. 36:499, 1977, with permission.)

821

and might be sufficient to damage fibers. The question of whether hydrostatic pressure was elevated in lead neuropathy was investigated by Low and Dyck103 using polyethylene matrix capsules implanted into the sciatic nerves of rats. The pressure did indeed begin to rise after 3 weeks of lead feeding. It did not reach values considered high enough to collapse capillaries. Windebank and co-workers162 performed studies to test the hypothesis that a toxic breakdown of the bloodnerve barrier might lead to Schwann cell intoxication by lead in the endoneurial fluid. In one series of experiments, they evaluated the lead and water content of nerve and looked at the association of these changes with rate of teased fibers showing demyelination and remyelination. Lead accumulated in the endoneurium very rapidly and reached a maximum between 20 and 35 days. With time the lead content fell. The onset of demyelination occurred at about the time when the endoneurial lead level was at its highest. Edema lagged behind maximal lead levels. Horseradish peroxidase, dextran of various molecular weights, and albumin normally excluded from the endoneurial space have been shown to enter the endoneurial fluid after induction of lead neuropathy. Windebank and Dyck161 examined the kinetics of lead entry into the endoneurium. After various time periods of feeding with 4% lead carbonate, tracer doses of 210Pb were given as single intravenous pulses. The entry of this tracer into the endoneurium was then followed by autoradiography. The tracer entered the endoneurium at a low rate in both control animals and animals that had been on a lead

822

Pathology of the Peripheral Nervous System

diet for 4 weeks. However, there was a 15-fold increase in the rate of entry of the lead tracer after 10 weeks of the lead diet. Thus there is a change in rate of bulk lead entry from the blood to the endoneurium, but this does not occur until 5 to 10 weeks after the onset of chronic lead ingestion. All of these studies produced remarkably consistent results using a wide range of tracers and detection techniques. During chronic lead intoxication, there is a change in the blood-nerve barrier between 6 and 12 weeks, and this change in barrier function parallels the accumulation of endoneurial edema and rising endoneurial fluid pressure. However, the actual accumulation of lead3,77,123,162 starts within 1 week of starting lead feeding and is maximal after about 5 weeks of lead diet; in addition, two studies have shown a fall in the endoneurial or whole-nerve lead content after 6 weeks of lead feeding. Segmental demyelination begins at the time of maximal lead accumulation at 5 weeks162 and before any changes are seen in the bloodnerve barrier. Thus it appears that the breakdown of the blood-nerve barrier is a late or epiphenomenon rather than a change of primary importance in the development of experimental lead neuropathy. It is appropriate to consider the complexities of barrier function when considering heavy metals such as lead. Previously we have considered compartments solely as single units—blood, extracellular space, endoneurium, and others. This is clearly an oversimplification. In the blood, for example, about 95% of lead is bound to red cells. The remainder, in the plasma, is bound to various protein species, and presumably a very small amount exists in some free solute form. In the same way, within the endoneurial compartment it is known that lead exists in a bound intracellular form. Lead-containing intranuclear inclusion bodies (Fig. 32–104) have been described in both capillary endothelial and Schwann cells.112,119 These dense, speculated inclusion bodies are identical to those found in other tissues during lead intoxication.110 In the kidney, these bodies have been shown to have a high lead and protein content,110,137 and it is assumed that they represent an insoluble lead-protein complex in vivo. Thus it is reasonable to postulate that intranuclear, intracytoplasmic, and extracellular bound and diffusible forms exist in the endoneurial compartment. At present we have no knowledge of which of these forms is biologically active in determining cell damage and what proportion each represents of the whole endoneurial lead content previously discussed. Cellular Mechanisms. To this point the major discussion has concentrated on the circumstances of lead entry into the endoneurium and its contact with the Schwann cell. In terms of concentration at the site of action and time course, it is reasonable to propose that

FIGURE 32–104 A, Schwann cell nucleus (unstained with lead; ⫻21,000). B, Endothelial cell nucleus (stained with lead; ⫻28,000). Both contain inclusion bodies. (From Ohnishi, A., and Dyck, P. J.: Retardation of Schwann cell division and axonal regrowth following nerve crush in experimental lead neuropathy. Ann. Neurol. 10:469, 1981, with permission.)

lead has a direct effect on Schwann cell function. There are several proposed mechanisms by which lead may exert this effect. However, direct evidence concerning the effect of lead on isolated enzyme or subcellular fractions is difficult to interpret because lead has affected virtually every enzyme system in which it was tested. Thus identifying a lead effect in vitro does not necessarily indicate a biologic effect in the whole animal. It is also true that identifying a cellular site of accumulation does not imply a cellular site of major action. Intranuclear inclusion bodies are one site at which lead is thought to accumulate within the cell nucleus, but it is suggested66,67 that this may represent a cellular protective response, sequestering the lead in an insoluble, and hence biologically inactive, complex.

Pathologic Alterations of Nerves

Our present knowledge suggests that the major site of action is either at the membrane level (myelin or mitochondrial) or at the level of mitochondrial oxidative metabolism. The spectrum of possible cellular sites of action is summarized. Lead might alter the composition of myelin either by binding directly to the membrane or by altering the synthesis of the normal lipid and protein components so that myelin structure can no longer be maintained. Preliminary results from electron microscopic autoradiographic studies suggest that 210Pb binds preferentially to myelin membranes. This binding may be reflected by the observed increase in rigidity in nuclear magnetic resonance studies of myelin membranes of chronically lead-intoxicated rats.160 A widespread membrane effect is suggested by the observation that red cell mechanical fragility is increased and osmotic fragility decreased as a result of increased membrane rigidity in lead-intoxicated animals 83,100 (Y. Wedmid, personal communication, 1993). These changes in membrane fluidity may reflect changes in the lysolecithin and cholesterol content of the lipid bilayer. This membrane effect might be implicated in lead interference with mitochondrial oxidative metabolism. Thus mitochondrial conformational changes thought to be essential for the coupling of oxidative phosphorylation might be prevented. Walton and Buckley157 found early mitochondrial changes in a cell culture system (chick embryo kidney tubule cells). Cell cultures exposed to concentrations of lead as low as 10⫺7 mol/L showed accumulation of electron-dense particles in cytoplasmic vesicles and mitochondria. Walton and Buckley confirmed by microincineration that the particles contained lead (or at least were mineral rich) and postulated that these particles mechanically interfered with normal mitochondrial membrane conformational changes. Abnormal mitochondrial conformations have been noted in renal tubular cells67 and in cultured muscle cells.74 This hypothesis is strengthened by the observation of Barltrop and colleagues7 that the radioactive isotope 210Pb was rapidly concentrated in the renal mitochondria of rat. That lead might interfere with the replication, transcription, or translation of genetic information has been suggested, although there is little direct evidence from other cell systems to support this. Ohnishi and Dyck demonstrated that Schwann cell division after crush injury is retarded in lead-intoxicated animals.118 However, this may not have a bearing on the primary process of demyelination, particularly since the number of Schwann cells is actually increased in the uncrushed nerve in lead-intoxicated animals. Intranuclear inclusion bodies have been reported as characteristic of lead poisoning for many years, but the significance of these bodies remains unclear. They appear to be separate from

823

chromatin,157 and, as discussed earlier, it has been suggested that they sequester lead in an insoluble and hence metabolically inactive protein complex. Several studies have suggested an effect on chromosomes and on passage of genetic information, but none of these has concerned the nervous system.10,62,156 It has been proposed that lead interferes with calcium-dependent metabolism by competitively binding to calcium receptor sites on enzymes and membrane proteins, thus interfering with substrate binding or inducing conformational changes in these proteins. Such interactions would explain the fact that decreased dietary calcium produces increased lead retention in the body, although the physiologic basis for this observation is not clear.9,111,113,143 It has also been suggested that calcium and lead interact at the cellular level.81 Walton and Buckley presented evidence that, in chick embryo renal cell culture, lead interacts with the sodium-dependent calcium ATPase pump at the level of the plasma membrane and with a postulated bivalent cation transport mechanism in the inner mitochondrial membrane.157 In studies of the CNS, Silbergeld and co-workers139–141 have demonstrated potential interactions of lead and calcium at both the synaptosomal and capillary endothelial mitochondrial levels. It still remains possible that lead interferes with some component outside the Schwann cells, with demyelination caused not directly by action of lead but by alterations in the endoneurial environment produced by lead. However, the possibility that a factor carried specifically in the edema fluid is responsible for the cellular changes has probably been excluded. Inherited Tendency to Pressure Palsy In this dominantly inherited disorder the peripheral nerves are unusually susceptible to pressure injury. Degrees of pressure not known to affect nerves cause focal neurologic abnormality. Uncompressed nerves may have abnormally low conduction velocities. Biopsy of nerves has shown short segments along internodes with excessively reduplicated myelin. Yoshikawa has found that a striking subclinical structure alteration characterizes unaffected myelin internodes.165 There are focal regions of uncompacted myelin adjacent to axolemma (Fig. 32–105). This may reflect an abnormality of adhesion molecules related to the susceptibility of these nerve fibers. Congenital Tomaculous Neuropathy Several patients have been described whose nerve fibers were characterized by hypomyelination and myelin thickening caused by reduplicated myelin.155,165

824

Pathology of the Peripheral Nervous System

FIGURE 32–105 Longitudinal electron micrograph of myelinated fiber of sural nerve of a patient with inherited tendency to pressure palsy. The failure of myelin compaction may suggest a developmental abnormality of myelin that accounts for the inherited liability to pressure injury. (From Yoshikawa, Y., Dyck, P. J., Poduslo, J. F., and Giannini, C.: Polyglucosan body axonal enlargement increases myelin spiral length but not lamellar number. J. Neurol. Sci. 98:107, 1990, with permission.)

REFERENCES 1. Aguayo, A. J., Attiwell, M., Trecarten, J., et al.: Abnormal myelination in transplanted Trembler mouse Schwann cells. Nature 265:73, 1977. 2. Aitken, J. T., and Thomas, P. K.: Retrograde changes in fiber size following nerve section. J. Anat. 96:121, 1962. 3. Anders, E., Dietz, D. D., Bagnell, C. R. Jr., et al.: Morphological, pharmacokinetic, and hematological studies of lead-exposed pigeons. Environ. Res. 28:344, 1982. 4. Asbury, A. K.: Renaut bodies: a forgotten endoneurial structure. J. Neuropathol. Exp. Neurol. 32:334, 1973. 5. Asbury, A. K., Aldredge, H., Hershberg, R., and Fisher, C. M.: Oculomotor palsy in diabetes mellitus: a clinico-pathological study. Brain 93:555, 1970. 6. Asbury, A. K., Arnason, B. G., and Adams, R. D.: The inflammatory lesion in idiopathic polyneuritis: its role in pathogenesis. Medicine 48:173, 1969. 7. Barltrop, D., Barrett, A. J., and Dingle, J. T.: Subcellular distribution of lead in the rat. J. Lab. Clin. Med. 77:705, 1971. 8. Barnes, J. M., and Denz, F. W.: Experimental demyelination and organophosphorus compounds. J. Pathol. Bacteriol. 65:597, 1953. 9. Barton, J. C., Conrad, M. E., Harrison, L., and Nuby, S.: Effects of calcium on the absorption and retention of lead. J. Lab. Clin. Med. 91:366, 1978.

10. Beckman, S. S.: Intranuclear inclusion bodies in the kidney and liver caused by lead poisoning. Bull. Johns Hopkins Hosp. 58:384, 1936. 11. Benstead, T. J., Dyck, P. J., and Sangalang, V.: Inner perineurial cell vulnerability in ischemia. Brain Res. 489:177, 1989. 12. Besinger, U. A., Toyka, K. V., Anzil, A. P., et al.: Myeloma neuropathy: passive transfer from man to mouse. Science 213:1027, 1981. 13. Bignami, A., Dahl, D., Nguyen, B. T., and Crosby, C. J.: The fate of axonal debris in Wallerian degeneration of rat optic and sciatic nerves: electron microscopy and immunofluorescence studies with neurofilament antisera. J. Neuropathol. Exp. Neurol. 40:537, 1981. 14. Bischoff, A., Fierz, U., Regli, F., and Ulrich, J.: Peripheral neurological disorders in Fabry’s disease (angiokeratoma corporis diffusum universale): clinical and electron microscopic findings in a case [in German]. Klin. Wochenschr. 46:666, 1968. 15. Bourque, C. N., Anderson, B. A., Martin del Campo, C., and Sima, A. A.: Sensorimotor perineuritis—an autoimmune disease? Can. J. Neurol. Sci. 12:129, 1985. 16. Boyd, I. A., and Davey, M. R.: Composition of Peripheral Nerves. Edinburgh, E & S Livingston, 1968. 17. Bradley, W. G., and Jenkison, M.: Abnormalities of peripheral nerves in murine muscular dystrophy. J. Neurol. Sci. 18:227, 1973.

Pathologic Alterations of Nerves 18. Bunge, R. P., Bunge, M. B., Williams, A. K., and Wartels, L. K.: Does the dystrophic mouse nerve lesion result from an extracellular matrix abnormality? In Schotland, D. L. (ed.): Disorders of the Motor Unit. New York, Wiley, p. 23, 1982. 19. Carpenter, S.: Proximal axonal enlargement in motor neuron disease. Neurology 18:841, 1968. 20. Cavanagh, J. B.: The toxic effects of tri-ortho-cresyl phosphate on the nervous system: an experimental study in lens. J. Neurol. Neurosurg. Psychiatry 17:163, 1954. 21. Chou, S. M., and Hartmann, H. A.: Axonal lesions and waltzing syndrome after IDPN administration in rats. Acta Neuropathol. (Berl.) 3:428, 1964. 22. Chou, S. M., and Hartmann, H. A.: Electron microscopy of focal neuroaxonal lesions produced by beta-betaiminodipropionitrile (IDPN) in rats. Acta Neuropathol. (Berl.), 4:590, 1965. 23. Clark, A. W., Griffin, J. W., and Price, D. L.: The axonal pathology in chronic IDPN intoxication. J. Neuropathol. Exp. Neurol. 39:42, 1980. 24. Cragg, G. B., and Thomas, P. K.: Changes in conduction velocity and fibre size proximal to peripheral nerve lesions. J. Physiol. (Lond.) 157:315, 1961. 25. Dayan, A. D., Graveson, G. S., Robinson, P. K., and Woodhouse, M. A.: Globular neuropathy: a disorder of axons and Schwann cells. J. Neurol. Neurosurg. Psychiatry 31:552, 1968. 26. Denny-Brown, D., and Brenner, C.: Paralysis of nerve induced by direct pressure and by tourniquet. Arch. Neurol. Psychiatry 51:1, 1944. 27. Doinikow, B.: Beitrage zur histologie und histopathologie der peripheren nerven. Histol. Histopathol. Arb. 4:445, 1911. 28. Donaghy, M., Brett, E. M., Ormerod, I. E., et al.: Giant axonal neuropathy: observations on a further patient. J. Neurol. Neurosurg. Psychiatry 51:991, 1988. 29. Donaghy, M., King, R. H., Thomas, P. K., and Workman, J. M.: Abnormalities of the axonal cytoskeleton in giant axonal neuropathy. J. Neurocytol. 17:197, 1988. 30. Dreyfus, R. M., Hakim, S., and Adams, R. D.: Diabetic ophthalmoplegia: report of a case with postmortem study and comments on vascular supply of human oculomotor nerve. Arch. Neurol. Psychiatry 77:337, 1957. 31. Dyck, P. J.: Detection, characterization, and staging of polyneuropathy: assessed in diabetics. Muscle Nerve 11:21, 1988. 32. Dyck, P. J.: Experimental hypertrophic neuropathy: pathogenesis of onion-bulb formations produced by repeated tourniquet applications. Arch. Neurol. 21:73, 1969. 33. Dyck, P. J.: Is there an axonal variety of GBS? Neurology 43:1277, 1993. 34. Dyck, P. J., Conn, D. L., and Okazaki, H.: Necrotizing angiopathic neuropathy: three-dimensional morphology of fiber degeneration related to sites of occluded vessels. Mayo Clin. Proc. 47:461, 1972. 35. Dyck, P. J., Ellefson, R. D., Lais, A. C., et al.: Histologic and lipid studies of sural nerves in inherited hypertrophic neuropathy: preliminary report of a lipid abnormality in nerve and liver in Dejerine-Sottas disease. Mayo Clin. Proc. 45:286, 1970.

825

36. Dyck, P. J., Ellefson, R. D., Yao, J. K., and Herbert, P. N.: Adult-onset of Tangier disease: 1. Morphometric and pathologic studies suggesting delayed degradation of neutral lipids after fiber degeneration. J. Neuropathol. Exp. Neurol. 37:119, 1978. 37. Dyck, P. J., Giannini, C., and Lais, A.: Pathologic alterations of nerves. In Dyck, P. J., Thomas, P. K., Griffin, J. W., et al. (eds.): Peripheral Neuropathy, 3rd ed. Philadelphia, W. B. Saunders, p. 514, 1993. 38. Dyck, P. J., Johnson, W. J., Lambert, E. H., and O’Brien, P. C.: Segmental demyelination secondary to axonal degeneration in uremic neuropathy. Mayo Clin. Proc. 46:400, 1971. 39. Dyck, P. J., Karnes, J., Sparks, M., and Low, P. A.: The morphometric composition of myelinated fibers by nerve, level, and species related to nerve microenvironment and ischemia. Electroencephalogr. Clin. Neurophysiol. Suppl. 36:39, 1982. 40. Dyck, P. J., Kawamura, Y., Low, P. A., et al.: The number and sizes of reconstructed peripheral autonomic, sensory and motor neurons in a case of dysautonomia. J. Neuropathol. Exp. Neurol. 27:741, 1978. 41. Dyck, P. J., Kratz, K. M., Lehman, K. A., et al.: The Rochester Diabetic Neuropathy Study: design, criteria for types of neuropathy, selection bias, and reproducibility of neuropathic tests. Neurology 41:799, 1991. 42. Dyck, P. J., and Lais, A. C.: Electron microscopy of teased nerve fibers: method permitting examination of repeating structures of same fiber. Brain Res. 23:418, 1970. 43. Dyck, P. J., and Lais, A. C.: Evidence for segmental demyelination secondary to axonal degeneration in Friedreich’s ataxia. In Kakulas, B. A. (ed.): Clinical Studies in Myology. Amsterdam, Excerpta Medica, p. 253, 1973. 44. Dyck, P. J., Lais, A. C., Giannini, C., and Engelstad, J. K.: Structural alterations of nerve during cuff compression. Proc. Natl. Acad. Sci. U. S. A. 87:9828, 1990. 45. Dyck, P. J., Lais, A. C., Hansen, S. M., et al.: Technique assessment of demyelination from endoneurial injection. Exp. Neurol. 77:359, 1982. 46. Dyck, P. J., and Lambert, E. H.: Dissociated sensation in amyloidosis: compound action potential, quantitative histologic and teased fiber, and electron microscopic studies of sural nerve biopsies. Arch. Neurol. 20:490, 1969. 47. Dyck, P. J., and Lambert, E. H.: Polyneuropathy associated with hypothyroidism. J. Neuropathol. Exp. Neurol. 29:631, 1970. 48. Dyck, P. J., Lambert, E. H., and Nichols, P. C.: Quantitative measurement of sensation related to compound action potential and number and sizes of myelinated and unmyelinated fibers of sural nerves in health, Friedreich’s ataxia, hereditary sensory neuropathy, and tabes dorsalis. In Cobb, W. A. (ed.): Handbook of Electroencephalography and Clinical Neurophysiology. Amsterdam, Elsevier, p. 83, 1971. 49. Dyck, P. J., and Lofgren, E. P.: Nerve biopsy: choice of nerve, method, symptoms, and usefulness. Med. Clin. North Am. 52:885, 1968. 50. Dyck, P. J., Low, P. A., Sparks, M. F., et al.: Effect of serum hyperosmolality on morphometry of healthy human sural nerve. J. Neuropathol. Exp. Neurol. 39:285, 1980.

826

Pathology of the Peripheral Nervous System

51. Dyck, P. J., Nukada, H., Lais, A. C., and Karnes, J. L.: Permanent axotomy: a model of chronic neuronal degeneration preceded by axonal atrophy, myelin remodeling and degeneration. In Dyck, P. J., Thomas, P. K., Lambert, E. H. and Bunge, R. P. (eds.): Peripheral Neuropathy, 2nd ed. Philadelphia, W. B. Saunders, p. 666, 1984. 52. Dyck, P. J., and O’Brien, P. C.: Meaningful degrees of prevention or improvement of nerve conduction in controlled clinical trials of diabetic neuropathy. Diabetes Care 12:649, 1989. 53. Dyck, P. J., O’Brien, P. C., and Ohnishi, A.: Lead neuropathy: II. Random distribution of segmental demyelination among “old internodes” of myelinated fibers. J. Neuropathol. Exp. Neurol. 36:570, 1977. 54. Dyck, P. J. B., and Dyck, P. J.: Chronic inflammatory demyelinating mononeuropathy (CIDM): a focal form of CIDP? Neurology 58:A303, 2002. 55. Dyck, P. J. B., Engelstad, J., Norell, J., and Dyck, P. J.: Microvasculitis in non-diabetic lumbosacral radiculoplexus neuropathy (LSRPN): similarity to the diabetic variety (DLSRPN). J. Neuropathol. Exp. Neurol. 59:525, 2000. 56. Dyck, P. J. B., Norell, J. E., and Dyck, P. J.: Microvasculitis and ischemia in diabetic lumbosacral radiculoplexus neuropathy. Neurology 53:2113, 1999. 57. Elzholz, A.: Zur Kenntnis der Veranderungen im centralen Strumpf ladierter gemischter Nerven. Jahrb. Psychiatr. 17:323, 1898. 58. Erlanger, J., and Gasser, H. S.: Electrical Signs of Nervous Activity. Philadelphia, University of Pennsylvania Press, 1937. 59. Evans, M. J., Finean, J. B., and Woolf, A. L.: Ultrastructural studies of human cutaneous nerve with special reference to lamellated cell inclusions and vacuole-containing cells. J. Clin. Pathol. 18:188, 1965. 60. Feasby, T. E., Hahn, A. F., and Gilbert, J. J.: Passive transfer of demyelinating activity in Guillain-Barre polyneuropathy. Neurology (N. Y.) 30:363, 1980. 61. Fullerton, P.: The relation between fibre diameter and internodal length in chronic neuropathy. J. Physiol. (Lond.) 178:26, 1965. 62. Gerber, G. B., Leonard, A., and Jacquet, P.: Toxicity, mutagenicity and teratogenicity of lead. Mutat. Res. 76:115, 1980. 63. Gombault, A.: Contribution a l’etude anatomique de la nevrite parenchymateuse subaigue et chronique: nevrite sementaire peri-axile. Arch. Neurol. (Paris) 1:11, 1880. 64. Gombault, A.: Sur les lesions de la nevrite alcoolique. C. R. Soc. Biol. (Paris) 102:439, 1886. 65. Gonatas, N. K., Evangelista, I., and Martin, J.: A generalized disorder of nervous system, skeletal muscle and heart resembling Refsum’s disease and Hurler’s syndrome. Am. J. Med. 42:169, 1967. 66. Goyer, R. A.: Lead and the kidney. Curr. Top. Pathol. 55:147, 1971. 67. Goyer, R. A., and Rhyne, B. C.: Pathological effects of lead. Int. Rev. Exp. Pathol. 12:1, 1973. 68. Griffin, J. W., Hoffman, P. N., Clark, A. W., et al.: Slow axonal transport of neurofilament proteins: impairment by ␤,␤⬘-iminodipropionitrile. Science 202:633, 1978.

69. Griffin, J. W., and Price, D. L.: Proximal axonopathies induced by toxic chemicals. In Spencer, P. S., and Schaumberg, H. H. (eds.): Experimental and Clinical Neurotoxicology. Baltimore, Williams & Wilkins, p. 161, 1980. 70. Grundfest, H.: The effects of hydrostatic pressure upon excitability, the recovery, and the potential sequence of frog nerve. Cold Spring Harb. Symp. Quant. Biol. 4:179, 1936. 71. Gutmann, E., and Sanders, F. K.: Recovery of fibre numbers and diameters in the regeneration of peripheral nerves. J. Physiol. (Lond.) 101:489, 1943. 72. Haas, L. F., Austad, W. I., and Bergin, J. D.: Tangier disease. Brain 97:351, 1974. 73. Hall, S. M., and Gregson, N. A.: The in vivo and ultrastructural effects of injection of lysophosphatidylcholine into myelinated peripheral nerve fibers of the adult mouse. J. Cell Sci. 9:769, 1971. 74. Harary, I., and Berliner, J.: Electron microscopic examination of lead treated L6 skeletal muscle line cells in culture. J. Environ. Pathol. Toxicol. 4:305, 1980. 75. Haymaker, W., and Kernohan, J. W.: The Landry-GuillainBarre syndrome: clinicopathologic report of 50 fatal cases and a critique of the literature. Medicine 28:59, 1949. 76. Hess, K., Eames, R., Darveniza, P., and Gilliatt, R. W.: Acute ischemic neuropathy in the rabbit. J. Neurol. Sci. 44:19, 1979. 77. Hietanen, E., Kilpio, J., Narhi, M., et al.: Biotransformational and neurophysiological changes in rabbits exposed to lead. Arch. Environ. Contam. Toxicol. 9:337, 1980. 78. Hirano, A.: Aspects of the ultrastructure of amyotrophic lateral sclerosis. In Rowland, L. P. (ed.): Advances in Neurology: Human Motor Neuron Diseases. New York, Raven Press, p. 75, 1982. 79. Hoffer, J. A., Stein, R. B., and Gordon, T.: Differential atrophy of sensory and motor fibers following section of cat peripheral nerves. Brain Res. 178:347, 1979. 80. Hoffman, P. N., and Lasek, R. J.: Axonal transport of the cytoskeleton in regenerating motor neurons: constancy and change. Brain Res. 202:317, 1980. 81. Jacquet, P., and Gerber, G. B.: Teratogenic effects of lead in the mouse. Biomedicine 30:223, 1979. 82. Johnson, P. B., Brendel, K., and Meezan, E.: Human diabetic perineurial cell basement membrane thickening. Lab. Invest. 44:265, 1981. 83. Karai, I., Fukumoto, K., and Horiguchi, S.: Studies on osmotic fragility of red blood cells determined with a coil planet centrifuge for workers occupationally exposed to lead. Int. Arch. Occup. Environ. Health 48:273, 1981. 84. Karnes, J., Robb, R., O’Brien, P. C., et al.: Computerized image recognition for morphometry of nerve attribute of shape of sampled transverse sections of myelinated fibers which best estimates their average diameter. J. Neurol. Sci. 34:43, 1977. 85. Kawamura, Y., and Dyck, P. J.: Evidence for three populations by size in L5 spinal ganglion in man. J. Neuropathol. Exp. Neurol. 37:269, 1978. 86. Kawamura, Y., Dyck, P. J., Shimono, M., et al.: Morphometric comparison of the vulnerability of peripheral motor and sensory neurons in amyotrophic lateral sclerosis. J. Neuropathol. Exp. Neurol. 40:667, 1981.

Pathologic Alterations of Nerves 87. Kawamura, Y., O’Brien, P., Okazaki, H., and Dyck, P. J.: Lumbar motoneurons of man: II. The number and diameter distribution of large- and intermediate-diameter cytons in “motoneuron columns” of spinal cord of man. J. Neuropathol. Exp. Neurol. 36:861, 1977. 88. Kernohan, J. W., and Woltman, H. W.: Periarteritis nodosa: a clinicopathologic study with special reference to the nervous system. Arch. Neurol. 39:655, 1938. 89. Kernohan, J. W., and Woltman, H. W.: Amyloid neuritis. Arch. Neurol. Psychiatry 47:132, 1942. 90. Key, A., and Retzius, G.: Studien in der Anatomie des Nerven-systems und des Bingegewebes, Vol. II. Stockholm, Nordstedt and Soner, 1876. 91. King, R. H., Llewelyn, J. G., Thomas, P. K., et al.: Diabetic neuropathy: abnormalities of Schwann cell and perineurial basal laminae. Implications for diabetic vasculopathy. Neuropathol. Appl. Neurobiol. 15:339, 1989. 92. Kocen, R. S., Lloyd, J. K., Lascelles, P. T., et al.: Familial alpha-lipoprotein deficiency (Tangier disease) with neurological abnormalities. Lancet 1:1341, 1967. 93. Kocen, R. S., and Thomas, P. K.: Peripheral nerve involvement in Fabry’s disease. Arch. Neurol. 22:81, 1970. 94. Korthals, J. K., Gieron, M. A., and Wisniewski, H. M.: Nerve regeneration patterns after acute ischemic injury. Neurology 39:932, 1989. 95. Korthals, J. K., Korthals, M. A., and Wisniewski, H. M.: Peripheral nerve ischemia. Part 2. Accumulation of organelles. Ann. Neurol. 4:487, 1978. 96. Korthals, J. K., and Wisniewski, H. M.: Peripheral nerve ischemia. Part 1. Experimental model. J. Neurol. Sci. 24:65, 1975. 97. Krucke, W.: Erkrankungen der peripheren Nerven. In Henke, F., Labarsch, O. and Rossle, R. (eds.): Handbuch der speziellen pathologischen Anatomie und Histologie. Berlin, Springer-Verlag, p. 105, 1955. 98. Kummer, H., Laissue, J., Spiess, H., et al.: Familial alphalipoproteinemia (Tangier disease) [in German]. Schweiz. Med. Wochenschr. 98:406, 1968. 99. Lampert, P. W., and Schochet, S. S. Jr.: Demyelination and remyelination in lead neuropathy: electron microscopic studies. J. Neuropathol. Exp. Neurol. 27:527, 1968. 100. Levander, O. A., Fisher, M., Morris, V., and Ferretti, R. J.: Morphology of erythrocytes from vitamin E-deficient leadpoisoned rats. J. Nutr. 107:1828, 1978. 101. Lewis, R. A., Sumner, A. J., Brown, M. J., and Asbury, A. K.: Multifocal demyelinating neuropathy with persistent conduction block. Neurology 32:958, 1982. 102. Low, P. A., Baumann, W. J., and Parthasarathy, S.: Structure-demyelinating-activity study of lysolecithin analogs in mammalian peripheral nerve. Neurology 32:A104, 1982. 103. Low, P. A., and Dyck, P. J.: Increased endoneurial fluid pressure in experimental lead neuropathy. Nature 269:427, 1977. 104. Low, P. A., Okazaki, H., and Dyck, P. J.: Splanchnic preganglionic neurons in man. I. Morphometry of preganglionic cytons. Acta Neuropathol. (Berl.) 40:55, 1977. 105. Low, P. A., Schmelzer, J. D., and Dyck, P. J.: Results of endoneurial injection of Guillain-Barre serum in Lewis rats. Mayo Clin. Proc. 57:360, 1982.

827

106. Lubinska, L.: Region of transection between preserved and regenerating parts of myelinated nerve fibers. J. Comp. Neurol. 113:315, 1959. 107. Lubinska, L.: Patterns of wallerian degeneration of myelinated fibres in short and long peripheral stumps and in isolated segments of rat phrenic nerve: interpretation of the role of axoplasmic flow of the trophic factor. Brain Res. 233:227, 1982. 108. Luna, L. G.: Manual of Histologic Staining Methods of the Armed Forces Institute of Pathology, 3rd ed. New York, McGraw-Hill, 1968. 109. Madrid, R. E., and Wisniewski, H. M.: Axonal degeneration in demyelinating disorders. J. Neurocytol. 6:103, 1977. 110. Moore, J. F., Goyer, R. A., and Wilson, M.: Lead-induced inclusion bodies: solubility, amino acid content, and relationship to residual acidic nuclear proteins. Lab. Invest. 29:488, 1973. 111. Moore, M. R.: Diet and lead toxicity. Proc. Nutr. Soc. 38:243, 1979. 112. Myers, R. R., Powell, H. C., Shapiro, H. M., et al.: Changes in endoneurial fluid pressure, permeability, and peripheral nerve ultrastructure in experimental lead neuropathy. Ann. Neurol. 8:392, 1980. 113. Mylroie, A. A., Moore, L., Olyai, B., and Anderson, M.: Increased susceptibility to lead toxicity in rats fed semipurified diets. Environ. Res. 15:57, 1978. 114. Nukada, H., and Dyck, P. J.: Microsphere embolization of nerve capillaries and fiber degeneration. Am. J. Pathol. 115:275, 1984. 115. Nukada, H., and Dyck, P. J.: Acute ischemia causes axonal stasis, swelling, attenuation, and secondary demyelination. Ann. Neurol. 22:311, 1987. 116. Ochoa, J., Danta, G., Fowler, T. J., and Gilliatt, R. W.: Nature of the nerve lesion caused by a pneumatic tourniquet. Nature 233:265, 1971. 117. Ochoa, J., Fowler, T. J., and Gilliatt, R. W.: Anatomical changes in peripheral nerves compressed by a pneumatic tourniquet. J. Anat. 113:433, 1972. 118. Ohnishi, A., and Dyck, P. J.: Loss of small peripheral sensory neurons in Fabry disease: histologic and morphometric evaluation of cutaneous nerves, spinal ganglia, and posterior columns. Arch. Neurol. 31:120, 1974. 119. Ohta, M., Offord, K., and Dyck, P. J.: Morphometric evaluation of first sacral ganglia of man. J. Neurol. Sci. 22:73, 1974. 120. Okada, E.: Ueber Zwiebelartige Gebilde im peripherischen Nerven (Renaut’sche Korperchen) bei einem Fall von Kakke (Beriberi). Tokyo Imperial Univ. Coll. Med. 6:93, 1903. 121. Okazaki, H., and Campbell, R. J.: Nervous system. In Ludwig, J. (ed.): Current Methods of Autopsy Practice, 2nd ed. Philadelphia, W. B. Saunders, p. 95, 1979. 122. Parry, G. J., and Brown, M. J.: Selective fiber vulnerability in acute ischemic neuropathy. Ann. Neurol. 11:147, 1982. 123. Powell, H. C., Myers, R. R., and Lampert, P. W.: Changes in Schwann cells and vessels in lead neuropathy. Am. J. Pathol. 109:193, 1982. 124. Powell, H. C., Ward, H. W., Garrett, R. S., et al.: Glycogen accumulation in the nerves and kidney of chronically diabetic rats: a quantitative electron microscopic study. J. Neuropathol. Exp. Neurol. 38:114, 1979.

828

Pathology of the Peripheral Nervous System

125. Prineas, J.: The pathogenesis of dying-back polyneuropathies: I. An ultrastructural study of experimental tri-ortho-cresyl phosphate intoxication in the cat. J. Neuropathol. Exp. Neurol. 28:571, 1969. 126. Prineas, J.: The pathogenesis of dying-back polyneuropathies: II. An ultrastructural study of experimental acrylamide intoxication in the cat. J. Neuropathol. Exp. Neurol. 28:598, 1969. 127. Reich, F.: Uber den zelligen Aufbau der Nervenfaser auf Grund mikrohistichemischer Untersuchungen. J. Psychol. Neurol. 8:244, 1907. 128. Renaut, J.: Systeme hyalin de soutenement des centres nerveux et de quelques organes des sens. Arch. Physiol. Normale Pathol. 8:854–860, 1881. 129. Renaut, M. J.: Recherches sur quelques points particuliers de l’histologie des nerfs. I. La gaine lamelleuse et le systeme hyalin intravaginal. Arch. N. Psych. 8:161–190, 1881. 130. Romeis, B.: Mikroskopische Technik, Vol. 15. Munchen, Leibniz, 1948. 131. Saida, G., Saida, K., Saida, T., and Asbury, A. K.: Axonal lesions in acute experimental demyelination: sequential teased nerve fiber study. Neurology (N. Y.) 31:413, 1981. 132. Saida, T., Saida, K., Silberberg, D. H., and Brown, M. J.: Transfer of demyelination by intraneural injection of experimental allergic neuritis serum. Nature 272:639, 1978. 133. Schlaepfer, W. W., and Hasler, M. B.: The persistence and possible externalization of axonal debris during wallerian degeneration. J. Neuropathol. Exp. Neurol. 38:242, 1979. 134. Schoene, W. C., Asbury, A. K., Astrom, K. E., and Masters, R.: Hereditary sensory neuropathy: a clinical and ultrastructural study. J. Neurol. Sci. 11:463, 1970. 135. Schuchmann, J. A.: Isolated sural neuropathy: report of two cases. Arch. Phys. Med. Rehabil. 61:329, 1980. 136. Seitelberger, F.: Neuroaxonal dystrophy: its relation to aging and neurologic diseases. In Vinken, P., Bruyn, G., and Klawans, H. (eds): Extrapyramidal Disorders. Amsterdam, Elsevier, p. 391, 1986. 137. Shelton, K. R., and Egle, P. M.: The proteins of leadinduced intranuclear inclusion bodies. J. Biol. Chem. 257:11802, 1982. 138. Shimono, M., Izumi, K., and Kuroiwa, Y.: 3, 3⬘-Iminodipropionitrile induced centrifugal segmental demyelination and onion bulb formation. J. Neuropathol. Exp. Neurol. 37:375, 1978. 139. Silbergeld, E. K., Adler, H. S., and Costa, J. L.: Subcellular localization of lead in synaptosomes. Res. Commun. Chem. Pathol. Pharmacol. 17:715, 1977. 140. Silbergeld, E. K., Hruska, R. E., Miller, L. P., and Eng, N.: Effects of lead in vivo and in vitro on GABAergic neurochemistry. J. Neurochem. 34:1712, 1980. 141. Silbergeld, E. K., Wolinsky, J. S., and Goldstein, G. W.: Electron probe microanalysis of isolated brain capillaries poisoned with lead. Brain Res. 189:369, 1980. 142. Sinnreich, M., Daube, J. R., Klein, C. J., and Dyck, P. J. B.: Inflammatory sensory polyradiculopathy. Neurology 60:A159, 2003. 143. Sobel, A. E., Yuska, H., and Kramer, B.: The biochemical behavior of lead. I. Influence of calcium, phosphorus and

144. 145.

146.

147.

148.

149.

150.

151.

152.

153. 154. 155.

156.

157.

158.

159.

160.

161.

vitamin D on lead in blood and bone. J. Biol. Chem. 132:239, 1940. Spencer, P. S.: The traumatic neuroma and proximal stump. Bull. Hosp. Joint Dis. 35:85, 1974. Spencer, P. S., and Schaumburg, H. H.: Central-peripheral distal axonopathy—the pathology of dying-back polyneuropathies. In Zimmerman, H. M. (ed.): Progress in Neuropathy. New York, Grune & Stratton, p. 252, 1976. Spencer, P. S., and Thomas, P. K.: The examination of isolated nerve fibres by light and electron microscopy, with observations on demyelination proximal to neuromas. Acta Neuropathol. (Berl.) 16:177, 1970. Spencer, P. S., and Thomas, P. K.: Ultrastructural studies of the dying-back process. II. The sequestration and removal by Schwann cells and oligodendrocytes of organelles from normal and diseased axons. J. Neurocytol. 3:763, 1974. Steward, V. W., and Hitchcock, C.: Fabry’s disease (angiokeratoma corporis diffusum): a report of five cases with pain in the extremities as the chief symptom. Pathol. Eur. 3:377, 1968. Stoll, G., Reiners, K., Schwendemann, G., et al.: Normal myelination of regenerating peripheral nerve sprouts despite circulating antibodies to galactocerebroside in rabbits. Ann. Neurol. 19:189, 1986. Stoner, G. L., Brosnan, C. F., Wisniewski, H. M., and Bloom, B. R.: Studies on demyelination by activated lymphocytes in the rabbit eye. I. Effects of a mononuclear cell infiltrate induced by products of activated lymphocytes. J. Immunol. 118:2094, 1977. Terzis, J. K., Daniel, R. K., Williams, H. B., and Spencer, P. S.: Benign fatty tumors of the peripheral nerves. Ann. Plast. Surg. 1:193, 1978. Thomas, P. K., and Slatford, J.: Lamellar bodies in the cytoplasm of Schwann cells. Proc. Anat. Soc. Great Britain Ireland 98:691, 1964. Thomas, P. K., and Young, J. Z.: Internode lengths in the nerves of fishes. J. Anat. 82:336, 1968. Tomonaga, M., and Sluga, E.: Zur Ultrastruktur der (pi)-granula. Acta Neuropathol. (Berl.) 15:56, 1970. Vallat, J. M., Gil, R., Leboutet, M. J., et al.: Congenital hypo- and hypermyelination neuropathy: two cases. Acta Neuropathol. (Berl.) 74:197, 1987. Verschaeve, L., Driesen, M., Kirsch-Volders, M., et al.: Chromosome distribution studies after inorganic lead exposure. Hum. Genet. 49:147, 1979. Walton, I., and Buckley, I. K.: The lead-poisoned cell: a fine structural study using cultured kidney cells. Exp. Mol. Pathol. 27:167, 1977. Weber, R. B., Daroff, R. B., and Mackey, E. A.: Pathology of oculomotor nerve palsy in diabetics. Neurology 20:835, 1970. Weller, R. O., and Das Gupta, T. K.: Experimental hypertrophic neuropathy: an electron microscope study. J. Neurol. Neurosurg. Psychiatry 31:34, 1968. Williams, I. R., Jefferson, D., and Gilliatt, R. W.: Acute nerve compression during limb ischaemia—an experimental study. J. Neurol. Sci. 46:199, 1980. Windebank, A. J., and Dyck, P. J.: Kinetics of 210Pb entry into the endoneurium. Brain Res. 225:67, 1981.

Pathologic Alterations of Nerves 162. Windebank, A. J., McCall, J. T., Hunder, H. G., and Dyck, P. J.: The endoneurial content of lead related to the onset and severity of segmental demyelination. J. Neuropathol. Exp. Neurol. 39:692, 1980. 163. Wisniewski, H. M., and Bloom, B. R.: Primary demyelination as a nonspecific consequence of a cell-mediated immune reaction. J. Exp. Med. 141:346, 1975.

829

164. Wohlfart, G.: Degenerative and regenerative axonal changes in the ventral horns, brain stem, and cerebral cortex in amyotrophic lateral sclerosis. Acta Univ. Lund. (Ner Ser. 2) 56:1, 1959. 165. Yoshikawa, H., Dyck, P. J., Poduslo, J. F., and Giannini, C.: Polyglucosan body axonal enlargement increases myelin spiral length but not lamellae number. J. Neurol. Sci. 98:107, 1990.

33 Diseases of the Neuromuscular Junction ANDREW G. ENGEL

The Safety Margin of Neuromuscular Transmission Autoimmune Disorders Myasthenia Gravis Clinical Features Pathogenesis Pathology Lambert-Eaton Myasthenic Syndrome Clinical Features Pathogenesis and Pathology

Congenital Myasthenic Syndromes Choline Acetyltransferase Deficiency Clinical Features End-Plate Studies Molecular Pathogenesis Paucity of Synaptic Vesicles and Reduced Quantal Release Congenital Myasthenic Syndrome Resembling LEMS

CMS with Reduced Quantal Release and Central Nervous System Symptoms End-Plate Acetylcholinesterase Deficiency Clinical Features Pathogenesis and Pathology Molecular Pathogenesis Postsynaptic CMSs The Slow-Channel CMSs Clinical Features End-Plate Studies Molecular Pathogenesis Fast-Channel Syndromes Clinical Features End-Plate Studies Molecular Pathogenesis Mutations Causing AChR Deficiency with or without Minor Kinetic Abnormality Clinical Features End-Plate Studies

In all diseases of the neuromuscular junction, the safety margin of neuromuscular transmission is compromised by one or more specific mechanisms. The clinical concomitant is abnormal weakness and fatigability on exertion. Table 33–1 shows a classification of currently recognized diseases of the neuromuscular junction. This chapter first reviews the factors that affect the safety margin of neuromuscular transmission and then summarizes the clinical, pathologic, and molecular aspects of the autoimmune and congenital disorders of the neuromuscular junction.

The Safety Margin of Neuromuscular Transmission The neuromuscular junction consists of presynaptic and postsynaptic regions separated by the synaptic space. Each presynaptic region is made up of a nerve terminal covered

Molecular Pathogenesis CMS Caused by Rapsyn Deficiency Clinical Features End-Plate Studies Molecular Pathogenesis Sodium Channel Myasthenia Clinical Features End-Plate Studies Molecular Pathogenesis Relation to Other Sodium Channel Disorders CMS Associated with Plectin Deficiency

Therapy Myasthenia Gravis Lambert-Eaton Myasthenic Syndrome The Congenital Myasthenic Syndromes

by a Schwann cell process. Each postsynaptic region is composed of junctional folds that overlie a specialized region of the muscle fiber, the junctional sarcoplasm. The synaptic space includes a single primary and multiple secondary synaptic clefts lined by basal lamina. Acetylcholine (ACh) is stored in the nerve terminal in synaptic vesicles.93,103 An average synaptic vesicle contains 6000 to 8000 ACh molecules.90,111 The contents of the synaptic vesicles are discharged into the synaptic space by exocytosis.93,94 A transmitter quantum is the amount of ACh released from a single synaptic vesicle. The end-plate (EP)–specific form of acetylcholinesterase (AChE) is positioned in the synaptic basal lamina14,83,132 at a density of about 2500 sites/␮m2.185 The acetylcholine receptor (AChR) is concentrated on the terminal expansions of the junctional folds, where its packing density is close to 104/␮m2.79,95,96,116,131 Voltage-gated sodium channels (Nav) of the Nav1.4 type are concentrated in the depths of the junctional folds.63,180 831

832

Pathology of the Peripheral Nervous System

Table 33–1. Classification of End-Plate Diseases Autoimmune Myasthenia gravis Lambert-Eaton myasthenic syndrome Congenital Presynaptic Synaptic basal lamina–associated Postsynaptic Toxic Botulism Drug induced

In the resting state, there is a steady but random release of transmitter quanta from the nerve terminal associated with exocytosis of individual synaptic vesicles.60,127 Focally released ACh molecules diffuse into the synaptic space, but most are not hydrolyzed initially by AChE because of saturation of AChE by the high concentration of ACh.62,183 The peak concentration of ACh reaching the postsynaptic membrane is close to 1 mM.200,201 Under normal conditions and with AChE fully active, ACh molecules in a single quantum exert their effect within a distance of 0.8 ␮m from their site of release.89 AChE limits the number of collisions of ACh with AChR and also the radius of spread of ACh. When AChE is inactive, the lateral spread of ACh is increased, so that each ACh molecule can sequentially bind to multiple AChRs and open multiple ion channels before leaving the synaptic space by diffusion.107 The depolarization and the concomitant current flow induced by a single quantum give rise to the miniature EP potential (MEPP) and miniature EP current (MEPC), respectively. The amplitude of the MEPP is a function of the number of ion channels opened by the quantum, the conductance change per channel opening, and the input resistance of the muscle fiber.108,126,127 The amplitude of the MEPC is a function of the number of ion channel openings and the mean current flow per opening. The number of channel openings is a function of the number of ACh molecules in the quantum, the number of available AChRs, and the geometry of the synaptic space, which allows rapid diffusion of most ACh molecules from their site of release to AChR. The duration of the MEPP depends on the duration of channel opening events,3,106 the functional state of AChE,107 and the cable properties of the muscle fiber surface membrane.60 The duration of the MEPC is independent of the cable properties of the muscle fiber surface membrane; otherwise, its duration is affected by the same factors that determine the duration of the MEPP.127 Once dissociated from AChR, ACh is rapidly hydrolyzed by AChE to choline and acetate. Choline is taken up by

the nerve terminal via a high-affinity, sodium-dependent, and hemicholinium-sensitive choline transporter.4,13,164 ACh is resynthesized by choline acetyltransferase (ChAT) and is then transported into the synaptic vesicles by the vesicular ACh transporter (VAChT)169,174 in exchange for protons delivered to the synaptic vesicle by a proton pump.133,174 Depolarization of the nerve terminal by a nerve impulse is followed by an influx of calcium ions into the terminal through voltage-gated calcium channels (Cav), and it is this influx that mediates transmitter release.27,105,122,135 The calcium ingress into the nerve terminal also results in facilitation, so that the probability of release by a subsequent impulse is increased. The effects of calcium are antagonized by magnesium.99,104 The details of the mechanism by which calcium triggers synaptic vesicle exocytosis are still not fully understood. However, it is now clear that the increased calcium concentration in the nerve terminal releases the synaptic vesicles from cytoskeletal constraints and affects the interaction of several proteins that together form a synaptic vesicle fusion complex. Nerve stimulation results in synaptic vesicle exocytosis adjacent to the active zones of the presynaptic membrane.94 In the freeze-fractured presynaptic membrane, the active zones are represented by double parallel rows of large (10- to 12-nm) intramembrane particles. The binding pattern of fluorescent ␻-conotoxin, a ligand for Cav at the frog motor nerve terminal,176,204 and the spatial disposition of calcium concentration microdomains visualized in the stimulated squid giant synapse with n-aequorin124 support this assumption. The number of quanta released by a nerve impulse (m) depends on the probability of release (p) and on at least one additional factor designated as n, according to the formula m ⫽ np.31 The factor n was originally defined as the number of quantal units capable of responding to the nerve impulse, but it more likely indicates the number of readily releasable quanta,42 the number of active release sites in the nerve terminal,212,218 or a combination of these variables. The amplitude of the end-plate potential (EPP) is affected by the same factors that affect the amplitude of the MEPP and by the number of quanta released by a nerve impulse. When the EPP exceeds the threshold for activating the Nav1.4 sodium channels in the depth of the synaptic folds, it triggers a muscle fiber action potential.127 A high concentration of AChRs on crests of the folds184 and of Nav1.4 in the depth of the folds63,180 ensures that excitation is propagated beyond the EP.128,216 The safety margin of neuromuscular transmission is a function of the difference between the depolarization caused by the EPP and the depolarization required to activate Nav1.4. All congenital or acquired defects of neuromuscular transmission identified to date have been traced to one or more factors

833

Diseases of the Neuromuscular Junction

Myasthenia gravis (MG) is caused by EP AChR deficiency. In most patients the disease stems from an autoimmune response against AChR. Consistent with this, 80% to 90% of MG patients have circulating antibodies against AChR.121,190 A recent report that MG patients without anti-AChR antibodies have a significant titer of antibodies against MuSK, a muscle-specific tyrosine kinase that plays a role in the aggregation of AChR at the EP, implies that MG could also arise from an autoimmune response against MuSK.97 The AChR deficiency diminishes the synaptic response to ACh and hence the amplitude of the MEPP; this reduces the amplitude of the EPP and hence the safety margin of neuromuscular transmission.

in approximately 15%. Close to 90% of the generalizations occur within 13 months after the onset. In approximately 25% of patients, MG is associated with another autoimmune disease. The clinical classification proposed by Osserman in 1958166 and by Osserman and Genkins in 1971167 divides the disease into ocular (group 1, 15% to 20%), mild generalized (group 2A, 30%), moderately severe generalized (group 2B, 20%), acute fulminating (group 3, 11%), and late severe (group 4, 9%) forms. This classification remains useful for defining stages of the disease, but the distinctions between groups are based on subjective criteria. Susceptibility to develop MG depends on AChR epitope presentation to particular major histocompatibility complex human leukocyte antigen (HLA) proteins whose expression on immune cells is genetically determined. This is reflected in the higher incidence of HLA-A1, -B8, and -DR3 determinants in young white MG patients, and of different HLA determinants in other age and racial groups. Other susceptibility genes in MG include those encoding cytotoxic lymphocyte–associated antigen, interleukin-1␤, interleukin-1 receptor antagonist, and tumor necrosis factor-␣ and -␤ (reviewed by Li et al.120).

Clinical Features

Pathogenesis

The clinical features of MG are well described in classic papers.24,71,78,167,196,207,214 The cardinal clinical finding is abnormal weakness and fatigability of some or all voluntary muscles. The weakness increases with repeated or sustained exertion and over the course of the day. Menses, viral or other infections, and emotional upsets can worsen the symptoms. The symptoms respond to anticholinesterase drugs. The external ocular muscles are affected initially in approximately 50% and eventually in approximately 90% of the cases. Voluntary muscles innervated by cranial nerves (facial, masticatory, lingual, pharyngeal, and laryngeal muscles) and cervical, pectoral girdle, and hip flexor muscles are also frequently affected; proximal limb muscles are usually more severely affected than distal ones. The tendon reflexes are brisk or normally active but may diminish if repeatedly elicited; reflexes can be absent from muscles that cannot be activated by voluntary effort. There are no objective sensory deficits. The symptoms can fluctuate from day to day, from week to week, or over longer periods of time. Spontaneous remissions can occur during the first 3 years of the disease but seldom thereafter; however, long and complete remissions are rare. In patients progressing from mild to more severe disease, the weakness tends to spread from ocular to facial to lower bulbar muscles, and then to truncal and limb muscles, but this sequence may vary and different muscles may be affected either together or in succession. The disease is initially ocular in approximately 40% of patients, but remains confined to the ocular muscles only

The basic event that breaks tolerance to self-AChR remains unknown. Three predisposing conditions have been recognized: treatment with penicillamine,61 treatment with ␣- or ␤-interferon,80,84 and bone marrow transplantation.12 As in other autoimmune diseases, the afferent limb of the immune response involves presentation of processed antigen (peptide fragments of AChR) by HLA class II positive antigen-presenting cells to specific autoreactive CD4⫹ T-helper cells, which, in turn, stimulate production by B cells and plasma cells of antibodies that recognize specific epitopes of AChR. The thymus gland is likely involved in autoimmunity to MG for the following reasons: it contains myoid cells that express AChR; the myasthenic thymus gland contains lymph nodes with germinal centers that contain AChR-specific B cells that secrete anti-AChR antibodies; and the gland is hyperplastic in approximately 70% of patients and harbors epithelial tumors in approximately 15% of patients.98 These findings suggest that a thymic abnormality could result in recognition of self-AChR components as nonself and thereby trigger the afferent limb of the immune response. The efferent limb of the autoimmune response is mediated by anti-AChR antibodies that reduce the number of EP AChRs by antibody-dependent complement-mediated lysis of the junctional folds,143,182 by accelerated internalization and destruction of AChRs (antigenic modulation),199 and by blocking the binding of ACh to AChR.35,102 The AChR deficiency decreases the amplitude of the MEPP and hence

that render the EPP subthreshold for activating Na v1.4, or to a defect in Nav1.4 that renders it unresponsive to EPPs of normal amplitude.

Autoimmune Disorders MYASTHENIA GRAVIS

834

Pathology of the Peripheral Nervous System

that of the EPP, which, in turn, reduces the safety margin of neuromuscular transmission. The basic event that breaks tolerance to self-MuSK is also unknown. Anti-MuSK antibodies inhibit agrin-induced clustering of extrajunctional AChR expressed by myotubes97; this suggests that they also decrease the density of AChR at the EP. Studies of MuSK antibody–positive MG, however, are still incomplete: the number of AChRs per EP has not been determined, no immune deposits have been demonstrated at the EP, EP fine structure has not been analyzed, and in vitro electrophysiologic studies of patient EPs have not been performed. Thus the pathogenesis of MuSK antibody–positive MG is less well understood than that of AChR antibody–positive MG.

Pathology When visualized at the light microscopic level by the cholinesterase reaction, the normal EP is pretzel shaped; in contrast, the MG EP consists of multiple small and often dotlike regions dispersed over an extended length of the muscle fiber. This appearance is not specific to MG; it occurs with any condition that causes recurrent cycles of degeneration and regeneration of the postsynaptic region,182 and also in congenital EP AChR deficiency, where it is not accompanied by degeneration of the junctional folds. Both conditions exist at some EPs in MG, and one or both conditions can exist at EPs in the congenital myasthenic syndromes (CMSs).137,158 At the ultrastructural level, morphometric analysis of individual EP regions shows that the nerve terminals and postsynaptic regions are smaller than normal. Many postsynaptic regions are simplified and have too few and too short junctional folds.57 Focal or diffuse degeneration of the junctional folds with concomitant widening of the synaptic space occurs at nearly 90% of the EPs56 (see Figs. 33–1B, 33–2, and 33–3). The widened synaptic spaces harbor degenerate residues of the junctional folds surrounded by loops of basal lamina that invested the preexisting folds. Small nerve sprouts are present near highly degenerate postsynaptic regions; the nerve sprouts presumably seek new sites of contact with the muscle fiber to initiate formation of a new postsynaptic region that, in turn, will degenerate when the AChR comes under immune attack. Ultrastructural localization of AChR with peroxidaselabeled ␣-bungarotoxin reveals marked attenuation of the density of AChR on the junctional folds, and the length of the postsynaptic membrane reacting for AChR normalized for the length of the primary synaptic cleft (the AChR index) is markedly reduced51 (Fig. 33–1). Immunoelectron microscopy demonstrates immunoglobulin G (IgG) (Fig. 33–2) and complement components C3 and C9 (Fig. 33–3) on the crests of the junctional folds, where AChR is normally concentrated, and on debris shed into the synaptic space.49,182 The presence of C9 on the

junctional folds implies that the membranolytic complement membrane attack complex (MAC) has been activated. Consistent with this, MAC has been immunolocalized at all EPs encountered in intercostal muscle specimens of 30 MG patients.45,143 Activation of the lytic phase of the complement reaction sequence results in formation of transmembrane ion channels,15,59 uncontrolled influx of extracellular ions, focal calcium excess,23 protease activation, and destruction of cytoskeletal elements within the junctional folds.

LAMBERT-EATON MYASTHENIC SYNDROME In the Lambert-Eaton myasthenic syndrome (LEMS), pathogenic autoantibodies deplete the P/Q type of Cav of the motor nerve terminal by antigenic modulation. The deficiency of the Cav restricts the ingress of calcium into the nerve terminal during activity. This reduces the probability of quantal release, which decreases the number of quanta released by nerve impulse, and hence the safety margin of neuromuscular transmission.

Clinical Features The essential clinical and electromyographic (EMG) features of the syndrome, including its association with small cell carcinoma of the lung, were described in six patients by Lambert, Eaton, and Rooke112 in 1956 and by Eaton and Lambert39 in 1957. Subsequent investigations established that LEMS occurs more frequently in men than women41 and that a higher proportion of males than females have an associated malignancy, but malignancy is uncommon under age 40. Close to 80% of the associated tumors are small cell carcinomas of the lung.165 The symptoms of LEMS can antedate those of the malignancy by many months. However, only about 3% of patients with small cell carcinoma of the lung develop LEMS.43,92,114 The characteristic symptoms of LEMS consist of weakness and increased fatigability of the truncal and limb muscles. The proximal lower limb muscles are selectively and severely affected. Seventy percent of the patients have mild or transient ocular symptoms. Severe respiratory muscle weakness requiring mechanical ventilation is uncommon. Some patients complain of myalgias that involve chiefly the thigh muscles, or have peripheral paresthesias.6,28,82,115,165,178,186 Eighty percent have autonomic nervous system abnormalities. Decreased salivation, producing dryness of the mouth, is the most common autonomic symptom; decreased lacrimation and sweating, orthostatism, impotence, and abnormal pupillary light reflexes can also occur. AChE inhibitors produce only slight or no improvement, but curare sensitivity is increased.115,178 Strength is reduced in rested muscles but increases over a few seconds during the beginning of maximal voluntary contraction. This is best

FIGURE 33–1 AChR localization with peroxidase-labeled ␣-bungarotoxin in external intercostal muscles of control subject (A) and patients with moderately severe generalized myasthenia gravis (MG) (B–D). At the normal neuromuscular junction, AChR is localized on the terminal expansions of the junctional folds. At the MG junctions the postsynaptic regions are simplified. Degeneration of the junctional folds and widening of the synaptic space are evident in D. In B and C, only segments of the simplified postsynaptic membrane react for AChR. No AChR can be detected in D. Presynaptic staining on Schwann cell (arrow in A) and on nerve terminal (arrowhead in A) is caused by diffusion artifact. (A–D, ⫻22,300.) (From Engel, A. G., Lindstrom, J. M., Lambert, E. H., and Lennon, V. A.: Ultrastructural localization of the acetylcholine receptor in myasthenia gravis and in its experimental autoimmune model. Neurology 27:307, 1977, with permission.)

836

Pathology of the Peripheral Nervous System

FIGURE 33–2 Ultrastructural localization of immunoglobulin G (IgG) at a human myasthenic neuromuscular junction. IgG deposits occur on short segments of some junctional folds and on degenerating material in the synaptic space (arrow). Asterisk indicates degenerate material not reacting for IgG. Reciprocal staining of presynaptic membrane represents diffusion artifact. (⫻39,000.) (From Engel, A. G., Lambert, E. H., and Howard, F. M.: Immune complexes [IgG and C3] at the motor end-plate in myasthenia gravis: ultrastructural and light microscopic localization and electrophysiologic correlations. Mayo Clin. Proc. 52:267, 1977, with permission.)

Diseases of the Neuromuscular Junction

837

FIGURE 33–3 Ultrastructural localization of the lytic C9 complement component at a myasthenic neuromuscular junction. The junctional folds are degenerating on the right, and here globular material that has been shed into the synaptic space reacts vividly for C9. The most degenerate folds are not covered by nerve terminal (asterisk). On the left, the junctional folds are well preserved and only sparse C9-reactive material appears between them. Presynaptic staining (arrow) is caused by diffusion artifact. (⫻20,100.) (From Sahashi, K., Engel, A. G., Lambert, E. H., and Howard, F. M. Jr.: Ultrastructural localization of the terminal and lytic ninth complement component [C9] at the motor end-plate in myasthenia gravis. J. Neuropathol. Exp. Neurol. 39:160, 1980, with permission.)

observed by asking the patient to squeeze the examiner’s fingers or to continue to exert force with a proximal muscle, such as the biceps or iliopsoas, against resistance, or by repeated testing of the same muscle at short intervals. On continued maximal exertion, strength again decreases. The tendon reflexes are hypoactive or absent in nearly all patients but may briefly return to normal immediately after exercise.115,178,179 Approximately one third of LEMS cases are nonneoplastic, and these present at any age. Non-neoplastic LEMS can be associated with other autoimmune disease(s), such as pernicious anemia, hypothyroidism, hyperthyroidism, Sjögren’s syndrome, vitiligo, celiac disease, juvenile-onset diabetes mellitus, and MG.25,81,117,146,177 Subacute cerebellar degeneration has been observed with neoplastic28,186 and less often with non-neoplastic LEMS.178 In some patients with a paraneoplastic cerebellar syndrome, LEMS may be clinically occult.28 In whites with non-neoplastic LEMS, there is also an increased association with the HLA-B8, -DR3, and -DRQ2 immune response genes.168,215 In clinical EMG studies, the amplitude of the first-evoked compound muscle action potential (CMAP) is abnormally low because low quantal release by nerve impulse blocks neuromuscular transmission at many EPs. Repetitive stimulation of rested muscle at 2 to 3 Hz further decreases the CMAP amplitude, but recovery sets in after the fourth to

eighth stimulus. Voluntary effort or nerve stimulation at frequencies higher than 10 Hz enhances the calcium concentration in the nerve terminal. This facilitates quantal release, decreases the number of blocked EPs, and enhances the amplitude of the CMAP. Single-fiber EMG205 shows increased jitter and frequent blocking of neuromuscular transmission even in minimally affected muscles. The abnormalities are worst on minimal activity and improve at higher rates of firing or at high rates of repetitive axonal stimulation. The opposite occurs in MG, in which blocking increases with increasing neural activity.189,205

Pathogenesis and Pathology In 1981, Lang and co-workers117 obtained crucial evidence that LEMS is an autoimmune disease by showing that IgG transferred from patients to mice decreased the quantal content of the EPP in recipient animals. Evidence that the Cav of the nerve terminal represented the target of the LEMS antibodies soon followed. The Cav correspond to the active zone particles noted in freeze-fractured presynaptic membranes. In 1982, Fukunaga and co-workers68 found that, in freeze-fractured presynaptic membranes of LEMS patients, the number of particles per active zone as well as the number of active zones per unit area were markedly reduced, and the active zone particles that remained were

838

Pathology of the Peripheral Nervous System

FIGURE 33–4 Freeze-fractured presynaptic membrane from a normal subject (A) and from a Lambert-Eaton syndrome patient (B). A, Several active zones appear along an arc (arrow). Each active zone comprises double parallel rows of large membrane particles that represent presynaptic voltage-gated calcium channels. A cluster of large particles occurs near the end of an active zone (arrowhead). B, The entire presynaptic membrane contains only a short active zone (arrow). A single cluster of large particles is also present (arrowhead). In comparison to A, the Lambert-Eaton presynaptic membrane shows a marked depletion of active zones and active zone particles. (A, ⫻98,000; B, ⫻61,000.) (From Fukunaga, H., Engel, A. G., Osame, M., and Lambert, E. H.: Paucity and disorganization of presynaptic membrane active zones in the Lambert-Eaton myasthenic syndrome. Muscle Nerve 5:686, 1982, with permission.)

aggregated into clusters (Fig. 33–4). Aggregation of the active zone particles was attributed to the particles being cross-linked by IgG, and the disappearance of the particles was attributed to antigenic modulation by the pathogenic IgG. Subsequent studies confirmed each of these postulates. That the membrane lesions in LEMS are indeed mediated by pathogenic IgG was established by showing that injection of LEMS IgG into mice reproduces the membrane lesions observed in LEMS patients.67 Stereometric analysis of freeze-fracture replicas of normal motor nerve terminals indicated that the active zone particles are sufficiently close to be cross-linked by IgG, and that the aggregation of the active zone particles precedes their depletion.46,70 Divalent IgG is required for antigenic modulation, and divalency of LEMS IgG was shown to be essential for aggregation and depletion of the active zone particles.142 Finally, immunoelectron microscopy studies of mice injected with LEMS IgG46,69 confirmed binding of LEMS IgG to the active zones.

Other morphologic features of the LEMS EPs are also noteworthy. When visualized at the light microscopic level by the cholinesterase reaction, the LEMS EP has a normal shape. Quantitative electron microscopy shows that the postsynaptic region of clefts and folds is significantly larger than normal, and that it is made up of structurally normal junctional folds. Thus the LEMS and MG EPs display opposite morphologic features.57

Congenital Myasthenic Syndromes CMSs arise from defects in presynaptic, synaptic basal lamina, and postsynaptic proteins. Table 33–2 shows a current classification of the CMSs and indicates the number of kinships with each type of CMS investigated at the Mayo Clinic to date.

839

Diseases of the Neuromuscular Junction

Table 33–2. Classification of the CMSs Based on Site of the Defect* Index Cases Presynaptic Defects (7%) ChAT deficiency† Paucity of synaptic vesicles and reduced quantal release Lambert-Eaton syndrome like Other presynaptic defects Synaptic Basal Lamina–associated Defects (14%) End-plate AChE deficiency† Postsynaptic Defects (79%) Kinetic abnormality of AChR with/without AChR deficiency† AChR deficiency with/without minor kinetic abnormality† Rapsyn deficiency† Kinetic defect in Nav1.4† Plectin deficiency Total (100%)

7 1 1 4

25 45 83 17 1 1 185

AChE ⫽ acetylcholinesterase; AChR ⫽ acetylcholine receptor; ChAT ⫽ choline acetyltransferase. *Classification based on cohort of CMS patients investigated at Mayo Clinic between 1988 and 2003. † Gene defects identified.

A generic diagnosis of a CMS is often possible on the basis of myasthenic symptoms since birth or early childhood, a typical pattern of the distribution of weakness with involvement of ocular and other cranial muscles, a high-arched palate, a history of similarly affected relatives, a decremental EMG response, and negative tests for AChR, MuSK, and calcium channel antibodies. Some CMSs, however, are sporadic or present in later life, a decremental EMG response may not be present in all muscles or at all times, and the weakness may be restricted in distribution and not involve cranial muscles. In some CMSs, a specific diagnosis can be made on the basis of simple histologic or EMG studies. In other CMSs, in vitro electrophysiologic, ultrastructural, and immunocytochemical investigations are needed for accurate diagnosis (Table 33–3).

CHOLINE ACETYLTRANSFERASE DEFICIENCY Clinical Features The distinguishing clinical feature of ChAT deficiency is sudden episodes of severe dyspnea and bulbar weakness leading to apnea precipitated by infections, fever, or

Table 33–3. Investigation of Congenital Myasthenic Syndromes* Clinical Data History, examination, response to AChE inhibitor EMG: conventional needle EMG, repetitive stimulation, SFEMG Serologic tests for AChR, MuSK, and calcium channel antibodies, and tests for botulism Morphologic Studies Routine histochemical studies Cytochemical and immunocytochemical localization of AChE, AChR, agrin, ␤2-laminin, utrophin, and rapsyn at the end plate Estimate of the size, shape, and configuration of AChE-reactive end plates or end-plate regions on teased muscle fibers Quantitative electron microscopy; electron cytochemistry Endplate-Specific Sites

125

I-Labeled ␣-Bungarotoxin Binding

In Vitro Electrophysiology Studies Conventional microelectrode studies: MEPP, MEPC, evoked quantal release (m, n, p) Single-channel patch-clamp recordings: channel types and kinetics Molecular Genetic Studies Mutation analysis (if candidate gene or protein identified) Linkage analysis (if no candidate gene or protein recognized) Expression studies (if mutation identified) AChE ⫽ acetylcholinesterase; AChR ⫽ acetylcholine receptor; EMG ⫽ electromyography; m ⫽ number of ACh quanta released by nerve impulse; MEPC ⫽ miniature end-plate current; MEPP ⫽ miniature end-plate potential; n ⫽ number of readily releasable ACh quanta; p ⫽ probability of quantal release; SFEMG ⫽ single-fiber EMG. *Not all studies need to be performed in all CMSs.

excitement. In some patients the disease presents at birth with hypotonia, and severe bulbar and respiratory weakness requiring ventilatory support that gradually improves, but is followed by apneic attacks and bulbar paralysis in later life.54 Other patients first experience the typical attacks during infancy or early childhood. Variable ptosis and fatigable weakness may persist between the attacks (Fig. 33–5). The clinical features of this disorder were recognized four decades ago,77 and later the disease was dubbed “familial infantile myasthenia,” but it was not differentiated from MG until the autoimmune origin of MG was established and until electrophysiologic and morphologic differences were demonstrated between MG and the congenital syndrome.140 Because all CMSs can be familial and because most CMSs present in infancy, the term familial infantile myasthenia has become a source of confusion and should be avoided.33 In clinical electrophysiology studies, a decremental response at 2-Hz stimulation and single-fiber EMG abnormalities are detected only when the tested muscles are weak. Weakness and a decremental response at 2-Hz stimulation can be induced in some but not all muscles either

840

Pathology of the Peripheral Nervous System

CMSs the EPP amplitude returns to baseline in less than 2 minutes.

Molecular Pathogenesis

FIGURE 33–5 A 5-year-old boy with heterozygous mutations in choline acetyltransferase attempting to look up. The boy had had numerous apneic episodes since birth and had mild to moderately severe myasthenic symptoms between these episodes. Note ptosis, ophthalmoparesis, compensatory head tilt, facial diplegia, tracheostomy, and percutaneous gastrostomy.

by exercise or by a conditioning train of 10-Hz stimuli for 5 to 10 minutes.47,85,88,140,175 This finding, however, is not specific for CMS with episodic apnea (CMS-EA); it can also occur in some CMS patients with EP AChE or EP AChR deficiency.

End-Plate Studies The number of AChRs per EP, estimated from the number of 125I-labeled ␣-bungarotoxin binding sites, and postsynaptic ultrastructure are normal, but morphometric analysis indicates that the synaptic vesicles are smaller than normal in rested muscle.140 In vitro microelectrode studies of intercostal muscle specimens reveal that the amplitude of the MEPP is normal in rested muscle but decreases abnormally after 10-Hz stimulation for 5 minutes. The amplitude of the EPP also decreases abnormally during 10-Hz stimulation and then recovers slowly over the next 10 to 15 minutes (Fig. 33–6A) but the quantal content of the EPP is essentially unaltered.22,140 An abnormal decline of the EPP during 10-Hz stimulation can also occur in other CMSs, but in these

That the MEPP and EPP amplitudes decline abnormally when neuronal impulse flow is increased and then recover slowly pointed to a defect in resynthesis or vesicular packaging of ACh and implicated four candidate genes: the presynaptic high-affinity choline transporter,4,164 ChAT,145 VAChT,58 and the vesicular proton pump.174 Mutation analysis in five CMS-EA patients uncovered no mutations in VACHT but revealed 10 recessive mutations in CHAT160 (see Fig. 33–6B). One mutation (523insCC) was a null mutation; three others (I305T, R420C, and E441K) markedly reduced ChAT expression in COS cells. Kinetic studies of nine bacterially expressed and purified missense mutants revealed that one (E441K) lacked catalytic activity, and eight (L210P, P211A, I305T, R420C, R482G, S498L, V506L, and R560H) had catalytic efficiencies for coenzyme A (CoA) (kcat/KmAcCoA) ranging from 1% to 39% of wild type. The overall catalytic efficiency for both CoA and choline, kcat/(KmAcCoA ⭈ Kmchol), ranged from less than 1% to 69% of wild type160 (see Fig. 33–6C). Because each patient carried two different mutations of different biochemical pathogenicity, no clear genotype-phenotype correlation emerged except that, in the most severely affected patient, both mutations (I305T and R420C) markedly reduced both the expression and the overall catalytic efficiency of ChAT. Four additional CHAT mutations (L210P188 and V194L, S694L, S694C, and R548X129) were recently detected in CMS patients with unexpected episodes of sudden apnea. None of these mutations has been functionally characterized. That patients harboring CHAT mutations have no central or autonomic nervous system symptoms was unexpected. The simplest explanation for this puzzling observation would be that the identified mutations occur in a neuromuscular junction–specific isoform of ChAT. This, however, is unlikely; although there are five alternative CHAT transcripts with at least three different promoters in human,40 the observed mutations localize to the coding region shared by all of the recognized ChAT isoforms. Tissue-specific vulnerability, however, could be affected by promoters of differing strength producing different levels of ChAT expression in different tissues. An alternative explanation is that ChAT is rate limiting for ACh synthesis at the neuromuscular synapse under conditions of increased neuronal impulse flow owing to presynaptic levels of ChAT, substrate availability, or rates of choline uptake. Although the exact reason for the selective neuromuscular junction vulnerability is not yet known, it is important to recognize that CHAT mutations can cause a potentially lethal

Diseases of the Neuromuscular Junction

841

FIGURE 33–6 End-plate (EP) choline acetyltransferase deficiency. A, 10-Hz stimulation for 5 minutes results in a rapid abnormal decline of the EP potential, which then recovers slowly over more than 10 minutes. 3,4-Diaminopyridine (3,4-DAP), which accelerates ACh release, enhances the defect, whereas a low-Ca2⫹, high-Mg2⫹ solution, which reduces ACh release, prevents the abnormal decline of the EP potential. B, Genomic structure of CHAT and identified mutations. Note that the gene encoding the vesicular ACh transporter VACHT is located in the first CHAT intron. C, Individually scaled kinetic landscapes of wild-type ChAT and of the L210P and R560H ChAT mutants. The L210P mutant shows no saturation over a practical range of acetyl-CoA (AcCoA) concentrations, indicating an extremely high Km for AcCoA. Similarly, the R560H mutant does not saturate with increasing concentrations of choline, indicating a very high Km for choline. (Reproduced from Engel, A. G, Ohno, K., and Sine, S. M.: Congenital myasthenic syndromes: progress over the past decade. Muscle Nerve 27:4, 2003, with permission.)

form of CMS without central or autonomic nervous system involvement, and to search for such mutations when clinically appropriate. It is also important to note that defects in the presynaptic high-affinity choline transporter,4,164 VAChT,58 or the vesicular proton pump174 may have similar phenotypic consequences, but no mutations of these proteins in humans have been detected to date.

PAUCITY OF SYNAPTIC VESICLES AND REDUCED QUANTAL RELEASE Only one patient suffering from this disorder has been reported to date. This patient’s clinical and electromyographic features were indistinguishable from those of patients with autoimmune MG but the onset was at birth, anti-AChR antibodies were absent, there was no EP AChR

842

Pathology of the Peripheral Nervous System

deficiency, and electron microscopy revealed no postsynaptic abnormality. A presynaptic defect was indicated by a decrease to approximately 20% of normal of the number of ACh quanta (m) released by a nerve impulse. The decrease in m was due to a decrease in the number of readily releasable quanta (n), which was associated with a decrease in the numerical density of synaptic vesicles to approximately 20% of normal in unstimulated nerve terminals.208 The patient’s symptoms were improved by pyridostigmine. In this disorder, the clinical consequences stem from the paucity of synaptic vesicles in the nerve terminal. Synaptic vesicle precursors associated with different sets of synaptic vesicle proteins are produced in the perikaryon of the anterior horn cell and are carried distally along motor axons to the nerve terminal by kinesinlike motors.17,109,123,163 Mature vesicles containing a full complement of vesicular proteins are assembled in the nerve terminal163 and are then packed with ACh. After ACh has been released by exocytosis, the vesicle membranes are recycled and then repacked with ACh.203 In the present syndrome, the reduction in synaptic vesicle density could arise from (1) a defect in the formation of synaptic vesicle precursors in the anterior horn cell, (2) a defect in the axonal transport of one or more species of precursor vesicles, (3) impaired assembly of the mature synaptic vesicles from their precursors, or (4) impaired recycling of the synaptic vesicles in the nerve terminal. That synaptic vesicle density is reduced even in unstimulated nerve terminals argues against a defect in vesicle recycling.

CONGENITAL MYASTHENIC SYNDROME RESEMBLING LEMS In one young child reported with this syndrome in 1987, the CMAP amplitude was abnormally small but facilitated severalfold on tetanic stimulation, and the symptoms were improved by guanidine.5 A second patient observed at the Mayo Clinic was a 6-month-old girl with severe bulbar and limb weakness, hypotonia, areflexia, and respirator dependency since birth. The EMG showed a low-amplitude CMAP that facilitated 500% on high-frequency stimulation and decremented 40% on low-frequency stimulation. Studies of an anconeus muscle specimen revealed no EP AChR deficiency. Electron microscopy of the EPs showed structurally intact pre- and postsynaptic regions and abundant synaptic vesicles in the nerve terminals. The MEPP amplitude was normal for muscle fiber size. The quantal content of the EPP (m) at 1-Hz stimulation was less than 10% of normal, and 40-Hz stimulation increased m by 300%. Thus the in vitro electrophysiologic findings were remarkably similar to those of LEMS.113 Consistent with this, the EMG abnormalities were improved by

3,4-diaminopyridine (3,4-DAP), an agent that increases the number of ACh quanta released by nerve impulse,125 but the patient remained weak and respirator dependent. The molecular basis of this CMS could reside in an abnormality of the presynaptic Cav or in a component of the synaptic vesicle release complex. Mutation analysis of CACNA1A, the gene encoding the pore-forming ␣1 subunit of the Cav2.1, or P/Q type, calcium channel expressed at the presynaptic membrane revealed no mutations (K. Ohno and A. G. Engel, unpublished data, 1999).

CMS WITH REDUCED QUANTAL RELEASE AND CENTRAL NERVOUS SYSTEM SYMPTOMS Maselli and co-workers130 reported three sporadic patients, two presenting in early infancy and one after the age of 5 years, with myasthenic symptoms that spared the external ocular muscles. All three had other neurologic symptoms that included truncal or limb ataxia and, in one case, horizontal nystagmus. Unlike in LEMS patients, none had an abnormally small first-evoked CMAP, and their decremental response on 2-Hz stimulation was not improved by stimulation at higher frequencies. None had EP AChR deficiency. Electron microscopy demonstrated nerve terminals of normal size containing a normal number of synaptic vesicles and displaying small doublemembrane-lined saccules filled with vesicles. In vitro microelectrode studies revealed marked decrease in the number of quanta (m) released by 1-Hz stimulation. In one patient, this was associated with a significantly decreased probability of quantal release (p); in another patient, there was a less than 50%, but still significant, decrease in the number of readily releasable quanta (n). A search for mutations in selected exons of CACNA1A was negative. One of the three patients responded well to combined treatment with pyridostigmine and 3,4-DAP, one showed only mild improvement in response to combined therapy with pyridostigmine and ephedrine, and one failed to respond adequately to pyridostigmine.

END-PLATE ACETYLCHOLINESTERASE DEFICIENCY Clinical Features This CMS is caused by the absence of AChE from the synaptic space basal lamina.48,100,156 In most patients the disease presents in the neonatal period and is highly disabling (Fig. 33–7), but in one kinship with partial AChE deficiency, it presented after the age of 6 years and became disabling only during the second decade of life.34 The cardinal clinical features are (1) a decremental EMG response

843

Diseases of the Neuromuscular Junction

FIGURE 33–7 Patient suffering from EP AChE deficiency at 4 months (left) and 8 years (right) of age. Left, Note hypotonia, ptosis, facial diplegia, open mouth, and nasogastric feeding tube. Right, Note Gowers sign and percutaneous gastrostomy.

(Fig. 33–8A); (2) a repetitive CMAP elicited from rested muscle by single-nerve stimuli (see Fig. 33–8A); (3) no effect of AChE inhibitors on the decremental response or the repetitive CMAP, and failure of AChE inhibitors to improve the clinical state; and (4) a slow pupillary light response in most adult cases. The diagnosis is confirmed by demonstrating absence of AChE from the EP (Fig. 33–9) or a pathogenic mutation in COLQ, the gene encoding the collagenic tail subunit of EP AChE.

Pathogenesis and Pathology AChE is the enzyme responsible for rapid hydrolysis of ACh released at cholinergic synapses. At the normal EP, AChE limits the number of collisions between ACh and AChR and, hence, the duration of the synaptic response.107 When the enzyme is inhibited or absent, each ACh molecule binds repeatedly to AChR until it leaves the synaptic space by diffusion. This prolongs the duration of the synaptic potentials and currents (see Fig. 33–8B). When the prolonged EPP exceeds the absolute refractory period of the muscle fiber surface membrane, it can evoke a second action potential detected by surface electrodes as a repetitive CMAP. The prolonged duration of the synaptic response results in cationic overloading of the postsynaptic region. Electron microscopy studies of the EP reveal abnormally small nerve terminals, frequently partially or totally isolated from the postsynaptic region by Schwann cell processes that extend into the synaptic cleft (Fig. 33–10). Smallness of the nerve terminals and their encasement by Schwann cells restrict the number of quanta that can be released by nerve impulse. This mitigates postsynaptic injury resulting from

overstimulation by unhydrolyzed ACh, but decreases the amplitude of the EPP. Despite this protective mechanism, cationic overloading of the postsynaptic region still causes an EP myopathy characterized by focal degeneration of the junctional folds with loss of AChR, and appearance of degenerating organelles and apoptotic nuclei in the junctional sarcoplasm.48,100 The safety margin of neuromuscular transmission is compromised by the decreased quantal content of the EPP, the EP myopathy, desensitization of AChR by prolonged exposure to ACh, and a depolarization block at physiologic rates of stimulation.

Molecular Pathogenesis The EP species of AChE is a heteromeric asymmetric enzyme composed of one, two, or three homotetramers of globular catalytic subunits (AChET) attached to a triplestranded collagenic tail (ColQ) (Fig. 33–11B). ColQ has an N-terminal proline-rich region attachment domain (PRAD), a collagenic central domain, and a C-terminal region enriched in charged residues and cysteines (Fig. 33–11A). Each ColQ strand can bind an AChET tetramer to its PRAD, giving rise to A4, A8, and A12 species of asymmetric AChE.16 Two groups of charged residues in the collagen domain (heparan sulfate proteoglycan binding domains)32 plus other residues in the C-terminal region110,153 assure that the asymmetric enzyme is inserted into the synaptic basal lamina. The C-terminal region is also required for initiating the triple-helical assembly of ColQ that proceeds from a C- to an N-terminal direction in a zipper-like manner.170 Expression of globular and asymmetric forms of AChE in muscle, or in COS cells

844

Pathology of the Peripheral Nervous System

FIGURE 33–8 End-plate AChE deficiency. A, Decremental electromyographic response and repetitive compound muscle action potential recorded from thenar muscle of patient during 2-Hz stimulation of the median nerve. At this rate of stimulation, the second response decrements more rapidly than the first and appears only once. B, Representative miniature EP currents (MEPCs) from a patient with EP AChE deficiency and a control subject. The best-fit exponential curve is superimposed on the decay phase of each current. Arrows indicate decay time constants. The MEPC is smaller and decays more slowly in the patient than in the control. (From Hutchinson, D. O., Walls, T. J., Nakano, S., et al.: Congenital endplate acetylcholinesterase deficiency. Brain 116:633, 1993, with permission.)

transfected with ACHET and COLQ complementary DNA (cDNA), is readily monitored by density gradient centrifugation of muscle or COS cell extracts (Fig. 33–11C and D). In 1998, human COLQ cDNA was cloned,34,151 the genomic structure of COLQ determined,151 and the molecular basis of EP AChE deficiency traced to recessive mutations in COLQ.34,151 Twenty-four COLQ mutations in 25 kinships have been identified to date34,101,150,151,153,191 (see Fig. 33–11B). The mutations are of three major types: (1) PRAD mutations preventing attachment of AChET to ColQ (Fig. 33–11E and I); (2) collagen domain mutations producing a short, single-stranded ColQ that binds a single AChET tetramer and is insertion incompetent (Fig. 33–11F and J); and (3) C-terminal mutations hindering the triple helical assembly of the collagen domain or producing an asymmetric species of AChE that is insertion incompetent, or both (Fig. 33–11H and K).

FIGURE 33–9 Electron cytochemical localization of AChE at a control EP (A) and at an AChE-deficient patient EP (B). At the control EP, heavy reaction product fills the synaptic space and extends into the adjacent regions. At the patient EP there is no reaction product for AChE in the synaptic space. (A, ⫻20,000; B, ⫻9000.)

POSTSYNAPTIC CMSs Most postsynaptic CMSs identified to date stem from a defect in the level of expression, kinetic properties, or aggregation of AChR. We therefore begin with a brief review of the structure and kinetic properties of the muscle species of AChR. Muscle AChR is a transmembrane macromolecule composed of five homologous subunits: two of ␣, one of ␤ and ␦, and one of ⑀ in adult AChR, or one of ␥ instead of ⑀ in fetal AChR. The genes coding for ␣, ␦, and ␥ subunits are at different loci on chromosome 2q, and those coding for ␤ and ⑀ subunits are at different loci on chromosome 17p. The subunits are highly homologous, have similar secondary structures, fold similarly, and are organized like barrel staves around a central cation channel. Each subunit has an Nterminal extracellular domain that comprises approximately

845

Diseases of the Neuromuscular Junction

FIGURE 33–10 End-plate AChE deficiency. A small nerve terminal completely encased by Schwann cell is applied against a small fraction of a postsynaptic region. The junctional folds are decorated with labyrinthine membranous networks. Mitochondria are absent, and a small focus of myofibrillar degeneration appears under the junctional folds (asterisk). (⫻25,900.)

50% of the primary sequence, four putative transmembrane domains (TMD1 through TMD4), and a small C-terminal extracellular domain. TMD2, which lines the ion channel, forms an ␣ helix interrupted by a short stretch of ␤ sheet. The TMDs are connected by an extracellular TMD2/TMD3 linker and by intracellular TMD1/TMD2 and TMD3/ TMD4 linkers. The TMD3/TMD4 linker forms a long cytoplasmic loop that serves as an attachment site for rapsyn, which links AChR to the subsynaptic cytoskeleton. The long cytoplasmic loops also bear phosphorylatable residues that may be important for desensitization, and a segment of the long cytoplasmic loop of the ⑀ subunit stabilizes the gating mechanism. Each AChR has two ACh binding pockets, one at the ␣/⑀ (or ␣/␥) and one at the ␣/␦ interface. Residues contributing to the binding pocket appear on three peptide loops on the ␣ subunit and on four peptide loops on the ⑀, ␦, and ␥ subunits. The recent x-ray analysis of the structure of the snail glial ACh binding protein confirmed that con-

served residues in all seven loops are present at the ligand binding sites.18 Activation of AChR by ACh can be described by a linear scheme that accounts for essential features of AChR activation in a variety of species197,217:

A+R

k+1 k–1

AR + A

k+2 k–2

A 2R

β α

A 2R*

where A is agonist, R is the resting state, R* is the openchannel state, ␤ the rate of channel opening, ␣ the rate of channel closure, and k⫹ and k⫺ the forward binding and dissociation rates of ACh. The dissociation constants of ACh from monoliganded and diliganded AChR, K1 and K2, obtained from the k⫺1/k⫹1 and k⫺2/k⫹2 ratios, are measures of the affinity of the monoliganded and diliganded receptor for ACh. The above scheme represents a

846

Pathology of the Peripheral Nervous System

FIGURE 33–11 A, Schematic diagram showing domains of a ColQ strand. B, Schematic diagram showing the A1 through A12 species of asymmetric AChE with 24 identified ColQ mutations. C–H, Density gradient profiles of AChE extracted from COS cells transfected with wild-type ACHET and different types of COLQ mutants. I–K, Schematic diagrams of the abnormal species of AChE formed after the transfections. In E and I, note that disruption of the PRAD domain produces a sedimentation profile identical to that obtained after transfection with ACHET alone. Thus ACHET fails to attach to ColQ and no asymmetric AChE is formed. In F and J, note that the asymmetric A4, A8, or A12 moieties are absent and there is a prominent 10.5-S mutant (M) peak, representing a G4 tetramer of the catalytic subunit linked to the truncated ColQ peptide. In G and the left diagram in K, note the presence of a small M peak but absence of peaks corresponding to triple-stranded asymmetric enzymes. In H and the right diagram in K, note that both an M peak and asymmetric AChE are present. HSPBD ⫽ heparan sulfate proteoglycan binding domain; PRAD ⫽ proline-rich attachment domain.

847

Diseases of the Neuromuscular Junction

Table 33–4. Kinetic Abnormalities of AChR Slow-Channel Syndromes

Fast-Channel Syndromes

End-plate currents Channel opening events Open states Closed states Mechanisms*

Slow decay Prolonged Stabilized Destabilized Increased affinity Increased ␤ Decreased ␣

Pathology Genetic background Response to therapy

Endplate myopathy from cationic overloading Dominant gain-of-function mutations Long-lived open channel blockade of AChR with quinidine or fluoxetine

Fast decay Brief Destabilized Stabilized Decreased affinity Decreased ␤ Increased ␣ Mode-switching kinetics No anatomic footprint Recessive loss-of-function mutations 3,4-DAP and AChE inhibitors

␣ ⫽ channel closing rate; ␤ ⫽ channel opening rate. *Different combinations of mechanisms operate in the individual slow- and fast-channel syndromes.

simplification because the unliganded and monoliganded receptors can open, the nonconducting blocked and desensitized states are not represented, and transitions can occur between all adjacent states. A more complete description of AChR activation is provided by the Monod-Wyman-Changeux description of allosteric protein function139:

R* + A

K1 *

AR* + A

θ1

θ0 R+A

K2 *

A2R* θ2 A2R

AR + A K1

affected patients, the cranial muscles tend to be spared. The weakness and fatigability can fluctuate, but not as rapidly as in autoimmune MG. The tendon reflexes are usually normal but can be reduced in severely affected limbs. The more severely affected muscles become atrophic. Progressive spinal deformities and respiratory embarrassment are common complications during the evolution of the illness.

K2

To date, two major kinetic abnormalities of AChR have emerged that give rise to the slow-channel and the fast-channel CMSs. The two syndromes are physiologic opposites. Table 33–4 summarizes the divergent properties of the two syndromes.

THE SLOW-CHANNEL CMSs Clinical Features The slow-channel syndromes are caused by dominant gain-of-function mutations. The clinical phenotypes vary. Some slow-channel CMSs present in early life and cause severe disability by the end of the first decade137; others present later in life and progress slowly, resulting in little disability even in the sixth or seventh decade.50,54,198 Most patients show selectively severe involvement of cervical and of wrist and finger extensor muscles (Fig. 33–12). Except for the more severely

FIGURE 33–12 Slow-channel syndrome patient is attempting to extend wrists and fingers as shown by examiner (with sleeve). Note atrophy of patient’s forearm muscles. (From Engel, A. G., Lambert, E. H., Mulder, D. M., et al.: A newly recognized congenital myasthenic syndrome attributed to a prolonged open time of the acetylcholine-induced ion channel. Ann Neurol 11:553, 1982, with permission.)

848

Pathology of the Peripheral Nervous System

FIGURE 33–13 Slow-channel syndrome. Repetitive and decrementing compound muscle action potential evoked from a limb muscle by a single nerve stimulus. The second and third responses are triggered by the prolonged EPP that outlasts the absolute refractory period of the muscle fiber. The second and third responses decrement faster than the first response.

End-Plate Studies As in EP AChE deficiency, single-nerve stimuli elicit one or more repetitive CMAPs (Fig. 33–13) but, unlike in EP AChE deficiency, AChE inhibitors increase the number of repetitive responses. Repetitive nerve stimulation reveals a

decremental response that is present at low stimulation frequency (Fig. 33–13) and increases progressively when stimulation frequency is increased. The repetitive CMAPs are of lower amplitude and decrement faster than the first CMAP. In vitro microelectrode studies demonstrate markedly prolonged EPPs and currents that decay biexponentially53,155,198 (Fig. 33–14C). Single-channel patch-clamp recordings reveal a dual population of AChR channels, one with normal and one with prolonged opening episodes, reflecting the presence of both wild-type and mutant receptors at the EP53,137,155 (Fig. 33–14B). In addition, the mutant AChR channels open even in the absence of ACh,137,155 resulting in a continuous cation leak into the postsynaptic region. The morphologic consequences stem from prolonged activation episodes of the AChR channel that cause cationic overloading of the postsynaptic region. Excessive accumulation of Ca2⫹ can be demonstrated at some EPs with alizarin red, which detects millimolar concentrations of Ca2⫹. The EP myopathy that develops is like that in EP AChE deficiency but is even more severe, sometimes causing massive destruction of the junctional folds (Fig. 33–15), nuclear

FIGURE 33–14 A, Schematic diagram of AChR subunits with slow-channel mutation. The mutations occur in the extracellular and transmembrane domains of the different subunits. B, Examples of single channel currents from wild-type and slow-channel (␣V249F) AChRs expressed in HEK cells. C, Miniature EP currents (MEPCs) recorded from EPs of a control subject and a patient harboring the ␣V249F slow-channel mutation. The slow-channel MEPC decays biexponentially as a result of expression of both wild-type and mutant AChRs at the EP, so one decay time constant is normal and the other is markedly prolonged.

Diseases of the Neuromuscular Junction

849

FIGURE 33–15 Slow-channel syndrome EP. Note degeneration of junctional folds and accumulation of degenerate material in the widened synaptic space. (⫻25,900.)

apoptosis (Fig. 33–16), and vacuolar degeneration near the EPs50,53,54,137 (Fig. 33–17).

Molecular Pathogenesis The abnormal kinetic properties of AChR predicted that the slow-channel syndrome stemmed from mutations in AChR subunits. Since 1995, 18 slow-channel mutations have been uncovered.29,53,74–76,91,137,156,162,192,198,209 The different mutations occur in different AChR subunits and in different functional domains of the subunits (see Fig. 33–14A). In the kinships observed to date, mutations in the channel domain had more severe phenotypic consequences than those at the ACh binding site, but there were also variations in phenotypic expressivity between and within kinships harboring the same mutation. Thus, although suggestive, phenotypic severity is not a fully reliable predictor of mutation site.

Patch-clamp studies at the EP, mutation analysis, and expression studies in human embryonic kidney (HEK) cells indicate that the ␣G153S mutation near the extracellular ACh binding site198 and the ␣N217K209 and ⑀L221F91 mutations in the N-terminal part of TMD1 act mainly by enhancing affinity for ACh. This slows dissociation of ACh from the binding site and results in repeated channel reopenings during the prolonged receptor occupancy. Another slow-channel mutation near the binding site region, ␣V156M, probably has similar effects, although its mechanistic consequences have not been investigated.29 Mutations in TMD2 (which lines the channel pore), such as ␤V266M, ⑀L269F, ⑀T264P, and ␣V249F, promote the open state by affecting channel opening and closing steps.53,137,155 Some TMD2 mutations also increase affinity for ACh; this is most marked in the case of ␣V249F,137 pronounced with ⑀L269F53 and ⑀T264P,155 and not apparent with

FIGURE 33–16 Slow-channel syndrome. Apoptotic nucleus appears near a degenerating postsynaptic region. (⫻15,500.)

FIGURE 33–17 Slow-channel syndrome. Autophagic vacuoles (V) containing degraded material near a postsynaptic region (X) abandoned by its nerve terminal. (⫻11,900.)

851

Diseases of the Neuromuscular Junction

␤V266M.53 These TMD mutations likely exert their effects through structures that couple the binding site to the channel gate. The safety margin of neuromuscular transmission is compromised by the altered EP geometry, the loss of AChR from degenerating junctional folds, and a depolarization block during physiologic activity owing to staircase summation of the markedly prolonged EPPs.

FAST-CHANNEL SYNDROMES Clinical Features The fast-channel mutations derange one or more of the following functions of AChR: affinity for ACh, efficiency of gating, and stabilization of channel kinetics. The clinical features resemble those of autoimmune MG, but symptoms are mild when the main effect is on gating efficiency,210 moderately severe when channel kinetics are unstable,138,211 and severe when affinity for ACh, or both affinity and gating efficiency, are impaired.161,195

End-Plate Studies The low-affinity fast-channel syndromes leave no anatomic footprint161,195: the structural integrity of the postsynaptic region is maintained, and the density and distribution of AChR on the junctional folds are normal. In the CMS that only affects gating efficiency210 and in the CMS with unstable channel kinetics,138 the number of AChRs per EP is also reduced. In these CMSs, multiple small EP regions are dispersed over an extended length of the fiber surface, some postsynaptic regions are simplified, and the expression of AChR on the junctional folds is patchy and attenuated. The common electrophysiologic features of the fast-channel CMSs are rapidly decaying EP currents (Fig. 33–18A), abnormally brief channel activation episodes (Fig. 33–18B), and a reduced quantal response owing to a decreased probability of channel opening (see Table 33–4).

Molecular Pathogenesis The fast-channel CMSs are caused by recessive, loss-offunction mutations. Several fast-channel mutations have

FIGURE 33–18 A, Miniature EP currents (MEPCs) recorded from EPs of a control subject and a patient harboring the ␣V285I fastchannel mutation. Arrows indicate decay time constants. B, Examples of single channel currents from wild-type and fast-channel (␣V285I) AChRs expressed in HEK cells. C, Schematic diagram of fast-channel mutations in the AChR ␣, ␤, and ␦ subunits. (From Engel, A. G., Ohno, K., and Sine, S. M.: Congenital myasthenic syndromes: progress over the past decade. Muscle Nerve 27:4, 2003, with permission.)

852

Pathology of the Peripheral Nervous System

been reported to date (see Fig. 33–18C). In each instance, the mutated allele causing the kinetic abnormality is accompanied by a null mutation in the second allele; therefore, the kinetic mutation dominates the clinical phenotype. Low-Affinity Fast-Channel Syndromes Four identified kinetic mutations fall in this group. Two different substitutions of residue 121 in the extracellular domain of the ⑀ subunit, ⑀P121L161 and ⑀P121T,194 both result in abnormally brief channel events, reduce the amplitude of the quantal response by decreasing the probability of channel openings, decrease the number of channel reopenings in bursts of openings, and decrease the affinity for ACh in the open channel state. The ⑀P121L mutation also reduces gating efficiency by reducing ␤, the channel opening rate. In contrast, the ⑀P121T mutation does not alter gating efficiency, but reduces closed-state affinity for ACh and thereby stabilizes the closed state. The third mutation in this group occurs in the ␣ subunit, in the most highly conserved domain of the AChR superfamily, the disulfide-bridged ␤-hairpin formed between cysteine 128 and cysteine 142, or the Cys-loop. The mutation consists of replacement of a conserved valine at position 132 by a leucine (␣V132L). The ␣V132L mutation markedly reduces closed-state affinity for ACh, and impairs gating efficiency by slowing ␤, the rate of channel opening, and speeding ␣, the rate of channel closing. The duration of channel opening events is only 15% of normal, and the amplitude of quantal response only 10% of normal.195 The fourth mutation in this group, ␦E59K, is in the extracellular domain of the ␦ subunit. It is of special interest because it causes multiple congenital joint contractures owing to fetal hypomotility in utero.20

the EP, ⑀1254ins18-mutated AChR shows abnormally brief activation episodes during steady-state ACh application. When expressed in HEK cells and exposed to desensitizing concentrations of ACh, openings of individual receptors occur in clusters separated by silent periods during which all receptors are desensitized. For the normal receptor, the kinetics of gating within a cluster is essentially uniform and highly efficient. By contrast, the kinetic behavior of the ⑀1254ins18-mutated AChR changes abruptly, so that a normal mode of activation is replaced by three abnormal and inefficient modes in which the receptor opens slowly and closes more rapidly than normal. In this disorder, the reduced gating efficiency and decreased AChR expression are partially offset by expression of fetal AChR harboring the ␥ instead of the ⑀ subunit (␥-AChR); this improves electrical activity at EP and likely rescues the phenotype.138 The second mutation that destabilizes channel kinetics is a nearby missense mutation in the ⑀ subunit, ⑀A411P. When this mutation is expressed in HEK cells, different clusters of channel openings differ widely in their activation kinetics, so that the spread in the distribution of the channel opening and closing rates is greatly expanded.211 Intracellular microelectrode and patch-clamp studies on EPs of patients harboring this mutation are still unavailable. That both ⑀1254ins18 and ⑀A411P occur in the amphipathic helix region of the long cytoplasmic loops of the ⑀ subunit implicates this region of the ⑀ subunit in stabilization of channel kinetics.

MUTATIONS CAUSING AChR DEFICIENCY WITH OR WITHOUT MINOR KINETIC ABNORMALITY Clinical Features

Fast-Channel Syndrome Caused by a Selective Abnormality of Gating Replacement of a valine by an isoleucine at residue 285 in TMD3 of the ␣ subunit (␣V285I) selectively reduces gating efficiency by depressing the channel opening rate (␤) and enhancing the channel closing rate (␣). The ␣V285I mutation also decreases AChR expression, which further impairs the safety margin of neuromuscular transmission.210 Fast-Channel Syndromes Caused by Unstable (Mode-Switching) Kinetics Two mutations causing unstable channel kinetics have been identified in expression studies, but only one of these, ⑀1254ins18, was investigated by in vitro microelectrode studies of the EP. The ⑀1254ins18 mutation, which is an in-frame duplication of codons 413 to 418 (STRDQE) in the long cytoplasmic loop of ⑀, also reduces AChR expression at the EP.138 At

The clinical phenotypes with these CMSs vary from mild to severe. Patients with recessive mutations in the ⑀ subunit are generally less affected than those with mutations in other subunits, but some harboring ⑀ subunit mutations can also be severely affected. The sickest patients have severe ocular, bulbar, and respiratory muscle weakness from birth and survive only with respiratory support and gavage feeding. They may be weaned from a respirator and begin to tolerate oral feedings during the first year of life, but have bouts of aspiration pneumonia and may need intermittent respiratory support during childhood and adult life. Motor milestones are severely delayed; they seldom learn to negotiate steps and can walk for only a short distance. Older patients close their mouth by supporting the jaw with the hand and elevate their eyelids with their fingers (Fig. 33–19). Facial deformities, prognathism, malocclusion, and scoliosis or kyphoscoliosis become noticeable during the second

853

Diseases of the Neuromuscular Junction

patchy (Fig. 33–20). The integrity of the junctional folds is preserved, but some EP regions are simplified and smaller than normal (Fig. 33–20A). The quantal response at the EP, indicated by the amplitude of MEPPs and MEPCs, is reduced, but quantal release by nerve impulse is frequently higher than normal. In patients with low-expressor or null mutations of the ⑀ subunit, single-channel patch-clamp recordings136,157,158 and immunocytochemical studies52 reveal the presence of fetal ␥-AChR at the EP.

Molecular Pathogenesis CMSs with severe EP AChR deficiency result from different types of homozygous or, more frequently, heterozygous recessive mutations in AChR subunit genes. The mutations are concentrated in the ⑀ subunit (Fig. 33–21). There are two possible reasons for this:

FIGURE 33–19 A severely affected 4-year-old girl who carries two heteroallelic mutations in the AChR ␤ subunit: a 9–base pair deletion in the long cytoplasmic loop (␤1276del9), which curtails AChR expression, and skipping of exon 8, which abolishes AChR expression. Note ptosis, exotropia, facial diplegia, and open mouth as well as tracheostomy and gastrostomy.

decade. Muscle bulk is reduced. The tendon reflexes are normal or hypoactive. The least affected patients pass their motor milestones with slight or no delay and only show mild ptosis and limited ocular ductions. They are clumsy in sports, fatigue easily, and cannot run well, climb rope, or do push-ups. In some instances, a myasthenic disorder is suspected only when the patient develops prolonged respiratory arrest on exposure to a curariform drug during a surgical procedure. Patients with intermediate clinical phenotypes experience moderate physical handicaps from early childhood. Ocular palsies and ptosis of the lids become apparent during the first year of life. These patients fatigue easily and cannot keep up with their peers in sports; they walk and negotiate stairs with difficulty, but can perform most activities of daily living.

End-Plate Studies Morphologic studies show an increased number of EP regions distributed over an increased span of the muscle fiber. AChR expression at the EP is markedly attenuated and

1. Expression of the fetal-type ␥ subunit, although at a low level, may compensate for absence of the ⑀ subunit,52,138,158 whereas patients harboring null mutations in subunits other than ⑀ might not survive for lack of a substituting subunit. 2. The gene encoding the ⑀ subunit, and especially exons coding for the long cytoplasmic loop, have a high GC content that likely predisposes to DNA rearrangements. Different types of recessive mutations causing severe EP AChR deficiency have been identified. Some mutations cause premature termination of the translational chain. These mutations are frameshifting,1,30,52,134,149,158,193,195 occur at a splice site,134,149 or produce a stop codon directly.159 An important mutation in this group is the 1369delG in the ⑀ subunit that results in loss of a C-terminal cysteine, C470, crucial to both maturation and surface expression of the adult receptor.38 Thus any mutation that truncates the ⑀ subunit upstream of C470 is predicted to inhibit ⑀ expression. A second type of recessive mutations are point mutations in the Ets binding site, or N-box, of the promoter region of the ⑀ subunit gene: ⑀-154G⬎A,2 ⑀-155G⬎A,147 and ⑀-156C⬎T.144 The N-box represents the end point of a signaling cascade driven by neuregulin through ERBB receptors. ERBB receptors phosphorylate mitogen-activated protein (MAP) kinases. Phosphorylated MAP kinases phosphorylate GABP␣ and GABP␤ (members of the Ets family of transcription factors), which bind to the N-box.36,65,187 That these mutations impair AChR expression is strong direct evidence that the neuregulin signaling pathway regulates synapse-specific transcription at the human neuromuscular junction. In addition to the above mutations, there are also missense mutations in a signal peptide region (⑀G-8R161 and ⑀V-13D134), and missense mutations involving residues essential for assembly of the pentameric receptor. Mutations of the latter type were observed in the

854

Pathology of the Peripheral Nervous System

FIGURE 33–20 Ultrastructural localization of AChR with peroxidase-labeled ␣-bungarotoxin at EP of patient homozygous for the ⑀553del7 null mutation (A) and at a control EP (B). The control EP shows heavy reaction for AChR on the terminal expansions of the junctional folds. At the patient’s EP, the junctional folds are simplified and the reaction for AChR is attenuated and patchy. (A, ⫻25,900; B, ⫻30,000.)

⑀ subunit at an N-glycosylation site (⑀S143L)161; in Cys 128 (⑀C128S), a residue that is an essential part of the C128-C142 disulfide loop in the extracellular domain138; in arginine 147 (⑀R147L), which is part of a short extracellular span of residues that contributes to subunit assembly158; in Thr 51 (⑀T51P)134; and in the long cytoplasmic loop of the ␤ subunit, causing the deletion of three codons.171 Finally, another group of missense mutations lead to the production of channels with minor kinetic effects. The kinetic consequences of these mutations are modest and overshadowed by reduced expression of the mutant gene. For example, ⑀R311W in the long cytoplasmic

loop between TMD3 and TMD4, and ⑀P245L in TMD1, decrease EP AChR below 10% of normal; ⑀R311W also reduces the open duration of channel events threefold158 and has thus mild fast-channel properties, whereas ⑀P245L increases the open duration of channel events threefold and has mild, nondominant slow-channel properties.158 Mutation of a corresponding proline in TMD1 of the ␦ subunit (␦P250Q) also markedly reduces AChR expression, but decreases rather than increases the duration of channel opening events by about twofold.193 It is also noteworthy that homozygous low-expressor ⑀ subunit mutations are endemic in Mediterranean or other

Diseases of the Neuromuscular Junction

855

FIGURE 33–21 Schematic diagram of low-expressor and null mutations reported in the ␣, ␤, ␦, and ⑀ subunits of AChR. Note that most mutations appear in the ⑀ subunit and especially in the long cytoplasmic loop between the third and fourth transmembrane domains (M). Square indicates a chromosomal microdeletion, hexagons are promoter mutations, open circles are missense mutations, closed circles are nonsense mutations, shaded circles are frameshifting mutations, and dotted circles are splicesite mutations. The most likely consequence of a splice-site mutation is skipping of a flanking exon; therefore, the splicesite mutations point to N-terminal codons of the predicted skipped exons. (From Engel, A. G., Ohno, K., and Sine, S. M.: Congenital myasthenic syndromes: progress over the past decade. Muscle Nerve 27:4, 2003, with permission.)

Near Eastern countries,134,148 and that the frameshifting ⑀1267delG mutation occurring at homozygosity is endemic in Gypsy families1,30,149 in whom it derives from a common founder.1 For a full list of AChR subunit gene mutations and the appropriate references, the reader is referred to a recently published gene table.152

CMS CAUSED BY RAPSYN DEFICIENCY In a subset of CMS patients with EP AChR deficiency but no mutation in any of the AChR subunits, the disease is traced to mutations in rapsyn. Rapsyn, under the influence of neurally supplied agrin, has a crucial role in concentrating the AChR in the postsynaptic membrane64

856

Pathology of the Peripheral Nervous System

and linking it to the subsynaptic cytoskeleton through dystroglycan.26 Rapsyn knockout mice cluster AChR poorly and fail to accumulate AChR at the EP.72 Thus rapsyn is an effector of agrin-induced clustering of AChR. In myotubes, agrin, MuSK, and possibly other myotube-specific mechanisms regulate rapsyn aggregation,213 but rapsyn expressed in heterologous systems self-aggregates and can then recruit AChRs, dystroglycan, and MuSK. The primary structure of rapsyn predicts distinct structural domains: a myristoylation signal at the N-terminus required for membrane association173; seven tetratricopeptide repeats (TPRs; codons 6 to 279) that subserve rapsyn self-association172,173; a coiled-coil domain (codons 298 to 331), the hydrophobic surface of which can bind to determinants within the long cytoplasmic loop of each AChR subunit10,172; a Cys-rich RING-H2 domain (codons 363 to 402) that binds to the cytoplasmic domains of ␤-dystroglycan11 and mediates the MuSK-induced phosphorylation of AChR118; and a serine phosphorylation site at codon 406 (Fig. 33–22A). Transcription of rapsyn in muscle is under the control of helix-loop-helix myogenic determination factors that bind to the cis-acting E-box sequence in the RAPSN promoter.159

Clinical Features Myasthenic symptoms can present in the neonatal period or later in life. Eight of 13 patients in one series21 and 1 of 4 patients in another series154 were born with arthro-

gryposis owing to hypomotility in utero. Other dysmorphic features are also common.7,21,154 Motor milestones are typically delayed, and fatigable weakness persists during life. Respiratory infections or other intercurrent illnesses precipitate increased weakness and respiratory insufficiency. A severely affected patient had episodes of respiratory compromise during infancy resulting in anoxic encephalopathy.7 Three patients were also reported in whom the disease was first noted at 13, 21, and 48 years of age. In the late-onset cases, the clinical course was mild and the disease mimicked seronegative autoimmune MG.21 Facial deformities associated with prognathism and malocclusion are pronounced in Near-Eastern Jewish patients who carry an E-box mutation (⫺38A⬎G) in the RAPSN gene159 but also can occur in other cases of rapsyn deficiency.8 The Near-Eastern Jewish patients have mild to severe weakness of the masticatory muscles, moderate to severe eyelid ptosis without ophthalmoparesis, facial weakness, and slurred or hypernasal speech. Cervical, truncal, and limb muscles are usually spared. Although these patients present at birth or soon thereafter, they show no signs of arthrogryposis. They have a stable and essentially benign course. EMG and single-fiber EMG studies may or may not reveal a defect of neuromuscular transmission. In some patients exercise or subtetanic stimulation is required to elicit a decremental EMG response.154 Similar EMG findings have been reported in the Near-Eastern Jewish patients with facial malformations.73,181

FIGURE 33–22 Schematic diagram showing structure of the RAPSN gene (A) and domains of rapsyn (B) with identified mutations. Seven tetratricopeptide repeats (TPRs) are required for rapsyn self-association; the coiled-coil domain binds to the long cytoplasmic loop of AChR subunits, and the RING-H2 domain links rapsyn to ␤-dystroglycan. Shaded areas in A indicate untranslated regions in RAPSN. E ⫽ E-box.

857

Diseases of the Neuromuscular Junction

End-Plate Studies The EPs, like those of patients with low-expressor mutation of the AChR, show a reduced expression of rapsyn and a proportionately reduced expression of AChR. In some patients, however, the number of AChRs per EP is only mildly reduced.154 Cholinesterase stains reveal multiple small EP regions dispersed over an increased span of individual muscle fibers. Ultrastructural studies show shallow postsynaptic folds and clefts, few secondary clefts, and smaller than normal nerve terminals and postsynaptic regions; the structural integrity of the pre- and postsynaptic regions is preserved.154,159 In vitro electrophysiologic studies show a higher than normal quantal release in some patients. Consistent with the EP AChR deficiency, the MEPP and MEPC amplitudes are reduced. Single-channel patch-clamp recordings show no kinetic abnormality of the AChR channel.

Molecular Pathogenesis Among the first four identified patients with rapsyn deficiency, one carried L14P in TPR1 and N88K in TPR3; two were homozygous for N88K; and one carried N88K and 553ins5 that frameshifts in TPR5.154 Several other rapsyn mutations were subsequently identified (see Fig. 33–22B). Expression of the first identified mutations in HEK cells showed that none hindered rapsyn self-association but all diminished co-clustering of AChR with rapsyn.154 That missense mutations in TPR domains decrease co-clustering of AChR with rapsyn implies that effects of these mutations propagate downstream to the coiled-coil domain, or that the mutations have an allosteric effect on the conformation of the coiled-coil or RING-H2 domains, or that the TPR mutations might compromise the cotransport of AChR with rapsyn to the cell surface. The N88K rapsyn mutation appears to be a relatively frequent cause of CMS.7,21,37,141,154,159 Haplotype analysis of 44 subjects in 10 kinships carrying the N88K mutation with four microsatellite markers revealed that only 20% of the mutant alleles shared the same haplotypes. Thus the N88K mutation either arose repeatedly and independently in different populations, or is an ancient mutation that arose even before the microsatellites diverged. A founder effect was noted in 10 independent European kinships,141 but the analysis was based on only one microsatellite.141 Other frameshift (46insC, 55ins5, C97X, Q124X and Y269X)37 and missense (Q3K, R91L, A142D, R151P, A246V, and G291D)141 mutations of rapsyn were also detected (see Fig. 33–22B). The E-box elements in the RAPSN promoter region with the consensus sequence of CANNTG are targets of myogenic determination factors and therefore par-

ticipate in modulating RAPSN expression. The two identified mutations in RAPSN E-box elements (⫺38A⬎G and ⫺27C⬎G) are the first examples of E-box mutations in humans. The ⫺38A⬎G mutation was homozygous in seven Oriental Jewish kinships with characteristic facial malformations and was traced to a common founder. The pathogenicity of both E-box mutations was confirmed in a luciferase reporter assay that showed attenuated reporter gene expression in C2C12 myotubes.159 The phenotype associated with the ⫺38A⬎G mutation appears to be homogeneous. However, there is no clear phenotype-genotype correlation in other rapsyndeficient patients. For example, among two patients homozygous for the same N88K mutation, one has severe myasthenic symptoms at the age of 2 years but the other has only mild weakness at the age of 27 years. A similar lack of genotype-phenotype correlation has been noted in other series.21

SODIUM CHANNEL MYASTHENIA Clinical Features Only one patient with this syndrome has been observed to date.206 This was the case of a 20-year-old normokalemic woman who had abrupt attacks of respiratory and bulbar paralysis since birth lasting 3 to 30 minutes and recurring one to three times per month. She survived only because since infancy she has been on an apnea monitor and received ventilatory support during apneic attacks. Her motor development was delayed, she always fatigued easily, and she had droopy eyelids. At age 20, she had eyelid ptosis and limited ocular ductions, as well as facial, truncal, and limb muscle weakness worsened by activity (Fig. 33–23); she could walk less than half a block, and could elevate her arms to the horizontal for only 20 seconds. She had a high-arched palate, adduction deformity of the knees and ankles, and increased lumbar lordosis. She was mentally retarded; brain magnetic resonance imaging revealed mild cerebral atrophy that was attributed to previous episodes of cerebral anoxia. Tests for anti-AChR antibodies were negative. The patient’s mother and sister were asymptomatic. No clinical data or DNA was available from the patient’s father. In EMG studies, a decremental response was elicited only by sustained high-frequency stimulation, or by low-frequency stimulation after a conditioning train of high-frequency stimulation. These findings were similar to those observed in the CMS caused by mutations in CHAT.

End-Plate Studies In an intercostal muscle specimen, type I fibers had a smaller mean diameter than type II fibers. The random

858

Pathology of the Peripheral Nervous System

from control muscle fibers. EPPs of the order of 40 mV depolarizing the membrane potential to ⫺40 mV or more failed to trigger action potentials and were recorded in the absence of d-tubocurarine. The MEPP amplitude and quantal content of the EPP were normal. Patch-clamp recordings from three EPs revealed AChR channels opening to a normal conductance of approximately 60 pS, and channel opening burst of normal duration. These findings established that the stimulation-dependent decrement of the CMAP was not caused by impaired resynthesis of acetylcholine and concomitant abnormal decrease of the EPP, and pointed to a defect in Nav1.4, the adult muscle fiber sodium channel. Other factors that compromise neuromuscular transmission in previously recognized CMS (e.g., decrease of evoked quantal release, diminished expression of AChE or AChR, a kinetic defect in AChR, or alteration of EP geometry) were excluded by the EP studies.

Molecular Pathogenesis

FIGURE 33–23 Patient with myasthenic syndrome caused by mutation of the Nav1.4 sodium channel. Note asymmetric ptosis, strabismus, lumbar lordosis, and adduction deformities of knees and ankles.

distribution of the histochemical fiber types was maintained. None of the intercostal muscle fibers harbored vacuoles of the type often observed in periodic paralysis muscle specimens. The configuration of the EPs on cholinesterase-reacted teased muscle fibers and the number of AChRs per EP, determined with 125I-labeled ␣-bungarotoxin, were normal. Electron microscopy showed normal nerve terminals and well-developed junctional folds at all EPs. One of the 63 observed EP regions was denuded of its nerve terminal. The density and distribution of AChR on the crests of the junctional folds, determined with peroxidase-labeled ␣-bungarotoxin, were also normal. Immunolocalization of sodium channels with an anti–pan sodium channel antibody showed similar surface membrane and similarly enhanced synaptic expression at patient and control muscle fibers. In vitro electrophysiology studies revealed that the muscle fiber membrane potential was similar to that recorded

That EPPs depolarizing the muscle fiber membrane potential to ⫺40 mV failed to activate action potentials prompted us to search for a mutation, SCN4A, that encodes Nav1.4. SCN4A harbored two heteroallelic missense mutations: S246L in the cytoplasmic link between the S4 and S5 segments of domain I, and V1442E in the extracellular link between the S3 and S4 segments of domain IV (Fig. 33–24). Neither mutation was observed in 400 normal alleles. Both S246 and V1442 are conserved across Nav1.4 channels of different species. Family analysis showed that the patient’s asymptomatic mother and sister are heterozygous for S246L. Expression studies in HEK showed that both sodium channel mutants expressed to normal levels in HEK cells. The salient finding consisted of a hyperpolarizing shift in the voltage dependence of fast inactivation; this was marked (⫺33 mV) for the V1442E mutant and milder (⫺7 mV) for the S246L mutant. The hyperpolarizing shift caused by the V1442E mutation predicts that nearly all V1442E channels at the EP are fast-inactivated, and hence inexcitable, at a normal resting membrane potential of ⫺80 mV. To determine whether use-dependent sodium channel inactivation might contribute to the abnormal decrement of the CMAP in the patient’s muscle at high-frequency stimulation, the response to a 50-Hz train of 3-ms pulses was measured in vitro. This revealed a precipitous drop of 30% of the normalized peak current amplitude during the first few pulses for the V1442E mutant, whereas wild-type and S246L channels showed only a 5% decrease. The rapid decrement can be attributed to trapping of V1442E channels in the fast-inactivated state

859

Diseases of the Neuromuscular Junction

FIGURE 33–24 Scheme of skeletal muscle sodium channel Nav1.4 encoded by SCN4A and the identified mutations in patient with sodium-channel myasthenia.

because recovery at ⫺100 mV was slowed threefold for this mutant.

Relation to Other Sodium Channel Disorders The present phenotype differs from that of periodic paralyses associated with hitherto identified mutations of SCN4A. The onset is neonatal, the disorder is normokalemic, the attacks selectively involve bulbar and respiratory muscles, physiologic rates of stimulation decrement the CMAP abnormally, and the muscle fiber membrane potential is normal when action potential generation fails. Periodic paralyses stemming from mutations in SCN4A present later in life, and these attacks typically spare cranial, bulbar, and respiratory muscles; the serum potassium level increases or declines during attacks in most cases, mild exercise for brief periods does not decrement the CMAP, and the resting membrane potential of the muscle fiber is decreased when action potential generation fails.55,119 The syndrome also differs from most other sodium channel disorders involving skeletal muscle. These are typically associated with a gain of function consisting of an excessive inward sodium current, whereas the present syndrome is associated with loss of function consisting of a markedly diminished inward sodium current. However, a subset of hypokalemic periodic paralyses caused by mutations of Arg 672 is also associated with loss of function and a hyperpolarizing shift of fast inactivation, but the shift is only ⫺10 mV and does not prevent action potential generation.202 The inheritance pattern for this CMS remains uncertain. The more severe V1442E mutation may be dominant, but this cannot be proven because this mutation derived from the patient’s father, who could not be evaluated, and because V1442E was observed only in combination with the heteroallelic S246L mutation. The S246L mutation alone was clinically silent and therefore either represents a rare polymorphism with detectable biophysical changes but no clinical or EMG phenotype, or is a recessive mutation. Despite this uncertainty, sodium channel myasthenia is a new allelic disorder to be grouped with periodic paralyses

and myotonias attributed to mutations in Nav1.4. This CMS is also the first in which the reduced safety margin of neuromuscular transmission occurs in the setting of a normal EP potential.

CMS ASSOCIATED WITH PLECTIN DEFICIENCY Plectin is a highly conserved and ubiquitously expressed intermediate filament–linking protein concentrated at sites of mechanical stress, such as the postsynaptic membrane of the EP, the sarcolemma, Z-discs in skeletal muscle, hemidesmosomes in skin, and intercalated disks in cardiac muscle. Pathogenic mutations in plectin are associated with a simplex variety of epidermolysis bullosa, a progressive myopathy, and a myasthenic syndrome.9 In a patient with epidermolysis bullosa simplex, a progressive myopathy, abnormal fatigability involving the ocular, facial, and limb muscles, as well as a decremental EMG response and no anti-AChR antibodies, revealed that plectin expression was absent in muscle and severely decreased in skin. Morphologic studies of muscle demonstrated necrotic and regenerating fibers and a wide spectrum of ultrastructural abnormalities. Many EPs had an abnormal configuration with chains of small regions over the fiber surface, and a few EPs displayed focal degeneration of the junctional folds. The EP AChR content was normal. In vitro electrophysiologic studies showed normal quantal release by nerve impulse, small MEPPs, and expression of fetal as well as adult AChR at the EPs.9

Therapy MYASTHENIA GRAVIS The mainstays of therapy are anticholinesterase medications, thymectomy, and immunosuppressive measures. A detailed discussion of modalities of therapy appropriate

860

Pathology of the Peripheral Nervous System

for the different grades of autoimmune MG are beyond the scope of this chapter. For a detailed discussion of the treatment of autoimmune MG, the reader is referred to a recent review.44

LAMBERT-EATON MYASTHENIC SYNDROME In noncarcinomatous LEMS, the mainstays of therapy are 3,4-DAP, which augments quantal release from the nerve terminal, and immunosuppressive measures. In carcinomatous LEMS, removal of the neoplasm as well as 3,4-DAP is indicated. For a detailed discussion of the treatment of LEMS, the reader is referred to a recent review.44

5.

6. 7.

8.

THE CONGENITAL MYASTHENIC SYNDROMES 9. For the purposes of therapy, the CMSs can be classified into two major categories: those that decrease and those that increase the synaptic response to ACh. The CMSs that decrease the synaptic response to ACh are treated with cholinergic agonists, namely drugs that inhibit AChE, and, in some cases, with 3,4-DAP. 3,4-DAP increases the number of quanta released by nerve impulse, and AChE inhibitors increase the number of AChRs activated by each quantum. An increased synaptic response to ACh occurs in the slow-channel syndromes and in EP AChE deficiency. The slow-channel syndromes are treated with long-lived open-channel blockers of AChR. There is no satisfactory therapy for EP AChE deficiency. The following special considerations apply: 1. End-plate ChAT deficiency is treated prophylactically with an oral AChE inhibitor (e.g., pyridostigmine) to prevent respiratory crises, and with parenteral prostigmine methylsulfate during respiratory crises. The patient’s parents must be taught how to administer prostigmine parenterally, and must own and be familiar with the use of a portable respiratory device. 2. The CMS associated with paucity of the synaptic vesicles and decreased quantal release is treated with AChE inhibitors. 3. One case of the LEMS-like CMS has responded to guanidine,5 a medication that acts like 3,4-DAP, but another patient, observed by the author, failed to respond to 3,4-DAP clinically. 4. In EP AChE deficiency, it is important to avoid cholinergic agonists. Anecdotal reports suggest that

ephedrine may mitigate the course of the disease. One respirator-dependent patient benefited from intermittent blockade of AChR by atracurium, an agent that protects AChR from overexposure to ACh, allowing for temporary withdrawal of respiratory support.19 The slow-channel syndromes are treated with quinidine sulfate66,86 or fluoxetine.87 Both medications act as longlived open-channel blockers of AChR and gradually improve the disease. The fast-channel syndromes respond well to combined therapy with pyridostigmine and 3,4-DAP. Patients with rapsyn deficiency also respond to pyridostigmine, and some derive striking further benefit from the use of 3,4-DAP.8 A patient with sodium channel myasthenia showed improved endurance on treatment with pyridostigmine; additional therapy with acetazolamide, which is known to mitigate periodic paralysis caused by Nav1.4 mutations, prevented further attacks of respiratory and bulbar weakness.206 In a patient with a CMS associated with plectin deficiency, pyridostigmine failed to improve the patient’s symptoms, but 3,4-DAP improved her strength and endurance.9

For further discussion of the treatment of the CMSs, the reader is referred to a recent review.44

REFERENCES 1. Abicht, A., Stucka, R., Karcagi, V., et al.: A common mutation (⑀1267delG) in congenital myasthenic patients of Gypsy ethnic origin. Neurology 53:1564, 1999. 2. Abicht, A., Stucka, R., Schmidt, C., et al.: A newly identified chromosomal microdeletion and an N-box mutation of the AChRe gene cause a congenital myasthenic syndrome. Brain 125:1005, 2002. 3. Anderson, C. R., and Stevens, C. F.: Voltage clamp analysis of acetylcholine produced end-plate current fluctuations at frog neuromuscular junction. J. Physiol. (Lond.) 235:655, 1973. 4. Apparsundaram, S., Ferguson, S. M., George, A. L. Jr., and Blakely, R. D.: Molecular cloning of a human, hemicholinium3-sensitive choline transporter. Biochem. Biophys. Res. Commun. 276:862, 2000. 5. Bady, B., Chauplannaz, G., and Carrier, H.: Congenital Lambert-Eaton myasthenic syndrome. J. Neurol. Neurosurg. Psychiatry 50:476, 1987. 6. Bady, B., Vial, C., and Chauplannaz, G.: Syndrome de Lambert-Eaton: étude clinique et électrophysiologique de 18 cas associes à un cancer du poumon. Rev. Neurol. (Paris) 148:513, 1992. 7. Banwell, B. L., Brengman, J. M., Ohno, K., et al.: Novel truncating mutation causing congenital myasthenic syndrome. Neurology 60(Suppl. 1):A420, 2003.

Diseases of the Neuromuscular Junction 8. Banwell, B. L., Ohno, K., Sieb, J. P., and Engel, A. G.: Novel truncating RAPSN mutation causing congenital myasthenic syndrome responsive to 3,4-diaminopyridine. Neuromuscul. Disord. 14:202, 2004. 9. Banwell, B. L, Russel, J., Fukudome, T., et al.: Myopathy, myasthenic syndrome, and epidermolysis bullosa simplex due to plectin deficiency. J. Neuropathol. Exp. Neurol. 58:832, 1999. 10. Bartoli, M., and Cohen, J. B.: Identification of the modular domains of rapsyn binding to nicotinic acetylcholine receptor (AChR) and to dystroglycan. Soc. Neurosci. Abstr. 27:904.16, 2001. 11. Bartoli, M., Ramarao, M. K., and Cohen, J. B.: Interactions of the rapsyn RING-H2 domain with dystroglycan. J. Biol. Chem. 276:24911, 2001. 12. Batocchi, A. P., Evoli, A., Servidei, S., et al.: Myasthenia gravis during interferon alpha therapy. Neurology 42:382, 1995. 13. Beech, R. L., Vaca, K., and Pilar, G.: Ionic and metabolic requirements for high-affinity choline uptake and acetylcholine synthesis in nerve terminals of a neuromuscular junction. J. Neurochem. 34:1387, 1980. 14. Betz, W., and Sakmann, B.: Effects of proteolytic enzymes on function and structure of frog neuromuscular junctions. J. Physiol. (Lond.) 230:673, 1973. 15. Bhakdi, S., and Trnum-Jensen, J.: Complement lysis: a hole is a hole. Immunol. Today 12:318, 1991. 16. Bon, S., Coussen, F., and Massoulié, J.: Quaternary associations of acetylcholinesterase. II. The polyproline attachment domain of the collagen tail. J. Biol. Chem. 272:3016, 1997. 17. Bööj, S., Larsson, P.-A., Dahllöf, A.-G., and Dahlström, A.: Axonal transport of synapsin I and cholinergic synaptic vesiclelike material: further immunohistochemical evidence for transport of axonal cholinergic transmitter vesicles in motor neurons. Acta Physiol. Scand. 128:155, 1986. 18. Brejc, K., van Dijk, W. V., Schuurmans, M., et al.: Crystal structure of ACh-binding protein reveals the ligand-binding domain of nicotinic receptors. Nature 411:269, 2001. 19. Breningstall, G. N., Kuracheck, S. C., Fugate, J. H., and Engel, A. G.: Treatment of congenital endplate acetylcholinesterase deficiency by neuromuscular blockade. J. Child Neurol. 11:345, 1996. 20. Brownlow, S., Webster, R., Croxen, R., et al.: Acetylcholine receptor ␦ subunit mutations underlie a fast-channel myasthenic syndrome and arthrogryposis multiplex congenita. J. Clin. Invest. 108:125, 2001. 21. Burke, G., Cossins, J., Maxwell, S., et al.: Rapsyn mutations in hereditary myasthenia: distinct early- and late-onset phenotypes. Neurology 61:826, 2003. 22. Byring, R. F., Pihko, H., Shen, X.-M., et al.: Congenital myasthenic syndrome associated with episodic apnea and sudden infant death. Neuromuscul. Disord. 12:548, 2002. 23. Campbell, A. K., Daw, R. A., and Luzio, J. P.: Rapid increase in intracellular free Ca2⫹ induced by antibody and complement. FEBS Lett. 107:55, 1979. 24. Campbell, H., and Bramwell, E.: Myasthenia gravis pseudoparalytica: review of 70 case reports, including nine new patients. Brain 23:277, 1900. 25. Caroni, P., and Grandes, P.: Nerve sprouting in innervated adult skeletal muscle induced by exposure to elevated levels of insulin-like growth factors. J. Cell Biol. 110:1307, 1990.

861

26. Cartaud, A., Coutant, S., Petrucci, T. C., and Cartaud, J.: Evidence for in situ and in vitro association between ␤-dystroglycan and the subsynaptic 43K rapsyn protein: consequence for acetylcholine receptor clustering at the synapse. J. Biol. Chem. 273:11321, 1998. 27. Charlton, M. P., Smith, S. J., and Zuker, R.: Role of presynaptic calcium ions and channels in synaptic facilitation and depression at the squid giant synapse. J. Physiol. (Lond.) 323:173, 1982. 28. Clouston, P. D., Saper, C. B., Arbizu, T., et al.: Paraneoplastic cerebellar degeneration. III. Cerebellar degeneration, cancer, and the Lambert-Eaton myasthenic syndrome. Neurology 42:1944, 1992. 29. Croxen, R., Newland, C., Beeson, D., et al.: Mutations in different functional domains of the human muscle acetylcholine receptor ␣ subunit in patients with the slow-channel congenital myasthenic syndrome. Hum. Mol. Genet. 6:767, 1997. 30. Croxen, R., Newland, C., Betty, M., et al.: Novel functional ⑀-subunit polypeptide generated by a single nucleotide deletion in acetylcholine receptor deficiency congenital myasthenic syndrome. Ann. Neurol. 46:639, 1999. 31. del Castillo, J., and Katz, B.: Quantal components of the end-plate potential. J. Physiol. (Lond.) 124:560, 1954. 32. Deprez, P. N., and Inestrosa, N. C.: Two heparin-binding domains are present on the collagenic tail of asymmetric acetylcholinesterase. J. Biol. Chem. 270:11043, 1995. 33. Deymeer, F., Serdaroglu, P., and Özdemir, C.: Familial infantile myasthenia: confusion in terminology. Neuromuscul. Disord. 9:129, 1999. 34. Donger, C., Krejci, E., Serradell, P., et al.: Mutation in the human acetylcholinesterase-associated gene, COLQ, is responsible for congenital myasthenic syndrome with end-plate acetylcholinesterase deficiency. Am. J. Hum. Genet. 63:967, 1998. 35. Drachman, D. B., Adams, R. N., Josifek, L. F., and Sel, S. G.: Functional activities of autoantibodies to acetylcholine receptors and the clinical severity of myasthenia gravis. N. Engl. J. Med. 307:769, 1982. 36. Duclert, A., Savatier, N., Schaeffer, L., and Changeux, J.-P.: Identification of an element crucial for the sub-synaptic expression of the acetylcholine receptor epsilon-subunit gene. J. Biol. Chem. 271:17433, 1996. 37. Dunne, V., and Maselli, R. A.: Identification of pathogenic mutations in the human rapsyn gene. Hum. Genet. 48:204, 2003. 38. Ealing, J., Webster, R., Brownlow, S., et al.: Mutations in congenital myasthenic syndromes reveal an e subunit C-terminal cysteine, C470, crucial for maturation and surface expressions of adult AChR. Hum. Mol. Genet. 11:3087, 2002. 39. Eaton, L. M., and Lambert, E. H.: Electromyography and electrical stimulation of nerves in diseases of the motor unit: observations on a myasthenic syndrome associated with malignant tumors. J. Am. Med. Assoc. 163:1117, 1957. 40. Eiden, L. E.: The cholinergic gene locus. J. Neurochem. 70:2227, 1998. 41. Elmqvist, D., and Lambert, E. H.: Detailed analysis of neuromuscular transmission in a patient with the myasthenic syndrome sometimes associated with bronchogenic carcinoma. Mayo Clin. Proc. 43:689, 1968.

862

Pathology of the Peripheral Nervous System

42. Elmqvist, D., and Quastel, D. M. J.: A quantitative study of end-plate potentials in isolated human muscle. J. Physiol. (Lond.) 178:505, 1965. 43. Elrington, G. M., Murray, N. M., Spiro, S. G., and NewsomDavis, J.: Neurological paraneoplastic syndromes in patients with small cell lung cancer: a prospective survey of 150 patients. J. Neurol. Neurosurg. Psychiatry 54:764, 1991. 44. Engel, A. G.: Myasthenia gravis and myasthenic syndromes. In Noseworthy, J. N. (ed.): Neurological Therapeutics: Principles and Practice. New York, Martin Dunitz, p. 2378, 2003. 45. Engel, A. G., and Arahata, K.: The membrane attack complex of complement at the endplate in myasthenia gravis. Ann. N. Y. Acad. Sci. 505:326, 1987. 46. Engel, A. G., Fukuoka, T., Lang, B., et al.: Lambert-Eaton myasthenic syndrome IgG: early morphologic effects and immunolocalization at the motor endplate. Ann. N. Y. Acad. Sci. 505:333, 1987. 47. Engel, A. G., and Lambert, E. H.: Congenital myasthenic syndromes. Electroencephalogr. Clin. Neurophysiol. Suppl. 39:91, 1987. 48. Engel, A. G., Lambert, E. H., and Gomez, M. R.: A new myasthenic syndrome with end-plate acetylcholinesterase deficiency, small nerve terminals, and reduced acetylcholine release. Ann. Neurol. 1:315, 1977. 49. Engel, A. G., Lambert, E. H., and Howard, F. M.: Immune complexes (IgG and C3) at the motor end-plate in myasthenia gravis: ultrastructural and light microscopic localization and electrophysiologic correlations. Mayo Clin. Proc. 52:267, 1977. 50. Engel, A. G., Lambert, E. H., Mulder, D. M., et al.: A newly recognized congenital myasthenic syndrome attributed to a prolonged open time of the acetylcholine-induced ion channel. Ann. Neurol. 11:553, 1982. 51. Engel, A. G., Lindstrom, J. M., Lambert, E. H., and Lennon, V. A.: Ultrastructural localization of the acetylcholine receptor in myasthenia gravis and in its experimental autoimmune model. Neurology 27:307, 1977. 52. Engel, A. G., Ohno, K., Bouzat, C., et al.: End-plate acetylcholine receptor deficiency due to nonsense mutations in the ⑀ subunit. Ann. Neurol. 40:810, 1996. 53. Engel, A. G., Ohno, K., Milone, M., et al.: New mutations in acetylcholine receptor subunit genes reveal heterogeneity in the slow-channel congenital myasthenic syndrome. Hum. Mol. Genet. 5:1217, 1996. 54. Engel, A. G., Ohno, K., and Sine, S. M.: Congenital myasthenic syndromes. In Engel, A. G. (ed.): Myasthenia Gravis and Myasthenic Disorders. New York, Oxford University Press, p. 251, 1999. 55. Engel, A. G., Potter, C. S., and Rosevear, J. W.: Clinical and electromyographic studies in a patient with primary hypokalemic periodic paralysis. Am. J. Med. 38:626, 1965. 56. Engel, A. G., Sahashi, K., and Fumagalli, G.: The immunopathology of acquired myasthenia gravis. Ann. N. Y. Acad. Sci. 377:158, 1981. 57. Engel, A. G., and Santa, T.: Histometric analysis of the ultrastructure of the neuromuscular junction in myasthenia gravis and in the myasthenic syndrome. Ann. N. Y. Acad. Sci. 183:46, 1971.

58. Erickson, J. D., Varoqui, H., Eiden, L. E., et al.: Functional identification of a vesicular acetylcholine transporter and its expression from a ‘cholinergic’ gene locus. J. Biol. Chem. 269:21929, 1994. 59. Esser, A. F.: Big MAC attack: complement proteins cause leaky patches. Immunol. Today 12:316, 1991. 60. Fatt, P., and Katz, B.: Spontaneous subthreshold activity at motor nerve endings. J. Physiol. (Lond.) 117:109, 1952. 61. Fawcett, P. R. W., McLachlan, S. M., Nicholson, L. V. B., et al.: D-penicillamine associated myasthenia gravis: immunological and electrophysiological studies. Muscle Nerve 5:328, 1982. 62. Fertuck, H. C., and Salpeter, M. M.: Quantitation of junctional and extrajunctional acetylcholine receptors by electron microscope autoradiography after 125I-␣-bungarotoxin binding at mouse neuromuscular junctions. J. Cell. Biol. 69:144, 1976. 63. Flucher, B. E, and Daniels, M. P.: Distribution of Na⫹ channels and ankyrin in neuromuscular junctions is complementary to that of acetylcholine receptors and the 43 kd protein. Neuron 3:163, 1989. 64. Froehner, S. C., Luetje, C. W., Scotland, P. B., and Patrick, J.: The postsynaptic 43K protein clusters muscle nicotinic acetylcholine receptors in Xenopus oocytes. Neuron 5:403, 1990. 65. Fromm, L., and Burden, S. J.: Synapse-specific and neuregulin-induced transcription require an Ets site that binds GABP␣/GAPB␤. Genes Dev. 12:3074, 1998. 66. Fukudome, T., Ohno, K., Brengman, J. M., and Engel, A. G.: Quinidine normalizes the open duration of slow-channel mutants of the acetylcholine receptor. Neuroreport 9:1907, 1998. 67. Fukunaga, H., Engel, A. G., Lang, B., and Vincent, A.: Passive transfer of Lambert-Eaton myasthenic syndrome with IgG from man to mouse depletes the presynaptic membrane active zones. Proc. Natl. Acad. Sci. U. S. A. 80:7636, 1983. 68. Fukunaga, H., Engel, A. G., Osame, M., and Lambert, E. H.: Paucity and disorganization of presynaptic membrane active zones in the Lambert-Eaton myasthenic syndrome. Muscle Nerve 5:686, 1982. 69. Fukuoka, T., Engel, A. G., Lang, B., and Vincent, A.: Lambert-Eaton myasthenic syndrome: II. Immunoelectron microscopy localization of IgG at the mouse motor endplate. Ann. Neurol. 22:200, 1987. 70. Fukuoka, T., Engel, A. G., Lang, B., et al.: Lambert-Eaton myasthenic syndrome: I. Early morphological effects of IgG on the presynaptic membrane active zones. Ann. Neurol. 22:193, 1987. 71. Fulpius, B. W., Zurn, A. D., Granato, D. A., and Leder, R. M.: Acetylcholine receptor and myasthenia gravis. Ann. N. Y. Acad. Sci. 274:116, 1976. 72. Gautam, M., Noakes, P. G., Mudd, J., et al.: Failure of postsynaptic specialization to develop at neuromuscular junctions of rapsyn-deficient mice. Nature 377:232, 1995. 73. Goldhammer, Y., Blatt, I., Sadeh, M., and Goodman, R. M.: Congenital myasthenia associated with facial malformations in Iraqi and Iranian Jews. Brain 113:1291, 1990.

Diseases of the Neuromuscular Junction 74. Gomez, C. M., Maselli, R., Gammack, J., et al.: A beta-subunit mutation in the acetylcholine receptor gate causes severe slow-channel syndrome. Ann. Neurol. 39:712, 1996. 75. Gomez, C. M., Maselli, R., Staub, J., et al.: Novel ␦ and ␤ subunit acetylcholine receptor mutations in the slow-channel syndrome demonstrate phenotypic variability. Soc. Neurosci. Abstr. 24:484, 1998. 76. Gomez, C. M., Maselli, R., Vohra, B. P. S., et al.: Novel delta subunit mutation in slow-channel syndrome causes severe weakness by novel mechanism. Ann. Neurol. 51:102, 2002. 77. Greer, M., and Schotland, M.: Myasthenia gravis in the newborn. Pediatrics 26:101, 1960. 78. Grob, D., Brunner, N. G., and Namba, T.: The natural course of myasthenia gravis and effect of therapeutic measures. Ann. N. Y. Acad. Sci. 377:652, 1981. 79. Grohovaz, F., Limbrick, A. R., and Miledi, R.: Acetylcholine receptors at the rat neuromuscular junction as revealed by deep etching. Proc. R. Soc. Lond. Ser. B 215:147, 1982. 80. Gurtubay, I. G., Morales, G., Arechaga, O., and Gallego, J.: Development of myasthenia gravis after interferon alpha therapy. Electromyogr. Clin. Neurophysiol. 39:75, 1999. 81. Gutmann, L., Crosby, T. W., Takamori, M., and Martin, J. D.: The Eaton-Lambert syndrome and autoimmune disorders. Am. J. Med. 53:354, 1972. 82. Gutmann, L., Weidman, D., and Gutierrez, A.: LambertEaton myasthenic syndrome with prominent postexercise exhaustion. Muscle Nerve 16:716, 1993. 83. Hall, Z. W., and Kelly, R. B.: Enzymatic detachment of endplate acetylcholinesterase from muscle. Nature 232:62, 1971. 84. Harada, H., Tamaoka, A., Kohno, Y., et al.: Exacerbation of myasthenia gravis in a patient after interferon-beta treatment for chronic active hepatitis C. J. Neurol. Sci. 165:182, 2001. 85. Harper, C. M.: Electrodiagnosis of endplate disease. In Engel, A. G. (ed.): Myasthenia Gravis and Myasthenic Disorders. New York, Oxford University Press, p. 65, 1999. 86. Harper, C. M., and Engel, A. G.: Quinidine sulfate therapy for the slow-channel congenital myasthenic syndrome. Ann. Neurol. 43:480, 1998. 87. Harper, C. M., Fukudome, T., and Engel, A. G.: Treatment of slow channel congenital myasthenic syndrome with fluoxetine. Neurology 60:170, 2003. 88. Hart, Z., Sahashi, K., Lambert, E. H., et al.: A congenital, familial, myasthenic syndrome caused by a presynaptic defect of transmitter resynthesis of mobilization. Neurology 29:559, 1979. 89. Hartzell, H. C., Kuffler, S. W., and Yoshikami, D.: Postsynaptic potentiation: interaction between quanta of acetylcholine at the skeletal neuromuscular synapse. J. Physiol. (Lond.) 251:427, 1975. 90. Hartzell, H. C., Kuffler, S. W., and Yoshikami, D.: The number of acetylcholine molecules in a quantum and the interaction between quanta at the subsynaptic membrane of the skeletal neuromuscular synapse. Symp. Quant. Biol. 40:175, 1976. 91. Hatton, C. J., Shelley, C., Brydson, M., et al.: Properties of the human muscle nicotinic receptor, and of the slow-channel

92.

93.

94.

95.

96.

97.

98.

99.

100. 101.

102.

103.

104.

105.

106.

107.

108.

863

myasthenic syndrome mutant ⑀L221F, inferred from maximum likelihood fits. J. Physiol. (Lond.) 547:729, 2003. Hawley, R. J., Cohen, M. H., Saini, N., et al.: The carcinomatous neuromyopathy of oat cell lung cancer. Ann. Neurol. 7:65, 1980. Heuser, J. E., and Reese, T. S.: Evidence for recycling of synaptic vesicle membrane during transmitter release at the frog neuromuscular junction. J. Cell Biol. 57:315, 1973. Heuser, J. E., Reese, T. S., Dennis, M. J., et al.: Synaptic vesicle exocytosis captured by quick freezing and correlated with quantal transmitter release. J. Cell Biol. 81:275, 1979. Heuser, J. E., and Salpeter, S. R.: Organization of acetylcholine receptors in quick-frozen, deep-etched, and rotary-replicated Torpedo postsynaptic membrane. J. Cell Biol. 82:150, 1979. Hirokawa, N., and Heuser, J. E.: Internal and external differentiations of the postsynaptic membrane at the neuromuscular junction. J. Neurocytol. 11:487, 1982. Hoch, W., McConville, J., Helms, S., et al.: Autoantibodies to the receptor tyrosine kinase MuSK in patients with myasthenia gravis without acetylcholine receptor antibodies. Nat. Med. 7:365, 2001. Hohlfeld, R., and Wekerle, H.: The immunopathogenesis of myasthenia gravis. In Engel, A. G. (ed.): Myasthenia Gravis and Myasthenic Disorders. New York, Oxford University Press, p. 87, 1999. Hubbard, J. I., Jones, S. F., and Landau, E. M.: On the mechanism by which calcium and magnesium affect the release of transmitter by nerve impulses. J. Physiol. (Lond.) 196:75, 1968. Hutchinson, D. O., Walls, T. J., Nakano, S., et al.: Congenital endplate acetylcholinesterase deficiency. Brain 116:633, 1993. Ishigaki, K., Nicolle, D., Krejci, E., et al.: Two novel mutations in the COLQ gene causing endplate acetylcholinesterase deficiency. Neuromuscul. Disord. 13:236, 2003. Jahn, K., Franke, C., and Bufler, J.: Mechanism of block of nicotinic acetylcholine receptor channels by purified IgG from seropositive patients with myasthenia gravis. Neurology 54:474, 2000. Jones, S. F., and Kwanbunbumpen, S.: The effects of nerve stimulation and hemicholinium on synaptic vesicles at the mammalian neuromuscular junction. J. Physiol. (Lond.) 207:31, 1970. Katz, B., and Miledi, R.: The release of acetylcholine from nerve endings by graded electrical pulses. Proc. R. Soc. Lond. Ser. B 167:23, 1967. Katz, B., and Miledi, R.: Tetrodotoxin-resistant electrical activity in presynaptic terminals. J. Physiol. (Lond.) 203:459, 1969. Katz, B., and Miledi, R.: The statistical nature of the acetylcholine potential and its molecular components. J. Physiol. (Lond.) 224:665, 1972. Katz, B., and Miledi, R.: The binding of acetylcholine to receptors and its removal from the synaptic cleft. J. Physiol. (Lond.) 231:549, 1973. Katz, B., and Thesleff, S.: On the factors which determine the amplitude of the miniature end-plate potential. J. Physiol. (Lond.) 137:267, 1957.

864

Pathology of the Peripheral Nervous System

109. Kiene, L.-M., and Stadler, H.: Synaptic vesicles in electromotoneurones. I. Axonal transport, site of transmitter uptake and processing of a core proteoglycan during maturation. EMBO J. 6:2209, 1987. 110. Kimbell, L. M., Ohno, K., Rotundo, R. L., and Engel, A. G.: Transplanting mutant human collagenic tailed acetylcholinesterase onto the frog neuromuscular junction: evidence for an attachment defect in a congenital myasthenic syndrome. Mol. Biol. Cell 12(Suppl.):161a, 2001. 111. Kuffler, S. W., and Yoshikami, D.: The number of transmitter molecules in the quantum: an estimate from iontophoretic application of acetylcholine at the neuromuscular synapse. J. Physiol. (Lond.) 251:465, 1975. 112. Lambert, E. H., Eaton, L. M., and Rooke, E. D.: Defect of neuromuscular transmission associated with malignant neoplasm. Am. J. Physiol. 187:612, 1956. 113. Lambert, E. H., and Elmqvist, D.: Quantal components of end-plate potentials in the myasthenic syndrome. Ann. N. Y. Acad. Sci. 183:183, 1971. 114. Lambert, E. H., and Rooke, E. D.: Myasthenic state and lung cancer. In Brain, W. R., and Norris, F. H. (eds): The Remote Effects of Cancer on the Nervous System. New York, Grune & Stratton, p. 67, 1965. 115. Lambert, E. H., Rooke, E. D., Eaton, L. M., and Hodgson, C. H.: Myasthenic syndrome occasionally associated with bronchial neoplasm: neurophysiologic studies. In Viets, H. R. (ed.): Myasthenia Gravis. Springfield, IL, Charles C Thomas, p. 362, 1961. 116. Land, B. R., Salpeter, E. E., and Salpeter, M. M.: Acetylcholine receptor site density affects the rising phase of miniature endplate currents. Proc. Natl. Acad. Sci. U.S.A. 77:3736, 1980. 117. Lang, B., Newsom-Davis, J., Wray, D. W., et al.: Autoimmune aetiology for myasthenic (Lambert-Eaton) syndrome. Lancet 2:224, 1981. 118. Lee, Y. I., Swope, S. L., and Ferns, M. J.: Rapsyn’s C-terminal domain mediates MuSK-induced phosphorylation of the AChR. Mol. Biol. Cell 13:395a, 2002. 119. Lehmann-Horn, F., and Jurkat-Rott, K.: Voltage-gated ion channels and hereditary disease. Physiol. Rev. 79:1317, 1999. 120. Li, F.-Y., Szobor, A., Croxen, R., et al.: Dominantly inherited myasthenia gravis as a separate genetic entity without involvement of defined candidate gene loci. Int. J. Mol. Med. 7:289, 2001. 121. Lindstrom, J. M., Seybold, M. E., Lennon, V. A., et al.: Antibody to acetylcholine receptor in myasthenia gravis: prevalence, clinical correlates and diagnostic value. Neurology 26:1054, 1976. 122. Llinás, R., and Nicholson, C.: Calcium in depolarization secretion coupling: an aequorin study in squid giant synapse. Proc. Natl. Acad. Sci. U.S.A. 72:187, 1975. 123. Llinás, R., Sugimori, M., Lin, J.-W., et al.: ATP-dependent directional movement of rat synaptic vesicles injected into the presynaptic terminal of squid giant synapse. Proc. Natl. Acad. Sci. U.S.A. 86:5656, 1989. 124. Llinás, R., Sugimori, M., and Siver, R. B.: Presynaptic calcium concentration microdomains and transmitter release. J. Physiol. (Paris) 86:135, 1992.

125. Lundh, H., Nilsson, O., and Rosen, I.: Treatment of LambertEaton syndrome: 3,4-diaminopyridine and pyridostigmine. Neurology 34:1324, 1984. 126. Martin, A. R.: Current concepts of pre- and post-junctional mechanisms in neuromuscular transmission. Ann. N. Y. Acad. Sci. 274:3, 1976. 127. Martin, A. R.: Junctional transmission. II. Presynaptic mechanisms. In Handbook of Neurophysiology. Bethesda, MD, American Physiological Society, p. 329, 1977. 128. Martin, A. R.: Amplification of neuromuscular transmission by postjunctional folds. Proc. R. Soc. Lond. B 258:321, 1994. 129. Maselli, R. A., Chen, D., Delores, M. O., et al.: Choline acetyltransferase mutations in myasthenic syndrome due to deficient acetylcholine resynthesis. Muscle Nerve 27:180, 2003. 130. Maselli, R. A., Kong, D. Z., Bowe, C. M., et al.: Presynaptic congenital myasthenic syndrome due to quantal release deficiency. Neurology 57:279, 2001. 131. Matthews-Bellinger, J., and Salpeter, M. M.: Distribution of acetylcholine receptors at frog neuromuscular junctions with a discussion of some physiological implications. J. Physiol. (Lond.) 279:197, 1978. 132. McMahan, U. J., Sanes, J. S., and Marshall, L. M.: Cholinesterase is associated with the basal lamina at the neuromuscular junction. Nature 271:172, 1978. 133. McMahon, H. T., and Nicholls, D. G.: The bioenergetics of neurotransmitter release. Biochim. Biophys. Acta 1059:243, 1991. 134. Middleton, L., Ohno, K., Christodoulou, K., et al.: Congenital myasthenic syndromes linked to chromosome 17p are caused by defects in acetylcholine receptor ⑀ subunit gene. Neurology 53:1076, 1999. 135. Miledi, R.: Transmitter release by injection of calcium ions into nerve terminals. Proc. R. Soc. Lond. Ser. B 183:421, 1973. 136. Milone, M., Ohno, K., Pruitt, J. N., et al.: Congenital myasthenic syndrome due to frameshifting acetylcholine receptor epsilon subunit mutation. Soc. Neurosci. Abstr. 22:1942, 1996. 137. Milone, M., Wang, H.-L., Ohno, K., et al.: Slow-channel syndrome caused by enhanced activation, desensitization, and agonist binding affinity due to mutation in the M2 domain of the acetylcholine receptor alpha subunit. J. Neurosci. 17:5651, 1997. 138. Milone, M., Wang, H.-L., Ohno, K., et al.: Mode switching kinetics produced by a naturally occurring mutation in the cytoplasmic loop of the human acetylcholine receptor ⑀ subunit. Neuron 20:575, 1998. 139. Monod, J., Wyman, J., and Changeux, J.-P.: On the nature of allosteric transitions: a plausible model. J. Mol. Biol. 12:88, 1965. 140. Mora, M., Lambert, E. H., and Engel, A. G.: Synaptic vesicle abnormality in familial infantile myasthenia. Neurology 37:206, 1987. 141. Müller, J. S., Mildner, G., Müller-Felber, W., et al.: Rapsyn N88K is a frequent cause of CMS in European patients. Neurology 60:1805, 2003. 142. Nagel, A., Engel, A. G., Lang, B., and Fukuoka, T.: LambertEaton myasthenic syndrome IgG depletes presynaptic

Diseases of the Neuromuscular Junction

143.

144.

145.

146. 147.

148.

149.

150.

151.

152. 153.

154.

155.

156.

157.

158.

membrane active zone particles by antigenic modulation. Ann. Neurol. 24:552, 1988. Nakano, S., and Engel, A. G.: Myasthenia gravis: quantitative immunocytochemical analysis of inflammatory cells and detection of complement membrane attack complex at the end-plate in 30 patients. Neurology 43:1167, 1993. Nichols, P. R., Croxen, R., Vincent, A., et al.: Mutation of the acetylcholine receptor ⑀-subunit promoter in congenital myasthenic syndrome. Ann. Neurol. 45:439, 1999. Oda, Y., Nakanishi, I., and Deguchi, T.: A complementary DNA for human choline acetyltransferase induces two forms of enzyme with different molecular weights in cultured cells. Brain Res. Mol. Brain Res. 16:287, 1992. Oh, S. J.: The Eaton-Lambert syndrome in ocular myasthenia gravis. Arch. Neurol. 31:183, 1974. Ohno, K., Anlar, B., and Engel, A. G.: Congenital myasthenic syndrome caused by a mutation in the Ets-binding site of the promoter region of the acetylcholine receptor ⑀ subunit gene. Neuromuscul. Disord. 9:131, 1999. Ohno, K., Anlar, B., Ozdemir, C., et al.: Frameshifting and splice-site mutations in acetylcholine receptor ⑀ subunit gene in 3 Turkish kinships with congenital myasthenic syndromes. Ann. N. Y. Acad. Sci. 841:189, 1998. Ohno, K., Anlar, B., Özdirim, E., et al.: Myasthenic syndromes in Turkish kinships due to mutations in the acetylcholine receptor. Ann. Neurol. 44:234, 1998. Ohno, K., Brengman, J. M., Felice, K. J., et al.: Congenital endplate acetylcholinesterase deficiency caused by a nonsense mutation and an A-to-G splice site mutation at position ⫹3 of the collagen-like tail subunit gene (COLQ): how does G at position ⫹3 result in aberrant splicing? Am. J. Hum. Genet. 65:635, 1999. Ohno, K., Brengman, J. M., Tsujino, A., and Engel, A. G.: Human endplate acetylcholinesterase deficiency caused by mutations in the collagen-like tail subunit (ColQ) of the asymmetric enzyme. Proc. Natl. Acad. Sci. U.S.A. 95:9654, 1998. Ohno, K., and Engel, A. G.: Congenital myasthenic syndromes: gene mutations. Neuromuscul. Disord. 13:283, 2003. Ohno, K., Engel, A. G., Brengman, J. M., et al.: The spectrum of mutations causing endplate acetylcholinesterase deficiency. Ann. Neurol. 47:162, 2000. Ohno, K., Engel, A. G., Shen, X.-M., et al.: Rapsyn mutations in humans cause endplate acetylcholine receptor deficiency and myasthenic syndrome. Am. J. Hum. Genet. 70:875, 2002. Ohno, K., Hutchinson, D. O., Milone, M., et al.: Congenital myasthenic syndrome caused by prolonged acetylcholine receptor channel openings due to a mutation in the M2 domain of the ⑀ subunit. Proc. Natl. Acad. Sci. U.S.A. 92:758, 1995. Ohno, K., Milone, M., Brengman, J. M., et al.: Slow-channel congenital myasthenic syndrome caused by a novel mutation in the acetylcholine receptor ⑀ subunit. Neurology 50:A432, 1998. Ohno, K., Milone, M., Shen, X.-M., and Engel, A. G.: A frameshifting mutation in CHRNE unmasks skipping of the preceding exon. Hum. Mol. Genet. 12:3055, 2003. Ohno, K., Quiram, P., Milone, M., et al.: Congenital myasthenic syndromes due to heteroallelic nonsense/missense

159.

160.

161.

162.

163.

164.

165.

166. 167.

168.

169.

170.

171.

172.

173.

174.

175.

865

mutations in the acetylcholine receptor ⑀ subunit gene: identification and functional characterization of six new mutations. Hum. Mol. Genet. 6:753, 1997. Ohno, K., Sadeh, M., Blatt, I., et al.: E-box mutations in RAPSN promoter region in eight cases with congenital myasthenic syndrome. Hum. Mol. Genet. 12:739, 2003. Ohno, K., Tsujino, A., Brengman, J. M., et al.: Choline acetyltransferase mutations cause myasthenic syndrome associated with episodic apnea in humans. Proc. Natl. Acad. Sci. U.S.A. 98:2017, 2001. Ohno, K., Wang, H.-L., Milone, M., et al.: Congenital myasthenic syndrome caused by decreased agonist binding affinity due to a mutation in the acetylcholine receptor ⑀ subunit. Neuron 17:157, 1996. Ohno, K., Wang, H.-L., Shen, X.-M., et al.: Slow-channel mutations in the center of the M1 transmembrane domain of the acetylcholine receptor ␣ subunit. Neurology 54(Suppl. 3):A183, 2000. Okada, Y., Yamazaki, H., Sekine-Aizawa, Y., and Hirokawa, N.: The neuron-specific kinesin superfamily protein KIF1A is a unique monomeric motor for anterograde axonal transport of synaptic vesicle precursors. Cell 81:769, 1995. Okuda, T., Haga, T., Kanai, Y., et al.: Identification and characterization of the high-affinity choline transporter. Nat. Neurosci. 3:120, 2000. O’Neill, J. H., Murray, N. M. F., and Newsom-Davis, J.: The Lambert-Eaton myasthenic syndrome: a review of 50 cases. Brain 3:577, 1988. Osserman, K. E.: Myasthenia Gravis. New York, Grune & Stratton, 1958. Osserman, K. E., and Genkins, G.: Studies in myasthenia gravis: review of a twenty-year experience in over 1200 patients. Mt. Sinai J. Med. 38:497, 1971. Parsons, K. T., Kwok, W. W., Gaur, L. K., and Nepom, G. T.: Increased frequency of HLA class II alleles DRB1*0301 and DQB!*0201 in Lambert-Eaton myasthenic syndrome without associated cancer. Hum. Immunol. 61:828, 2000. Parsons, S. M., Carpenter, R. S., Koenigsberger, R., and Rothlein, J. E.: Transport in the cholinergic synaptic vesicle. Fed. Proc. 41:2765, 1982. Prockop, D. J., and Kivirikko, K. I.: Collagens: molecular biology, diseases, and potentials for therapy. Annu. Rev. Biochem. 64:403, 1995. Quiram, P., Ohno, K., Milone, M., et al.: Mutation causing congenital myasthenia reveals acetylcholine receptor ␤/␦ subunit interaction essential for assembly. J. Clin. Invest. 104:1403, 1999. Ramarao, M. K., Bianchetta, M. J., Lanken, J., and Cohen, J. B.: Role of rapsyn tetratrichopeptide repeat and coiledcoil domains in self-association and nicotinic acetylcholine receptor clustering. J. Biol. Chem. 276:7475, 2001. Ramarao, M. K., and Cohen, J. B.: Mechanism of nicotinic acetylcholine receptor cluster formation by rapsyn. Proc. Natl. Acad. Sci. U.S.A. 95:4007, 1998. Reimer, R. J., Fon, A. E., and Edwards, R. H.: Vesicular neurotransmitter transport and the presynaptic regulation of quantal size. Curr. Opin. Neurobiol. 8:405, 1998. Robertson, W. C., Chun, R. W. M., and Kornguth, S. E.: Familial infantile myasthenia. Arch. Neurol. 37:117, 1980.

866

Pathology of the Peripheral Nervous System

176. Robitaille, R., Adler, E. M., and Charlton, M. P.: Strategic location of calcium channels at transmitter release sites of frog neuromuscular synapses. Neuron 5:773, 1990. 177. Rondepierre, P., Furby, A., Godefroy, O., et al.: Bloc neuromusculaire mixte pré- et post-synaptique. Rev. Neurol. (Paris) 148:193, 1992. 178. Rooke, E. D., Eaton, L. M., Lambert, E. H., and Hodgson, C. H.: Myasthenia and malignant intrathoracic tumor. Med. Clin. North Am. 44:977, 1960. 179. Rubenstein, A. E., Horowitz, S. H., and Bender, A. N.: Cholinergic dysautonomia and Lambert-Eaton syndrome. Neurology 29:720, 1979. 180. Ruff, R. L.: Sodium channel slow inactivation and the distribution of sodium channels on skeletal muscle fibres enable the performance properties of different skeletal muscle fiber types. Acta Physiol. Scand. 156:159, 1996. 181. Sadeh, M., Blatt, I., and Goldhammer, Y.: Single fiber EMG in a congenital myasthenic syndrome associated with facial malformations. Muscle Nerve 16:177, 1993. 182. Sahashi, K., Engel, A. G., Lambert, E. H., and Howard, F. M. Jr.: Ultrastructural localization of the terminal and lytic ninth complement component (C9) at the motor end-plate in myasthenia gravis. J. Neuropathol. Exp. Neurol. 39:160, 1980. 183. Salpeter, M. M.: Molecular organization of the neuromuscular synapse. In Albuquerque, E. X., and Eldefrawi, A. T. (eds): Myasthenia Gravis. New York, Chapman and Hall, p. 105, 1983. 184. Salpeter, M. M.: Vertebrate neuromuscular junctions: general morphology, molecular organization, and functional consequences. In Salpeter, M. M. (ed.): The Vertebrate Neuromuscular Junction. New York, Alan Liss, p. 1, 1987. 185. Salpeter, M. M., Rogers, A. W., Kasprzak, H., and McHenry, F. A.: Acetylcholinesterase in the fast extraocular muscle of the mouse by light and electron microscopy autoradiography. J. Cell Biol. 78:274, 1978. 186. Satoyoshi, E., Kowa, H., and Fukunaga, N.: Subacute cerebellar degeneration in Eaton-Lambert syndrome with bronchogenic carcinoma. Neurology 23:764, 1973. 187. Schaeffer, L., Duclert, N., Huchet-Dymanus, M., and Changeux, J.-P.: Implication of a multisubunit Ets-related transcription factor in synaptic expression of the nicotinic acetylcholine receptor. EMBO J. 17:3078, 1998. 188. Schmidt, C., Abicht, A., Krampfl, K., et al.: Congenital myasthenic syndrome due to a novel missense mutation in the gene encoding choline acetyltransferase. Neuromuscul. Disord. 13:245, 2003. 189. Schwartz, M. S., and Stålberg, E.: Myasthenic syndrome studied with single fiber electromyography. Arch. Neurol. 32:815, 1975. 190. Seybold, M. E.: Diagnosis of myasthenia gravis. In Engel, A. G. (ed.): Myasthenia Gravis and Myasthenic Disorders. New York, Oxford University Press, p. 167, 1999. 191. Shapira, Y. A., Sadeh, M. E., Bergtraum, M. P., et al.: Three novel COLQ mutations and variation of phenotypic expressivity due to G240X. Neurology 58:603, 2002. 192. Shen, X.-M., Ohno, K., Aams, C., and Engel, A. G.: Slowchannel congenital myasthenic syndrome caused by a novel epsilon subunit mutation in the second AChR transmembrane domain. J. Neurol. Sci. 199(Suppl. 1):S96, 2002.

193. Shen, X.-M., Ohno, K., Fukudome, T., et al.: Congenital myasthenic syndrome caused by low-expressor fast-channel AChR ␦ subunit mutation. Neurology 59:1881, 2002. 194. Shen, X.-M., Ohno, K., Milone, M., et al.: Fast-channel syndrome. Neurology 56(Suppl. 3):A60, 2001. 195. Shen, X.-M., Ohno, K., Tsujino, A., et al.: Mutation causing severe myasthenia reveals functional asymmetry of AChR signature Cys-loops in agonist binding and gating. J. Clin. Invest. 111:497, 2003. 196. Simpson, J. F., Westbery, M. R., and Magee, K. R.: Myasthenia gravis: an analysis of 295 cases. Acta Neurol. Scand. Suppl. 23:1, 1966. 197. Sine, S. M., Claudio, T., and Sigworth, F. J.: Activation of Torpedo acetylcholine receptors expressed in mouse fibroblasts: single-channel current kinetics reveal distinct agonist binding affinities. J. Gen. Physiol. 96:395, 1990. 198. Sine, S. M., Ohno, K., Bouzat, C., et al.: Mutation of the acetylcholine receptor ␣ subunit causes a slow-channel myasthenic syndrome by enhancing agonist binding affinity. Neuron 15:229, 1995. 199. Stanley, E. F., and Drachman, D. B.: Effect of myasthenic immunoglobulin in acetylcholine receptors of intact mammalian neuromuscular junctions. Science 200:1285, 1978. 200. Stiles, J. R., Kovyazina, I. V., Salpeter, E. E., and Salpeter, M. M.: The temperature sensitivity of miniature endplate currents is mostly governed by channel gating: evidence from optimized recordings and Monte Carlo simulations. Biophys. J. 77:1177, 1999. 201. Stiles, J. R., Van Helden, D., Bartol, T. M., et al.: Miniature endplate current rise times ⬍100 ␮s from improved dual recordings can be modeled with passive acetylcholine diffusion from a synaptic vesicle. Proc. Natl. Acad. Sci. U.S.A. 93:5747, 1996. 202. Struyk, A. F., Scoggan, K. A., Bulman, D. E., and Cannon, S. C.: The human muscle Na channel mutation R699H associated with hypokalemic periodic paralysis enhances slow inactivation. J. Neurosci. 20:8610, 2000. 203. Südhof, T. C., and Jahn, R.: Proteins of synaptic vesicles involved in exocytosis and membrane recycling. Neuron 6:665, 1991. 204. Tarelli, F. T., Passafaro, M., Clementi, F., and Sher, E.: Presynaptic localization of omega-conotoxin-sensitive calcium channels at the frog neuromuscular junction. Brain Res. 547:331, 1991. 205. Trontelj, J., and Stålberg, E.: Jitter measurement by axonal micro-stimulation: guidelines and technical notes. Electroencephalogr. Clin. Neurophysiol. 85:30, 1992. 206. Tsujino, A., Maertens, C., Ohno, K., et al.: Myasthenic syndrome caused by mutation of the SCN4A sodium channel. Proc. Natl. Acad. Sci. U.S.A. 100:7377, 2003. 207. Viets, H. R., and Schwab, R. S.: Thymectomy for Myasthenia Gravis. Springfield, IL, Charles C Thomas, 1960. 208. Walls, T. J., Engel, A. G., Nagel, A. S., et al.: Congenital myasthenic syndrome associated with paucity of synaptic vesicles and reduced quantal release. Ann. N. Y. Acad. Sci. 681:461, 1993. 209. Wang, H.-L., Auerbach, A., Bren, N., et al.: Mutation in the M1 domain of the acetylcholine receptor alpha subunit decreases the rate of agonist dissociation. J. Gen. Physiol. 109:757, 1997.

Diseases of the Neuromuscular Junction 210. Wang, H.-L., Milone, M., Ohno, K., et al.: Acetylcholine receptor M3 domain: stereochemical and volume contributions to channel gating. Nat. Neurosci. 2:226, 1999. 211. Wang, H.-L., Ohno, K., Milone, M., et al.: Fundamental gating mechanism of nicotinic receptor channel revealed by mutation causing a congenital myasthenic syndrome. J. Gen. Physiol. 116:449, 2000. 212. Wernig, A.: Estimates of statistical release parameters from crayfish and frog neuromuscular junctions. J. Physiol. (Lond.) 244:207, 1975. 213. Willman, R., and Fuhrer, C.: Neuromuscular synaptogenesis. Cell. Mol. Life Sci. 59:1296, 2002.

867

214. Wilson, K. S. A.: Neurology. London, Butterworth, 1955. 215. Wirtz, P. W., Roep, B. O., Schreuder, G. M., et al.: HLA class I and II in Lambert-Eaton myasthenic syndrome without associated cancer. Hum. Immunol. 62:809, 2001. 216. Wood, S. J., and Slater, C. P.: Safety factor at the neuromuscular junction. Prog. Neurobiol. 64:393, 2001. 217. Zhang, Y., Chen, J., and Auerbach, A.: Activation of recombinant mouse acetylcholine receptors by acetylcholine, carbamylcholine and tetraethylammonium. J. Physiol. (Lond.) 486:189, 1995. 218. Zucker, R. S.: Changes in the statistics of transmitter release during facilitation. J. Physiol. (Lond.) 229:787, 1973.

34 Pathology and Quantitation of Cutaneous Innervation WILLIAM R. KENNEDY, GWEN WENDELSCHAFER-CRABB, MICHAEL POLYDEFKIS, AND JUSTIN C. MCARTHUR

Overview Anatomic Features of the Skin Methods of Cutaneous Nerve Analysis Choice of Biopsy/Blister Location Biopsy Methods Staining Skin Biopsy Specimens Skin Blister Method Epidermal Sheet Preparations Polymerase Chain Reaction Techniques in Skin Microscopy Quantitation of Cutaneous Nerves Epidermal Nerves Normal Epidermal Nerve Density Other Cutaneous Nerves Morphologic Changes in Sensory Neuropathy

Correlation with Tests of Unmyelinated Nerve Function Correlation between Epidermal Nerve and Sural Nerve Utility of Skin Biopsy Contributions of Skin Biopsy to Diagnosis Painful Sensory Neuropathy Diabetic Neuropathy HIV-Associated Sensory Neuropathies Friedreich’s Ataxia Restless Legs Syndrome Familial Dysautonomia (Riley-Day Syndrome, Hereditary Sensory and Autonomic Neuropathy Type III) Congenital Insensitivity to Pain with Anhidrosis (Hereditary Sensory and Autonomic Neuropathy Type IV)

OVERVIEW The validity of skin biopsy/blister for assessing the extent of cutaneous innervation, and especially epidermal nociceptors, has now been established. These techniques provide a reliable and reproducible means of assessing the numbers of C-fiber nociceptors. Easily applicable and robust quantitation of the dermal innervation, sweat glands, hair shafts, and other potentially instructive nerve fiber populations remains to be developed. Skin biopsy/blister methods have opened new opportunities to learn about the response of unmyelinated nerve fibers to a number of disease entities and experimental conditions. The minimally invasive nature of the biopsy, blister, and epidermal sheet techniques makes it possible to study reactions of unmyelinated nerves to experimental conditions directly in human subjects. The methods are useful for diagnosis and

Psoriasis Port-Wine Stains Leprosy Fabry’s Disease Sensory Ganglionopathies Postherpetic Neuralgia CADASIL Chronic Inflammatory Demyelinating Polyneuropathy Postural Tachycardia Syndrome Pediatric Neurologic Disorders Other Conditions Research Uses of Skin Biopsy Human Models of Nerve Regeneration Animal Models Using Skin Biopsy

may find a place in the longitudinal evaluation of disease progression. An area of particular interest is the potential value of serial skin biopsies in clinical trials of agents that are intended to promote regeneration of nerve fibers. Evaluation of unmyelinated cutaneous nerves obtained by skin biopsy has emerged as a useful method to diagnose and study peripheral nerve disorders.27,46,48,52,63 Until recently, work with peripheral nerves was restricted to studies of myelinated nerve fibers because methods available for assessment of small, unmyelinated fibers were limited. With standard electrodiagnostic studies, unmyelinated fibers remain “invisible” because nerve conduction studies assess only large sensory and motor fibers. Even assessment by sural nerve biopsy is problematic; electron microscopy is required to visualize unmyelinated fibers and nerve biopsy only gives a “window” into one location along the nerve at a single time point. Finally, from a functional standpoint, 869

870

Pathology of the Peripheral Nervous System

cutaneous sensation is transduced by identifiable nerve fibers that enter their “targets” in the skin; these cannot be identified by nerve biopsies. These limitations and the observations that individuals with sensory neuropathies have spontaneous acral neuropathic pain with marked allodynia stimulated development of methods to identify and quantify unmyelinated nerves in the skin. Early studies of cutaneous nerves obtained by punch skin biopsy defined the density and distribution of Meissner’s corpuscles and their myelinated nerves in different age groups and in several hereditary disorders. Unmyelinated nerves were not visualized.7,20 The development of sensitive immunohistochemical techniques offered a new approach to identify and quantify small sensory nerves. In particular, immunostaining of skin biopsies for protein gene product 9.5 (PGP 9.5), a neuronal ubiquitin carboxy-terminal hydrolase,106 has been used by a number of groups to visualize the subpapillary plexus of small myelinated and unmyelinated nerve fibers. Epidermal nerve fibers have received the greatest scrutiny, mainly because they appear to be early indicators of neuropathy and adequate samples can easily be obtained for quantitation.18,45,63,110 These fibers include both A and C fibers that convey “slow” nociception. We developed robust normative data,61 and have demonstrated a usually distally dominant pattern of epidermal nerve fiber loss in diabetic neuropathy,47 human immunodeficiency virus (HIV)–associated sensory neuropathies,82 and idiopathic small fiber sensory neuropathies.35 In patients with severe neuropathy, decreased epidermal densities are most severe distally, with less marked reductions found at progressively more rostral levels and prominent predegenerative axonal swellings identified even at asymptomatic sites. In fact, abnormalities of cutaneous innervation are found in some individuals with normal tendon reflexes at the ankles, normal sural nerve action potential amplitudes, and normal quantitative sensation tests (QSTs). Epidermal denervation also occurs in individuals with spontaneous allodynic pain.34,78 While the precise structural correlates of allodynia remain uncertain, it can clearly occur even with marked depletion of A and C fibers in the skin. The association of neuropathic pain with epidermal nerve loss is, en face, paradoxical because nerve fiber loss is traditionally related to loss of sensation without pain. Neuropathic pain has been correlated with epidermal nerve loss in small fiber sensory neuropathy, postherpetic neuralgia, HIV-associated sensory neuropathies, and diabetic neuropathy. Explanations for this apparent paradox include possible changes in both peripheral and central nervous systems. In peripheral neuropathy caused by persistent action of an etiologic agent, continuous spontaneous pain can result from ectopic discharges in peripheral nociceptive fibers.77 Alternatively, sensitized nociceptors responding with reduced threshold to weak stimuli that normally evoke touch, warm, or cool sensations also evoke

pain (hyperalgesia). Capsaicin-induced mechanical and heat hyperalgesia is one example. Another is the neuronal excitability that results from increased voltage-gated Na currents after exposure to serotonin, prostaglandin E2, or adenosine, agents that produce tenderness or hyperalgesia.26 Likewise, products of nerve degeneration might increase nerve excitation by unmasking proteins normally hidden from immune surveillance (e.g., P0, P2). Resultant immune cell activation exposes surviving nerves to inflammatory cytokines and other substances that alter threshold to stimuli.111 There is growing evidence that injury to nociceptors leads to increased firing not only in injured axons but also in uninjured neighboring axons115 and perhaps in dorsal root ganglia neurons at an adjacent level.4 In fact, selective injury of a ventral root causes hypersensitivity of C fiber afferents in an adjacent root.116 Subsequent activation of nociceptors is believed to produce central sensitization and secondary hyperalgesia. One mechanism is by central terminals of injured A fibers that develop nerve sprouts from deep dorsal horn lamina into superficial laminae. If these synapse with nociceptive neurons, the neurons could become susceptible to activation by otherwise innocuous peripheral stimuli.114 In addition, a shift of descending modulation from inhibition to facilitation can contribute to hypersensitivity.84

ANATOMIC FEATURES OF THE SKIN Skin separates the external environment from internal tissue while providing sensory input, protection from pathogens and allergens, and a water barrier. The layered structure of skin imparts characteristics essential to these functions (Fig. 34–1). The distinct dermal and epidermal layers of skin comprise the major compartments of cutaneous structure. The epidermis is the topmost living layer of skin and consists of layers of keratinocytes that are replaced in approximately a 28-day cycle in humans as they are desquamated to form the cornified surface layer of dead cells that provides the water barrier. Intermixed among the vital keratinocytes are basal melanocytes, dendritic Langerhans cells, and sensory nerve terminals. The epidermis is separated from the dermis by the dermalepidermal basement membrane. The area of dermis immediately internal to the epidermis, the papillary dermis, contains a vascular plexus with periodic capillary loops that reside within dermal protrusions into the epidermis called “dermal papillae.” Within the matrix of the dermis, cutaneous structures include hair follicles with associated arrector pilorum muscles and sebaceous glands, blood vessels, nerve bundles, and sweat glands (Fig. 34–1). The nerves to skin arise from sensory and motor neurons residing in dorsal root ganglia and sympathetic ganglia. Bundles of nerves enter the skin deep in the dermis and course toward the skin surface, giving off axons to

871

Pathology and Quantitation of Cutaneous Innervation

Normal Skin

FIGURE 34–2 Normal human epidermal and papillary dermis innervation. Nerves are localized with antibody to PGP 9.5, and basement membrane is demarcated with antibody to type IV collagen. Vasculature is labeled with Ulex europaeus agglutinin type I. Epidermal nerve fibers appear aqua and lie within the blue epidermis (E). The subepidermal neural plexus (SNP) appears green or yellow. The dermal-epidermal junction appears red. Capillaries (C) appear magenta. Nerve fibers (green and aqua) course in bundles through the dermis and branch in the papillary dermis to form the subepidermal neural plexus. Fibers arise from this plexus and penetrate the epidermal-dermal basement membrane to enter the epidermis. Note that some non-neuronal fibroblasts appear green. Epidermal nerve fibers are abundant and uniformly distributed in normal human skin. See Color Plate

FIGURE 34–1 Human skin innervation and vasculature. Confocal image of 150-m-thick section of human skin immunostained to demonstrate nerves (PGP 9.5—green and yellow) and basement membrane (type IV collagen—red). Image is a projection of 31 5-m optical sections acquired with a 10  objective lens. Basement membrane outlines cutaneous structure, including the basement membrane separating the dermis from the epidermis (Ep), capillary (Cap) loops of the papillary dermis and vasculature throughout the dermis, the central hair follicle (HF) and associated sebaceous gland, arrector pili muscles (AP), and sweat glands (SG). Nerve bundles (NB) arise from the deep dermis to innervate cutaneous structures. Characteristic nerve networks associated with arteries (Ar), sweat glands, arrector pili muscles, the hair follicle, and the subepidermal neural plexus provide sufficient morphologic criteria for identification of these structures based on their innervation patterns. See Color Plate

innervate the associated end organs. Unmyelinated nerve fibers comprise the vast majority of cutaneous innervation to the above dermal structures. The few myelinated nerve fibers terminate at hair follicles, Meissner’s corpus-

cles, and Merkel complexes. When the vertically oriented nerve bundles enter the superficial papillary dermis, they form a horizontal subepidermal neural plexus (Fig. 34–2). Epidermal nerve fibers branch from this plexus and, while penetrating the dermal-epidermal basement membrane to enter the epidermis, they lose their Schwann cell ensheathment and collagen collar.12 Within the epidermis they extend between keratinocytes to the surface of the vital epidermis. Autonomic nerves to sweat glands enwrap the coiled sweat tubules in such a dense pattern that they virtually cover much of the surface (see Fig. 34–1, SG). The autonomic innervation of the arrector pilorum muscles is easily distinguished within the muscle by a characteristic wavy pattern. The complex innervation pattern of hair follicles is composed of both myelinated and unmyelinated fibers, with specialized nerve endings at the base and along the shaft of the hair.22,28,85 Fine unmyelinated nerve endings form a meshlike network covering larger arteries in the deep dermis, but only one or two nerves are on the more superficial smaller arterioles (see Fig. 34–1, Ar). Epidermal nerve fibers are constantly remodeling in concert with the migration and shedding of keratinocytes during the turnover of epidermal layers. In addition,

872

Pathology of the Peripheral Nervous System

epidermal nerve fibers respond to injury by regenerative regrowth and collateral sprouting. The neurotrophic requirements of epidermal nerve fibers have been studied, and at least three trophic factors have been found to affect C-fiber nociceptors: nerve growth factor (NGF), glial cell line–derived neurotrophic factor (GDNF), and insulin-like growth factor type 1. The neurons that depend on NGF, approximately half the small sensory neurons in rat dorsal root ganglia, express the high-affinity NGF receptor TrkA as well as the low-affinity receptor p75. The other half are predominantly GDNF-responsive neurons that express the receptor c-Ret65 as well as a series of other markers. These markers include the ability to bind a specific Griffonia lectin for fucosyl moieties, isolectin B4, which provides a convenient, but not completely specific, marker for the GDNF-dependent population. Finally, most nociceptors bear the receptor for the chili pepper toxin capsaicin. This receptor, vanilloid receptor type 1 (VR1), is the basis for the burning pain elicited by topical application of capsaicin. It normally responds to heat and to protons; mice genetically engineered to lack VR1 fail to respond to heat and acid.10 A related receptor, vanilloid receptor–like receptor type 1, has recently been identified, primarily on somewhat larger sensory neurons probably giving rise to A fibers. Vanilloid receptor–like receptor type 1 responds to very high temperatures (52° C).11 In the human, almost all of the epidermal axons are VR1 positive.82 Epidermal fibers containing the peptides calcitonin gene–related peptide (CGRP) and substance P (SP) are NGF-dependent axons. In some animals these commonly enter the epidermis, but in human skin they usually terminate near capillary loops in the dermal papilla. Attempts to immunostain for other peptides in human nerves, including the isolectin B4, have been unsuccessful.

reduced number of epidermal nerves or may be devoid of epidermal nerve fibers altogether. The foot has the advantage of being more distal, but its nerves are subject to trauma. Supplementary biopsies at more proximal levels (e.g., calf, thigh, or upper extremity) can yield additional information about the severity and distribution of small fiber loss (Fig. 34–3). If the plan is to follow the course of neuropathy or the response to therapy, it is best to select a biopsy location with a definite reduction of epidermal nerves, but with a sufficient nerve residual in the epidermis and subepidermal neural plexus to either provide a substrate for nerve regeneration or allow documentation of further degeneration. This can usually be accomplished by removing biopsies from the upper sensory level, if present, and from one or two more proximal locations, keeping in mind that the loss of epidermal nerves is generally more severe than would be inferred from the clinical symptoms or findings. Future biopsies for comparison should be performed horizontally adjacent to the original biopsy at least 5 to 10 mm away from scarring and within the same peripheral nerve distribution. Normal epidermal nerve fiber values have been determined for proximal and distal anterolateral thigh, proximal and distal posterolateral calf, dorsum of the foot over the extensor digitorum brevis muscle, over the first dorsum interosseus muscle, proximal volar forearm, and the paraspinal area at T4-T5. The most frequently biopsied locations selected to assess a suspected length-dependent neuropathy are thigh and calf. The flexibility of the technique allows biopsy or blister to be performed safely almost anywhere on the body. Thus a localized sensory complaint (e.g., a thoracic radiculopathy) can be examined by biopsy of the symptomatic site and a mirror-image control, with postherpetic neuralgia being a possible exception because there may be contralateral changes in the unaffected side.76

METHODS OF CUTANEOUS NERVE ANALYSIS

Biopsy Methods

Choice of Biopsy/Blister Location The minimal invasiveness and negligible scarring remove most cosmetic objections to skin biopsy and skin blister procedures. When choosing a biopsy site, the susceptibility to nerve loss with neuropathy and the availability of normative data should be considered. Areas subject to trauma or with scars should be avoided. Site selection also depends upon the purpose for which the biopsy/blister is being considered (e.g., for diagnosis or for longitudinal study). Diagnosis of small fiber pathology can be made from a distal skin location where the clinical examination shows decreased sensation, particularly decreased sensitivity to mechanical (pin) and thermal (hot) pain stimuli. In patients with neuropathy, biopsies from the dorsum of the foot, distal calf, and proximal calf frequently contain a

Skin biopsy is commonly performed with a 3-mm diameter punch instrument (Fig. 34–4) (Acupunch; Acuderm, Inc., Ft. Lauderdale, FL) with local anesthetic. The wound rarely requires cautery or suture, and the infection rate in the lower extremity is less than 1:200. A shallow biopsy is adequate if interest is limited to epidermal nerves. A deeper biopsy (3 to 4 mm) is necessary to acquire sweat glands, the full hair follicle, subcutaneous fat, and larger arterioles. The biopsy is immediately placed into 4° C fixative and stored overnight, then transferred to 20% sucrose in phosphatebuffered saline for cryoprotection and storage. The biopsy can be held in sucrose in phosphate-buffered saline at 4° C for up to 1 month or frozen for more prolonged storage. Formaldehyde-based fixatives provide the best tissue and antigen preservation. Glutaraldehyde destroys the antigens and should not be used. Zamboni (2% paraformaldehyde and picric acid), Lana (4% formaldehyde and picric acid),

Pathology and Quantitation of Cutaneous Innervation

873

FIGURE 34–3 Cutaneous innervation of normal and diabetic subjects. Confocal microscope images of sectioned superficial skin of a normal (A and B) and a diabetic (C and D) subject. Epidermal nerve fibers arise from a subepidermal neural plexus (SNP) that extends horizontally immediately below the dermal-epidermal basement membrane. Numerous nerve fibers cross the basement membrane to innervate the epidermis of a control subject (A, thigh; B, calf). Diabetic subjects display a range of epidermal nerve fiber density depending on the body location and the degree of neuropathy. Skin biopsies from a diabetic subject have normal density of innervation in the thigh (C); however, several abnormalities are observed. Swellings in the nerve fibers are common in the SNP, and some nerve fibers are longer and more branched. The bulbous nerve endings are not seen at other depths of the dermis. The calf biopsy of the same subject contains no epidermal nerve fibers (D). A single nerve fiber in the SNP branches at the dermal-epidermal junction but does not cross into the epidermis. (From Kennedy, W. R., Wendelschafer-Crabb, G., and Walk, D.: Use of skin biopsy and skin blister in neurological practice. J. Clin. Neuromuscul. Dis. 1:196, 2000, with permission.) See Color Plate

and PLP (paraformaldehyde, lysine, and periodate) fixatives all provide optimal preservation. If these fixatives are not available, 10% formalin preservation for 12 to 18 hours is an alternative, but produces a more fragmented appearance of the epidermal nerves and is suboptimal. For multicenter studies or trials, in which the biopsy will be performed at a location geographically removed from the processing laboratory, the specimen should be placed into fresh cold fixative for 12 to 24 hours, then transferred into cryoprotectant for overnight shipping on wet ice.

Staining Skin Biopsy Specimens Nerve evaluation from skin biopsy relies on immunohistochemical localization of neural antigens within skin tissue sections. This is accomplished by applying a primary antibody directed against the antigen of interest followed by a labeled secondary antibody directed to the primary antibody. The major primary antibody used for the localization

of nerves in skin biopsy recognizes the pan-neuronal marker PGP 9.5 (Ultraclone, Wellow, UK; Chemicon, Temecula, CA) that is present in all cutaneous nerves. Localization of additional antigens can aid in diagnosis but may not be necessary in all situations. Several antigens can be localized within a single tissue section by using secondary antibodies that recognize and react with the immunoglobulin G (IgG) of the primary antibodies. Typically rabbit, mouse, and goat primary antibodies are localized with donkey antiserum specifically prepared to recognize IgG from only a single species (rabbit, mouse, or goat). Complementary markers such as fluorophores or enzymes label each of the speciesspecific secondary antibodies. Each fluorophore’s emission wavelength peak must be distinct so that it can be viewed individually with appropriate light filters (Fig. 34–5). Alternatively, enzymes such as peroxidase or alkaline phosphatase are reacted with a variety of available substrates to provide reaction products of different colors. All secondary antibodies should be derived from the same species

874

Pathology of the Peripheral Nervous System

Cy5

AMCA Cy2, FITC Cy3

400

500

600

700

Wavelength (nm)

FIGURE 34–4 Punch biopsy tool. A 3-mm punch biopsy is made on clean, shaved skin that has been swabbed with povidone-iodine (Betadine) and numbed with local anesthetic. The biopsy tool is inserted with a twisting motion to approximately half the depth of the metal collar. The depth of the biopsy depends on the thickness of the skin and the sample content desired. Sweat glands lie deep in the dermis and require a full-thickness biopsy, while shallow biopsies are sufficient for analysis of epidermal nerve density.

(e.g., donkey) and must be from a species different than hosts of the primary antibodies. Secondary antibodies used for multiple staining must be preadsorbed with IgG from all species other than the primary antibody being localized. Frozen 50- to 100-m-thick sections are cut perpendicular to the epidermal surface. The thicker sections provide greater sampling and visualization of anisotropic nerve fibers as they weave through the skin. Sections are floated in staining reagents containing Triton X-100 to promote penetration of antibodies. Samples are washed and antibodies are diluted in a solution of phosphate-buffered saline, 1% normal donkey serum, and 0.3% Triton X-100. Sections are incubated in 5% normal donkey serum so that nonspecific binding of IgG will be “blocked.” Sections are incubated in primary antibodies (e.g., rabbit anti–PGP 9.5 and mouse anti–type IV collagen; Chemicon) with gentle rotation for 5 hours at room temperature and overnight at 4° C, then washed in three changes of wash buffer, 1 hour each, with gentle rotation. This is followed with overnight incubation in labeled secondary antibody to the rabbit and mouse primary antibodies (Jackson ImmunoResearch, West Grove, PA), then washed three times. The secondary antibodies are labeled with either peroxidase or a stable fluorescent marker. Peroxidase markers are reacted enzymatically to produce a precipitate at the site of the antigen. The peroxidase reaction product is visible by bright-field microscopy and by electron microscopy. With the perox-

FIGURE 34–5 Fluorophore emission wavelength profile. Fluorophores with distinct wavelengths that can be isolated visually by use of optical filters are used to label antibodies for multiantigen immunofluorescent staining applications. Nerves are usually labeled with Cy 3 (fluorescence peaking at 565 nm), while basement membrane is labeled with Cy 2, fluorescing at 520 nm. By applying a blocking filter between 540 and 550 nm, the two antigens can be viewed independently; while applying a dual-band filter, the two antigens can be viewed simultaneously and distinguished microscopically by their colors. (Courtesy of Jackson ImmunoResearch, West Grove, PA.)

idase reaction, the basement membrane is localized as being just below the basal membrane of basal keratinocytes. Fluorescent markers require no further reaction before being studied by fluorescence or confocal microscopy. A new generation of fluorescent markers, Cy dyes 2, 3, and 5 (Jackson ImmunoResearch), are very bright and remain stable for several years in permanent mountants. The markers are useful for labeling two or three antigens within the same section. Labeling nerve with Cy 3, basement membrane with Cy 2, and endothelial cells with Ulex europaeus agglutinin type I with Cy 5 provides a vivid image of skin structure, innervation, and vasculature (Fig. 34–6). Non–immune serum specificity controls are essential and should be run with each biopsy. Accurate counts of epidermal nerve fibers, interpretation, and quantitation are difficult when sections fold or curl during drying of the epidermis. For optimal viewing and quantification, the epidermis should lie flat on the slide. Premounting sections in a small drop of Nobel agar prevents drying, adheres the sections, and holds the surface of the stratum corneum perpendicular to the slide. The premounted section is dehydrated in alcohol, cleared with methyl salicylate, and mounted on a coverslip in DPX (Fluka, Buchs, Switzerland). This provides a permanent preparation and is useful for slides labeled with Cy fluorophores and some enzyme reactions. Glycerol can be used to coverslip tissue without dehydrating the specimens and is useful for short-term storage.

Pathology and Quantitation of Cutaneous Innervation

875

FIGURE 34–6 Staining nerve and basement membrane. A, Nerves are immunostained with rabbit anti–PGP 9.5 and Cy 3. B, Basement membrane is immunostained with mouse anti–type IV collagen and Cy 2. C, Epidermis and vessels are stained with the lectin Ulex europaeus agglutinin type I and Cy 5. D, All images are combined in this colorized representation of the localization of nerves, vessels, dermal-epidermal basement membrane, and epidermis. Confocal images are acquired as gray-scale image stacks for each antigen-fluorophore combination. These are projected, pseudo colored, and combined to create a color image of the entire z series. See Color Plate

Skin Blister Method Skin blister is a painless, bloodless method to acquire, quantify, and plot distribution of epidermal nerve fibers.43 The blister roof, which consists of pure epidermis and epidermal nerve fibers, separates from the dermis on a plane between the basal membrane of basal keratinocytes and the dermal-epidermal basement membrane. Blisters are made by applying suction to a blister capsule (WR Medical Electronics, Stillwater, MN) that is adhered to shaved, cleansed skin with double-sided tape (Fig. 34–7). The surface of the capsule resting on skin contains one or more round openings to accommodate the desired number and size of blisters. Tegaderm tape (3M, Maplewood, MN) disks, cut slightly smaller than the size of the desired blister (e.g., 3-mm diameter), are placed on the skin surface to conform to the openings in the blister capsule. These prevent overstretch of the blister roof. The capsule is secured to skin with an elastic bandage, leaving the transparent top of the capsule uncovered to permit the examiner and subject to observe blister formation, and the capsule is evacuated to a negative pressure of 300 mm Hg. The small volume of the capsule makes it imperative to have tight seals to prevent leakage. A reservoir placed in series with the tubing between the pump and capsule enhances blister formation by equilibrating small leaks. We use a commercially available mechanical pump to evacuate the capsule,

but an inexpensive hand-held pump is adequate when a single capsule is used. Initially small blisters form, then these coalesce to form a full blister after 20 to 40 minutes. The time is less with increasing subject age and skin temperature, but also depends upon skin location and successful maintenance of negative pressure. Warming the area

FIGURE 34–7 A blister formed by application of negative pressure in the blister capsule. The blister conforms to the 3-mm opening in the base of the capsule. (From Kennedy, W. R., Nolano, M., WendelschaferCrabb, G., et al.: A skin blister method to study epidermal nerves in peripheral nerve disease. Muscle Nerve 22:360, 1999, with permission.)

876

Pathology of the Peripheral Nervous System

with a heating pad hastens blistering. After the capsule is removed, the blister roof with tape disk intact is excised with microscissors and placed in cold Zamboni fixative overnight at 4° C. Fixed blisters are cryoprotected with 20% sucrose in 0.1 M phosphate-buffered saline until processed. Prior to processing, blister roofs are frozen on dry ice. Specimens are processed as whole mounts without sectioning; otherwise the procedure is identical to that for thick sections.

Epidermal Sheet Preparations This is a modification of the standard punch technique that allows for separation of the epidermis from the dermis before fixation.5 The technique is particularly useful for examining the epidermis in the horizontal plane without dermal vessels or nerves. In addition, the epidermal sheet can be used for protein or messenger RNA (mRNA) assays, as described below, rather than immunocytochemical staining. Briefly, after obtaining the punch biopsy, the specimen is bathed in ethylenediaminetetraacetic acid for 2 hours; then the epidermis can be easily removed from the dermis, fixed, and immunostained as above. This preparation permits the identification of the nerve fiber terminals within the epidermis without blood vessels or the dermal structures. The sheet can be stained in exactly the same way to identify PGP-immunoreactive nerve fibers.

Polymerase Chain Reaction Techniques in Skin It is feasible to quantify neurotransmitter or neurotrophin mRNA levels in either the epidermis (using the sheet preparation) or the dermis (using the standard punch biopsy). In the field of HIV/acquired immunodeficiency syndrome (AIDS), there is interest in the concept that specific antiretrovirals may affect mitochondrial DNA levels by inhibition of gamma DNA polymerase.13 Subcutaneous fat is a rich source of mitochondria. We use the subcutaneous fat, which is usually trimmed and discarded from punch skin biopsies, to measure levels of mitochondrial DNA with real-time polymerase chain reaction23 in HIV-seropositive patients receiving antiretrovirals. Specific antiretrovirals, the dideoxynucleosides, reduce mitochondrial DNA levels by inhibition of gamma DNA polymerase,60 and thus this technique may provide a convenient measure of the severity of dideoxynucleoside toxicity and the metabolic complications common in individuals treated with these drugs long term.14

MICROSCOPY Thick sections are advantageous for viewing the full meandering course of nerves from the nerve trunk in the dermis throughout their ramifications to the endings in

sweat glands, arrector pili, hair follicles, arterioles, or epidermis. Sections are first scanned under a 10  objective in the microscope to evaluate the orientation and flatness of the section and staining of nerves, blood vessels, and cutaneous organs. The course and ending of single unmyelinated nerve fibers are better visualized with a 20  (or higher) objective, as, for example, when following epidermal nerve fibers through basement membrane and the pathway between keratinocytes to their endings under stratum corneum. Images taken from thick sections are often marred by out-of-focus blur because the focal plane of a 20  objective is about 2 m. Objects in 50- to 100-m sections that are above and below this focal plane are out of focus. Some laboratories use thinner sections to obtain sharp images by conventional microscopy. This is counterbalanced by the inability to follow nerves for more than a short distance in a section and reduced sampling. Confocal microscopy of thick sections is one solution to this problem, although the added complexity may not always be necessary for simple clinical questions. The confocal microscope takes in-focus images while retaining the advantages of working with thick sections. This is accomplished by “optically sectioning” the fluorescentlabeled thick sections into a series of smaller increments that correspond to the focal length of the lens being used. A confocal z series comprising multiple images taken at incremental focal planes throughout the tissue eliminates out-of-focus blur and provides an in-focus three-dimensional view of the tissue (Fig. 34–8). The operator can select the top and bottom of the section to be imaged, the size of the z-focus increments, and the fluorophores to image before collecting the z series. A laser scanning confocal microscope requires about 30 minutes to scan double-stained sections (two antibodies) at 2-m increments through 50 m of tissue. The nonlaser spinning disk confocal system (e.g., CARV) can collect equivalent images in about 3 minutes. The z series is projected into a single in-focus image for viewing. The series can also be “paged through” to follow nerves throughout the tissue, or the series can be rendered into a three-dimensional object.

QUANTITATION OF CUTANEOUS NERVES Epidermal Nerves Standard punch biopsies removed from anatomic locations that are often sampled to diagnose neuropathy contain an adequate sample of unmyelinated epidermal nerves and subepidermal neural plexus for detecting neuropathic changes. Epidermal innervation can be readily quantified using either confocal microscopy46,78 or direct counting of chromogen-stained specimens.16,32 The diagnosis of neuropathy is dependent upon finding

Pathology and Quantitation of Cutaneous Innervation

877

FIGURE 34–8 Confocal versus nonconfocal microscopy. The confocal microscope uses focused light beams (laser or mercury vapor) to optically section thick tissue sections, much like a computed tomography scanner. A series of many images are acquired at focal increments specific to the objective lens’ depth of field. Out-of-focus blur is eliminated, and clear, in-focus images of sections up to 200 m thick can be collected and used for three-dimensional analysis of nerve morphology and quantification.

a reduced density of epidermal nerves. Abnormalities of nerve morphology provide valuable secondary help. Epidermal nerve fibers can be quantified accurately because, after they leave the nerve bundles of the subepidermal nerve plexus and before they enter the epidermis, they separate as single nerve fibers. In thick sections the entire intraepidermal segment of most epidermal nerve fibers can be followed rather than isolated nerve segments, as is necessary in thinner sections. A set of almost identical counting rules evolved simultaneously in laboratories at both the University of Minnesota and Johns Hopkins University (Fig. 34–9): A. Count each nerve as it crosses the basement membrane of the epidermis. Because epidermal nerve fibers often branch below the basement membrane as well as in the epidermis, it is necessary to establish a standard counting site. Most investigators have agreed in practice to count as one unit each nerve that penetrates the basement membrane, even if there is branching within the epidermis. In peroxidase-stained sections, the membrane and cytoplasm of basal keratinocytes stain darker than surrounding tissue. Because the basal

FIGURE 34–9 Nerve counting rules. A set of concise rules was developed to normalize counting among investigators.

membrane is only a few micrometers above the dermalepidermal basement membrane, it can be used as the site for counting penetrating epidermal nerve fibers. In sections that are double-stained for PGP 9.5 in nerve and type IV or type VII collagen for basement membrane and labeled with compounds that fluoresce at different wavelengths, it is possible to directly observe the epidermal nerve fibers as they penetrate basement membrane. If confocal images of thick sections are available, the optical section in which the epidermal nerve fiber penetrates the basement membrane is located by paging through the z series of optical sections. B. Nerves that branch after crossing the basement membrane are counted as one. Epidermal nerve endings in normal subjects are usually quite simple—often free of branches or with minimal branching as they course through the epidermis. Nerves of patients with neuropathy frequently appear to have increased fiber

878

C.

D. E.

F.

G.

Pathology of the Peripheral Nervous System

branching. To address both nerve loss and collateralization of surviving axons, the number of fibers crossing the basement membrane is used to provide counts (how many nerves) while analysis of tracing data provides information about the complexity (length and branching pattern) of the nerve (how much nerve). Nerves that split below the basement membrane are counted as two units. Single epidermal nerve fibers arise as branches from nerve bundles of the subepidermal neural plexus. Each branch is counted as it penetrates the basement membrane. Nerves that appear to branch at the basement membrane are counted as two or more units. Nerve fragments that cross the basement membrane are counted. Epidermal nerve fibers that appear short are counted. They may be short fibers or the roots of fragments that terminate in adjacent sections. Nerve fibers that approach but do not cross the basement membrane are not counted. Nerve fibers that terminate immediately proximal to the basement membrane are common in neuropathy. They frequently follow the basement membrane a short distance before ending with a terminal swelling. Epidermal nerve fragments that do not cross the basement membrane are not counted. This rule has been followed in all publications by the Minnesota group. In previous publications by the Johns Hopkins group, such epidermal nerve fragment were included in the total numbers, but they will not be included in future counts.

Tracings made of epidermal nerve fibers with a software program (Neurolucida; MicroBrightField, Williston, VT) provide the number of nerve fibers, branch points, nerve fiber length, and coordinates of basement membrane penetration (Fig. 34–10). This information is useful because in some patients nerve branching and length increase in apparent compensation for a decrease in nerve number. The data are also useful for analysis of nerve distribution.

Results of epidermal nerve fiber density are commonly expressed as number of epidermal nerve fibers per millimeter of epidermis, providing a “linear density.” Comparison of such values from different reporting laboratories requires that the thickness of the section be known because thicker sections contain more nerve fibers. Results can also be reported as epidermal nerve fibers per square millimeter area of skin surface for easier comparison.32,46 Measurement accuracy is potentially adversely influenced by compression of the section between the slide and coverslip and by tissue shrinkage during preparation. For accurate comparison, these factors must be taken into consideration. Our laboratories report epidermal nerve fiber density based on nerve counts made in the equivalent of 50-m-thick sectioned tissue, making adjustments for changes resulting from dehydration and coverslip compression. Manual counting should be performed on at least three to four nonadjacent sections taken from different places within the biopsy. After manually counting the number of individual epidermal nerve fibers, the length of the epidermis is measured using Bioquant V (R&M Biometrics, Nashville, TN). Counting rules should be established a priori for each laboratory, and preferably only one observer used for research studies involving serial specimens. We have compared manual counting with light microscopy to stereologic techniques and demonstrated that the manual counting produced comparable densities but was much quicker.101 Reliability is highest when at least four sections at each anatomic site are counted. A technician can be trained to perform the manual quantitation, with excellent inter- and intraobserver reliability (0.95).61 We have also demonstrated that different laboratories can achieve reasonable interlaboratory reliability.98 For example, mean ( standard error of the mean) interobserver reliability was 12.2%  1.1% for each biopsy site and mean intraobserver variability was 10.2%  1.5% for individual sections. The variability between laboratories was 25.5% and the correlation coefficient for both intraobserver and

FIGURE 34–10 Nerve quantification by tracing. Neurolucida software (MicroBrightField, Williston, VT) uses the confocal image series to identify and trace each nerve fiber from the point where it crosses the basement membrane to its termination within the epidermis. Nerve number, length, complexity, and density as well as epidermal length and volume can be determined from the acquired data.

Pathology and Quantitation of Cutaneous Innervation

interobserver reliability was .98. The correlation coefficient between sites was .94. There was no relationship between absolute intraepidermal nerve fiber counts and reliability.

Normal Epidermal Nerve Density The density of epidermal nerves varies depending upon the biopsy site.40 In general, a consistent gradient in intraepidermal nerve fiber density exists from proximal to distal sites in the lower extremities, with minimal effects of race or sex. Significantly higher intraepidermal nerve fiber densities were noted in a younger group, age 10 to 19 years, but over the age of 20 years there is little change up to 80 years of age (Fig. 34–11).15,78 Stereologic quantitation has been compared with linear density estimation and is significantly associated with a correlation coefficient of .79 (P  .001). With formalin fixation (which tends to underestimate by about 20% compared to PLP fixation), the normative range for intraepidermal nerve fiber density (including intraepidermal fragments) in healthy normal controls is 21.1  10.4 fibers/mm (mean and standard deviation) in the thigh, with a fifth percentile of 5.2 fibers/mm. At the distal leg the normative range is 13.8  6.7, with a fifth percentile reading of 3.8 fibers/mm. Using the cutoff values for the fifth percentile for intraepidermal nerve fiber density at the distal part of the leg, the diagnostic efficiency (percentage correctly classified) was 88%, with a specificity (percentage true negative) of 97% and a sensitivity (percentage true positive) of 45%. Positive predictive value was 92% and negative predictive value was 90%. Although there is a relatively wide biologic variation in the intraepidermal nerve fiber density, the distribution of densities at the distal part of the leg permits the delineation of lower limit of normal

879

based on the fifth percentile value for healthy controls. The high specificity of the measure increases the positive predictive value of the test, which is therefore ideal for its anticipated use, namely to verify the presence of a disease for which there may be little clinical or electrophysiologic evidence, or when the clinician must be virtually certain of a diagnosis, for example, before initiating some form of therapy. The relatively low sensitivity of the measure implies that a normal intraepidermal nerve fiber density does not rule out the presence of a sensory neuropathy. Normative values of epidermal nerve fiber density for six body locations have been determined at the Kennedy laboratory (Table 34–1). Epidermal Nerves in Skin Blisters and Epidermal Sheets Whole-mounted blister and epidermal sheet preparations have sampling advantages over sectioned skin biopsies. The “bird’s-eye” perspective of epidermal nerve fibers in these 3-mm diameter (area  7.1 mm2) preparations presents for visualization and analysis the same number of nerves contained in the combined sections of a 3-mm biopsy (Fig. 34–12). Using a 20  objective lens, we count the epidermal nerve fibers in five representative segments of the blister roof, each measuring 0.4  0.3 mm. The 0.6-mm2 skin surface area is approximately 12% of the whole blister roof. The results, expressed as counts of epidermal nerve fibers per surface area of skin, vary according to body location in the same manner as that observed with biopsies. Epidermal nerve fiber density determinations from biopsies and blisters show a high degree of correlation (r  .71; P .005).49 The blister roof also provides an advantageous oversight of the horizontal territory of individual epidermal nerve fibers and the distribution of all epidermal nerve fibers within the specimen. Changes in epidermal nerve distribution caused by reorientation of nerves in their course through the epidermis by elongation, branching, or loss of branching probably occur normally during the continual regeneration of the epidermis, perhaps similar to normal reshaping of the terminals of motor axons.57 We suspect that exaggerated redistribution of epidermal nerves may be the earliest reaction of epidermal nerves to neuropathy (L. Waller, personal observations, 1998).

Other Cutaneous Nerves FIGURE 34–11 Effects of age on intraepidermal nerve fiber density at the distal part of the leg by age decile for healthy controls. (From McArthur, J. C., Stocks, E. A., Hauer, P., et al.: Epidermal nerve fiber density: normative reference range and diagnostic efficiency. Arch. Neurol. 55:1513, 1998, with permission.)

In general, dermal nerves are more difficult to quantify than epidermal nerve fibers. The subepidermal neural plexus that occupies the papillary dermis is often less dense in subjects with neuropathy. These bundled nerve fibers cannot be accurately counted or traced. The volume of nerve in the subepidermal neural plexus has been

880

Pathology of the Peripheral Nervous System

Table 34–1. Epidermal Nerve Fiber Density* Normal

Foot

Calf

Thigh

Hand

Forearm

Back

Mean 95% Cutoff Standard Deviation Number

22.83 12.80 7.61 39

18.58 6.73 6.54 39

34.99 19.60 15.88 39

24.57 10.40 11.31 38

37.57 17.10 12.96 37

62.61 33.00 21.25 23

*Normative data have been determined at the University of Minnesota for proximal and distal anterolateral thigh, proximal and distal posterolateral calf, dorsum of the foot over the extensor digitorum brevis muscle, over the first dorsum interosseus muscle of the hand, the proximal volar forearm, and the paraspinal area of the back at T4 to T5.

estimated based on the total fluorescence,55 but accurate quantification of nerves in the subepidermal neural plexus by thresholding from background is complicated by the positive immunoreactivity of dermal fibroblasts for PGP 9.5. A semiquantitative rating system can be used to assess

the density of the plexus (Fig. 34–13): 1  hyperinnervation, 0  normal, 1  mild (50%) loss, 2  severe (75%) loss, 3  a trace, and 4  no nerve. A similar semiquantitative rating system can be used to quantify sudomotor nerves. These rating systems require that the

FIGURE 34–12 Blister analysis. A, Blister profile: confocal image of an immunostained sectioned blister. Nerves (PGP 9.5) are green or yellow, and basement membrane (type IV collagen) is red. Epidermis separated from dermis just above the dermal-epidermal basement membrane. Dermal capillaries and the subepidermal neural plexus remained intact (left box). Epidermal nerves, severed from their proximal segment, remained in the blister roof (right box). Bar: 100 m. B, Survey confocal image of a 3-mm blister roof immunostained for nerve with antibody to PGP 9.5. Rectangles indicate the areas imaged at 20  for nerve counts. Bar: 500 m. C, Higher magnification image is used for quantification of epidermal nerve fibers (from right rectangle in B). Bar: 100 m. (From Kennedy, W. R., Nolano, M., Wendelschafer-Crabb, G., et al.: A skin blister method to study epidermal nerves in peripheral nerve disease. Muscle Nerve 22:360, 1999, with permission.) See Color Plate

Pathology and Quantitation of Cutaneous Innervation

881

FIGURE 34–13 Subepidermal neural plexus (SNP) rating. A semiquantitative rating system can be used to assess the density of the plexus. SNP  0 indicates a normal innervation pattern in the papillary dermis with nerve bundles running subjacent and parallel to the epidermis. SNP  1 denotes hyperinnervation, a condition that is encountered in early stages of neuropathy when active regeneration is occurring. It is often associated with morphologic abnormalities such as swellings, clustering of nerves, and truncation of nerve fibers at the dermal-epidermal basement membrane. SNP  1 signifies a moderate (50%) loss of SNP innervation, and is usually accompanied by a reduction in epidermal nerve fiber density. SNP  2 indicates a severe (75%) loss. SNP  3 applies when only a trace of SNP remains. SNP  4 indicates that no nerve remains in the papillary dermis. See Color Plate

viewer have an excellent interpretation of “normal.” Other techniques utilizing image analysis to quantify cutaneous nerves have recently appeared107 but have not yet been widely accepted. The density of sudomotor nerves around secretory tubules of sweat glands can be quantified by calculating nerve volume per sweat gland volume in confocal images.46 Sudomotor nerve density is below normal in patients with severe neuropathy (Fig. 34–14), whereas sweat glands often appear well innervated in mild or moderate neuropathy, even though the secretion of sweat is decreased. This suggests that quantitation of sudomotor nerves may not be helpful for early diagnosis of diabetic neuropathy. A possible explanation for this apparent discrepancy is that sweat glands in nonhairy human and murine skin receive multiple sudomotor nerves. Multiple innervation facilitates collateral reinnervation after a

partial nerve lesion,44 but sweat volume may be compromised. The occasional findings of increased sweat gland innervation46 and of increased sweat secretion58 in proximal skin locations of diabetic subjects may represent localized nerve growth in response to a general stimulus for collateral reinnervation during the early denervation stage of neuropathy. We have not attempted to quantify the innervation of arrector pilorum muscles, hair follicles, or arteries. Quantitation of the nerves to arrector pilorum muscles is possible with the quantitative methods described for sweat glands. This may be useful in the clinical diagnosis of peripheral autonomic nerve disease if adequate sampling of these muscles is achieved. Hair follicles have an abundant supply of associated myelinated and unmyelinated nerve fibers, but the relatively few follicles visualized in biopsy of nonhairy skin create a sampling problem. Evaluation for

882

Pathology of the Peripheral Nervous System

clawlike terminals. In general, these morphologic features, particularly the intra-axonal swellings, are more evident at the thigh than in regions below the knee (see Fig. 34–3C). The composition of intra-axonal swellings remains uncertain. Neurofilament stains are negative and electron microscopy studies to date have been relatively inconclusive, suggesting that the swellings represent collections of organelles. Because of their location proximal to sites with epidermal denervation, we believe that swellings represent predegenerative changes. In models of epidermal regeneration, clusters of terminal nerve fiber swellings appear more commonly in neuropathic individuals, suggesting that such clusters of swellings may represent aberrant efforts at nerve sprouting, or microneuromas. The interaction of nerve with basement membrane may be altered in neuropathy as evidenced by morphologic irregularities found in patient biopsies. Some nerve fibers follow a course along the dermal side of the basement membrane but do not penetrate, frequently ending in a bulblike swelling (see Fig. 34–3D). In patient biopsies with low nerve fiber density, nerves entering the epidermis often branch at the point of penetration through the basement membrane and form a tuft of nerve endings that extend into epidermis. Similarly, clustering of epidermal nerve fibers is frequently seen in disease states, wherein multiple nerve fiber penetrations are interspersed with relatively long segments of nerve-free epidermis. FIGURE 34–14 Sweat gland nerve density. These images depict the innervation of sweat glands from normal (top) and diabetic (bottom) human skin. Samples used for quantification are immunostained with anti–PGP 9.5 and imaged confocally. Nerve volume is computer-rendered.

clinical purposes may be possible in the scalp, depending upon the number of follicles available for analysis. The wide distribution of cutaneous arteries and arterioles and the variable number of nerves per vessel complicate attempts to quantify vasomotor nerves.

Morphologic Changes in Sensory Neuropathy In healthy subjects, epidermal fibers arise from subpapillary nerve fiber bundles and run toward the stratum corneum in either branched or unbranched patterns. The fibers are usually thin, are sometimes varicose, and often end with a single clublike enlargement. In cases of sensory neuropathy, intraepidermal nerve fibers show distinct morphologic changes. Fibers frequently show more tortuous courses and more complex ramification.16,53 In addition to the increased branching complexity, there often is increased varicosity with stubby, rounded projections; intra-axonal swellings; and

Correlation with Tests of Unmyelinated Nerve Function Relatively few studies have examined the relationship of epidermal nerve fiber densities to other tests of nerve function. Heat pain has been examined in humans using a chemical (topical capsaicin) denervation model. While no correlation was found between heat pain sensitivity and epidermal nerve fiber density using a large (900-mm2) diameter probe, statistically significant correlations between fiber density and heat pain sensitivity were found when a small (7.1-mm2) probe delivered the stimulus.50 This apparent incongruity is believed to result from the spatial summation of surviving epidermal nerves plus stimulation of the proximal stumps of degenerated epidermal nerves in the dermis (some of which may be sensitized by the degeneration of the nerve ending) that occurs when the larger probe heats a deeper area. The smaller probe heats a more isolated shallow area, with fewer nerve endings activated. Comparisons of intraepidermal nerve fiber densities with just noticeable difference (JND) thresholds for cooling and vibration levels were made in HIV-associated sensory neuropathies using the CASE IV instrument (WR Company, Stillwater, MN) and were inversely correlated. Because high JND indicates decreased sensitivity to sensory stimuli, it is not unexpected that the correlation estimates were negative.

Pathology and Quantitation of Cutaneous Innervation

A correlation analysis based upon the percentile distribution rather than the JND is another statistical approach, but given that the sample size was relatively small and that many of the subjects had mild neuropathy, this analysis was based on the categorization of subjects as normal/abnormal based on their JNDs. Surprisingly, intraepidermal nerve fiber density correlated inversely more closely with vibratory threshold than with cooling threshold. Vibration is subserved by A fibers that are not present in the epidermis while cool threshold detection is mediated by A fibers, which may be included in measurements of intraepidermal nerve fiber densities. An association with C fibers, which subserve noxious heat and mechanical pain and represent the bulk of epidermal nerve fibers, was not included in the QST battery. The inverse correlation of intraepidermal nerve fiber density and vibratory threshold may be partially explained by concurrent involvement of large and small fiber nerves. Furthermore, the fact that only a minority of subjects (43%) had abnormal epidermal nerve fiber densities at the distal leg at baseline suggests that the severity of the neuropathy in these subjects was relatively mild (and perhaps that the epidermal denervation in these subjects had not progressed above the ankle).82 In another study of 32 patients with painful, burning feet there was no concordance between the results from quantitative sudomotor axon reflex testing (QSART), a QST (vibration and cooling by CASE IV) performed on the upper and lower extremities, and skin biopsy from the distal calf, in which 28 (87.5%) had a reduction of intraepidermal nerve fibers.78 Intraepidermal nerve fiber densities have also been assessed in small fiber and autonomic neuropathies, and compared with autonomic screening tests. In a study that was designed to quantify the severity of autonomic impairment in patients with painful neuropathy, excellent correlation was observed between QSART and cooling abnormalities and loss of intraepidermal nerve fiber density.73 The QSART quantitatively evaluates the postganglionic sympathetic sudomotor axon58 (see also Postural Tachycardia Syndrome below).

883

densities were highly correlated (r  .87, P .0001). Intraepidermal nerve fiber density and sural nerve small fiber measures were concordant in 73% of patients. In 23% of cases, reduced intraepidermal nerve fiber density was the only indicator of small fiber depletion. It was usually normal in acquired demyelinating neuropathies and where the clinical suspicion for neuropathy was low. The inferences are that determination of distal leg intraepidermal nerve fiber density may be more sensitive than sural nerve biopsy in identifying small fiber sensory neuropathies and that intraepidermal nerve fiber density and assessments of large nerve fibers by nerve biopsy and electrophysiology are all useful for characterizing sensory and sensorimotor neuropathies.31

Utility of Skin Biopsy It is important in clinical practice to have a good sense of the utility (and limitations) of skin biopsy for evaluation of the peripheral nervous system. The biopsy can only provide information about the most distal terminals of sensory and autonomic nerves where they meet their target organ. Interpretation can be affected by local trauma, other diseases, or biopsy artifacts. Nonetheless, biopsy of proximal and distal sites can provide inbuilt intraindividual controls. We have found the punch skin biopsy technique to be clinically useful for defining presence and severity of focal neuropathy and of distally dominant sensory neuropathy and for differentiation of neuropathy from radiculopathy. It remains to be conclusively demonstrated that skin biopsy can track changes in neuropathies over time, or with regenerative treatments, although the technique has this potential.

CONTRIBUTIONS OF SKIN BIOPSY TO DIAGNOSIS Quantitation of epidermal nerve fiber density and morphology has been helpful for evaluating several clinical disorders.

Correlation between Epidermal Nerve and Sural Nerve

Painful Sensory Neuropathy

Twenty-six patients with neuropathic complaints had sural nerve morphometry and determination of epidermal nerve fiber density at the distal part of the leg. Nonparametric correlations were used because of the nonlinearity of the values. Intraepidermal nerve fiber density correlated with the densities of total myelinated fibers within the sural nerve (r  .57, P  .0011), small myelinated fibers (r  .53, P  .029), and large myelinated fibers (r  .49, P  .0054). There was a trend toward an association between intraepidermal nerve fiber density and sural nerve unmyelinated nerve fiber densities (r  .32, P  .054). Sensory nerve action potential amplitudes and large myelinated nerve fiber

In 20 patients with painful sensory neuropathies, intraepidermal nerve fiber density was significantly reduced compared with age-matched controls, even in regions proximal to the areas of clinically identifiable sensory abnormalities. For example, patients with grade 1 neuropathy in whom sensory abnormalities were restricted to the toes and feet had significantly reduced intraepidermal nerve fiber densities at the calf. Cooling thresholds on QST (CASE IV device) did not correlate with intraepidermal nerve fiber density or the clinical grade of sensory neuropathy (Fig. 34–15). Intraepidermal nerve fiber density from calf and thigh correlated against clinical estimates of neuropathy severity.35

884

Pathology of the Peripheral Nervous System

25

Axonai Density

IENF Density

40,000

20 15 10 5

30,000

20,000

10,000 0

0

25

The clinical features, natural history, and neuropathology were described for 32 patients presenting with “burning feet,” thought to represent idiopathic small fiber sensory neuropathies. Two clinical patterns were apparent based on natural history and anatomic location of cutaneous denervation. Most patients (28 of 32) presented with neuropathic pain that was initially restricted to the feet and toes, but eventually extended more proximally to involve the legs and hands. Intraepidermal nerve fiber density was most severely reduced distally, with more normal intraepidermal nerve fiber densities at proximal sites. The minority (4 of 32) presented with abrupt onset of generalized cutaneous burning pain and hyperesthesia. In these patients intraepidermal nerve fiber densities were reduced at both proximal and distal sites. For the entire group, intraepidermal nerve fiber densities were reduced at the calf site below the fifth percentile in 81% of all patients.34 Figure 34–16 compares nerve densities in biopsies of the skin and sural nerve from one patient from each group. Skin biopsy was obtained in a separate study of 57 patients with painful, burning feet; minimal signs of neuropathy; and normal nerve conduction studies. Of these, 44 (77%) had a reduced number of intraepidermal nerves at the distal calf. Some nerves ended abruptly just below the basement membrane; others had enlarged, swollen terminal nerve endings. Reduced ankle reflexes were found in 11%, pinprick loss on toes and feet in 45%, and loss of vibration sense in 43%. Only three patients had conditions known to be associated with peripheral neuropathy.78

40,000 20

Axonai Density

IENF Density

FIGURE 34–15 Intraepidermal nerve fiber density from calf (black circles) and thigh (white squares) skin correlated against clinical estimates of neuropathy severity in patients with painful sensory neuropathy. The data shown are group means with standard error bars. (From Holland, N. R., Stocks, A., Hauer, P., et al.: Intraepidermal nerve fiber density in patients with painful sensory neuropathy. Neurology 48:708, 1997, with permission.)

15 10 5 0

30,000

20,000

10,000

0

Distal Proximal Relative Position of skin Biopsy Skin fibers

Unmyelinated

FIGURE 34–16 Intraepidermal nerve fiber (IENF) densities at various sites for chronic progressive idiopathic small fiber sensory neuropathy (SFSN) (left) and sural nerve morphometry for chronic progressive idiopathic SFSN patient 1 (top right) and patient 2 (bottom right). Left, IENF density (mm 1) at various sites for patients compared with normal controls (mean  standard error shown by shaded rectangle and 5th and 95th percentile range shown by bars). Right, Unmyelinated fiber density (axons/mm2) for patients compared to the 5th and 95th percentile range for normal controls (bars). Patient 1 (top) had normal densities of myelinated and unmyelinated fibers in the sural nerve biopsy specimen (taken from the calf) despite marked cutaneous denervation in skin biopsies. Patient 2 (bottom), with more severe progressive idiopathic SFSN, had reduced densities of smalldiameter myelinated and unmyelinated axons in the sural nerve biopsy specimen (taken from the ankle) in addition to marked length-dependent cutaneous denervation. (From Holland, N. R., Stocks, A., Hauer, P., et al.: Intraepidermal nerve fiber density in patients with painful sensory neuropathy. Neurology 48:708, 1998, with permission.)

Diabetic Neuropathy Earlier work in painful diabetic neuropathy suggested that there was a predominance of involvement of unmyelinated and small myelinated nerve fibers in the sural nerve.8 Small fibers have long been recognized to be involved in diabetic polyneuropathy, and skin biopsies have confirmed that epidermal denervation can occur relatively early, and generally

Pathology and Quantitation of Cutaneous Innervation

0.06

0.25

Nerve vol/sweat gland vol Nerve length/epidermal vol

0.20

0.05 0.04

0.15

0.03 0.10

0.02 0.05

0.00

0.01

0 (Mild)

25

50

75

Clinical Score

100

0.00 125

Length of epidermal nerve (μm) per epidermal volume (μm3)

Sweat gland nerve volume per volume of sweat gland

in a length-dependent manner. The total length of epidermal nerve per volume of epidermis correlates with the overall clinical severity (Fig. 34–17). Some of the surviving epidermal nerves have morphologic abnormalities, particularly axon swellings.47,80 The subepidermal neural plexus is often thinner than in control skin, and fibers within the subepidermal plexus may have a thickened, dystrophic appearance. In some neuropathies single nerve fibers are seen ending in a terminal enlargement just under the basement membrane.46 These fibers appear to be unsuccessfully attempting to penetrate basement membrane en route to regeneration into the epidermis, although it is also possible that they are in the process of dying back. Sweat gland innervation is sometimes abnormal or completely lost in diabetic patients (Fig. 34–17), but it appears not to be useful as an early indication of neuropathy.46 Epidermal nerve fibers also appear to be prominently affected in patients with impaired glucose tolerance (IGT). Several groups have reported an increased prevalence of IGT among patients with painful sensory neuropathy,74,97 and randomly selected patients with sensory neuropathy and IGT were found to have reduced epidermal nerve fiber densities.99 Of 97 patients with clinically diagnosed predominantly sensory neuropathy of undetermined etiology evaluated with oral glucose tolerance tests, 36% were classed

(Severe)

FIGURE 34–17 Neuropathy score correlated with epidermal nerve fiber density and sweat gland innervation density in diabetic subjects. Correlation of cutaneous nerve quantitation with clinical assessment. The amount of nerve in calf biopsies correlated with the clinical evaluation of most diabetic subjects. The data suggest that the epidermal nerve assessment is useful in evaluating mild to moderate neuropathy, and sweat gland nerve analysis is helpful in later stages of neuropathy. Reduction of epidermal nerve is seen in mild cases of neuropathy, whereas no epidermal nerve is seen in severe cases. Sweat gland innervation remains at normal levels until neuropathy is quite severe. The peripheral and autonomic systems of the diabetic subjects were evaluated from graded results of history and examination, nerve conduction, cardiorespiratory reflexes, warm and cold sensitivity, and sweat testing. (Kennedy, W. R., and Wendelschafer-Crabb, G.: Utility of skin biopsy in diabetic neuropathy. Semin. Neurol. 16:163, 1996, with permission.)

885

as having IGT and 20% as having diabetes. Both the IGT and diabetes groups had reduced distal leg intraepidermal nerve fiber densities. The neuropathy associated with IGT was less severe than the neuropathy associated with diabetes. Patients with predominantly small fiber neuropathy had a shorter duration of neuropathy symptoms, and most had IGT. The results suggest that small-caliber nerve fiber loss may be the earliest detectable sign of neuropathy in glucose dysmetabolism.103 Diabetic Truncal Neuropathy In diabetic truncal neuropathy, skin biopsies from symptomatic regions show a loss of intraepidermal nerve fibers. After clinical recovery there may be nerve fiber regeneration.53

HIV-Associated Sensory Neuropathies Sensory neuropathies of HIV/AIDS affect at least 30% of individuals with AIDS.90 The distal sensory polyneuropathy associated with HIV overlaps clinically with the toxic neuropathy provoked by specific antiretrovirals. The sensory neuropathies produce paresthesias or burning pain in the feet, often associated with hyperalgesia and lightning pains. Unlike HIV-associated dementia, the incidence rates have not declined with the advent of highly active antiretroviral therapy, probably because specific dideoxynucleoside analogues (ddI, d4T, and ddC) produce peripheral neurotoxicity.66 Punch skin biopsies were originally developed at the Johns Hopkins University laboratory to study HIV-associated sensory neuropathies that had previously been shown by sural nerve biopsy to be associated with prominent involvement of unmyelinated nerve fibers.17 The most systematic survey of skin biopsies in HIV-associated sensory neuropathy comes from a trial of recombinant human NGF. Skin biopsies were included in this trial as an outcome measure, the first reported application of skin biopsy to measure treatment effect. Sixty-two of 270 patients with sensory neuropathies who participated in the trial of recombinant human NGF were included in a substudy examining the density of intraepidermal nerve fibers.62 Fiber density was inversely correlated with neuropathic pain as measured both by patient and physician global pain assessments, but not using the Gracely Pain Scale. The reproducibility of the technique, as assessed from intrasubject correlation between baseline and the week 18 densities, was 81% in the distal part of the leg and 77% in the upper thigh. Decreased intraepidermal nerve fiber density at the distal leg was associated with lower CD4 counts and higher plasma HIV RNA levels. There was no treatment effect over the relatively short period of 18 weeks, and a few biopsies studied out to 70 weeks remained stable over time. Although no significant treatment effect was observed over the 18 weeks, the reproducibility of the technique appears to be good, suggesting that this technique could be incorporated as an outcome measure in future trials of regenerative agents for sensory neuropathies.

886

Pathology of the Peripheral Nervous System

Friedreich’s Ataxia Friedreich’s ataxia is an autosomal recessive inherited ataxia. Hyperexpression of a GAA trinucleotide on the causative gene (X25) localized to chromosome 9 results in reduction of the protein frataxin.9 Pathology of large-diameter myelinated nerve fibers in peripheral nerves is consistent with the loss of vibratory sense and deep tendon reflexes and severely reduced sensory action potentials. Small fiber involvement was postulated but unproven19 until severe loss of unmyelinated sensory nerves and of the autonomic innervation to sweat glands, arrector pili, and arterioles was uncovered in thick sections of skin biopsies.70 The small fiber loss correlated with reduced mechanical and thermal nociception.71

Restless Legs Syndrome The restless legs syndrome (RLS) is characterized by a compelling urge to move the extremities. It is often associated with paresthesias or dysesthesias, motor restlessness, worsening of symptoms with rest, relief by activity, and worsening of symptoms in the evening or night. Two forms are described: crurum dolorosum, characterized by pain, and crurum paresthetica, characterized by paresthesias.21 Numerous risk factors are reported for RLS, and there is a suggestion that early-onset (age 45 or younger) and late-onset RLS differ etiologically.1,2 Nerve biopsies performed in an unselected group of primary patients suggest that the incidence of peripheral neuropathy may be underestimated.37 Twenty-two consecutive patients with RLS were evaluated for evidence of large fiber neuropathy (LFN) and small sensory fiber loss (SSFL). Neuropathy was identified in eight (36%). Three patients had pure LFN, two had mixed LFN and SSFL, and three had isolated SSFL. Patients with SSFL had a later onset, reported pain in their feet more frequently, and tended not to have a family history of RLS. Patients without SSFL did not associate onset with age, family history, or presence of pain. The report suggested that two forms of RLS exist: one is triggered by painful dysesthesias associated with SSFL and has a later onset and a negative family history; the other does not have small sensory fiber involvement, has an earlier onset and a positive family history, and is without pain.79

Familial Dysautonomia (Riley-Day Syndrome, Hereditary Sensory and Autonomic Neuropathy Type III) Familial dysautonomia3,86 is an autosomal dominant disorder of sensory and autonomic nerves found in Ashkenazi Jews that is caused by mutation of the IKBKAP gene on chromosome 9q31.6 Clinical features include increased sweating, absent overflow lacrimation, vomiting crises, decreased sensitivity to pain, depressed

Achilles reflexes, postural hypotension, and absence of tongue fungiform papillae and of the axon flare response to intradermal histamine. Biopsy of glabrous skin shows severe loss of unmyelinated sensory and autonomic nerves at the calf and moderate loss in the paraspinal region of the upper back.33

Congenital Insensitivity to Pain with Anhidrosis (Hereditary Sensory and Autonomic Neuropathy Type IV) Congenital insensitivity to pain with anhidrosis (CIPA) is a rare condition caused by mutation of the TrkA (NTRK1) gene on the 1q 21-22 chromosome,38 characterized by mental retardation; congenital analgesia that leads to selfmutilation, multiple scars, and fractures; and anhidrosis with repeated bouts of fever.87 Nerve biopsy shows loss of unmyelinated and small myelinated fibers.25 Recently a 10-year-old girl was reported with a low IQ, pain, and thermal insensitivity. The remainder of the neurologic examination, sensory nerve conduction velocities, cardiovascular reflexes, and visual, brainstem, and somatosensory evoked potentials were normal. Sural nerve biopsy was devoid of small myelinated nerves and contained only rare unmyelinated nerves. Epidermal nerve fibers were essentially absent in skin from the thigh, back, and calf. The few sweat glands present were hypotrophic and without innervation.69

Psoriasis Psoriasis is a common skin disorder characterized histologically by epidermal hyperproliferation, dilated tortuous capillaries, and a lymphocytic infiltrate. Skin biopsy shows that the epidermal nerves are greatly elongated through a thickened epidermis to the stratum corneum.39 Several treatment modalities are available: therapy that targets epidermal proliferation, immunomodulators, or physical treatment, such as the use of lasers. The pulsed dye laser at 585-nm wavelength targets blood vessels and selectively destroys dilated vessels in the dermal papillae. Laser therapy is followed by remodeling of the epidermis and subsequent regeneration of capillaries and epidermal nerves with return to a more normal appearance.118

Port-Wine Stains It is proposed that the pathogenesis of port-wine stains might be related to a lack of innervation around the ectatic blood vessels.89 Biopsy of the involved skin shows a deficiency of innervation that is inversely proportional to the size of the vascular spaces. Moreover, nerve density may be predictive of the response of port-wine stain lesions to laser treatment.117

887

Pathology and Quantitation of Cutaneous Innervation

Leprosy

Postherpetic Neuralgia

Severe loss of immunoreactivity to PGP 9.5 and neurofilament antibodies occurs in leprosy. Loss is greatest in the tuberculoid variety of leprosy, less in lepromatous, and least in the indeterminate variety.41 Curiously, there is almost complete loss of reactivity for the neuropeptides CGRP, SP, and neuropeptide Y. This raises the possibility that neuropathy may be detected early by loss of immunoreactivity to some antigens while reactivity to other antigens remains, at least temporarily.

Shingles represents reactivation of latent varicella-zoster virus in a sensory ganglion, causing loss of a variable proportion of the sensory neurons. In some patients, neuropathic pain persists for months or years (postherpetic neuralgia). Hyperalgesia and allodynia to light touch are often prominent. Two studies have demonstrated that the density of epidermal nerves in areas of allodynia is reduced, although they came to differing conclusions. One study found that higher fiber densities correlated with pain, an observation that might suggest that intact but hyperactive nociceptors were responsible. The reduction in intraepidermal nerve fiber density correlated with the magnitude of thermal sensory deficits.88 By contrast, a second study75 found that pain correlated with lower intraepidermal nerve fiber densities, and interpreted the results as more compatible with central sensitization. Interestingly, there was also a reduction of intraepidermal nerve fiber density in the contralateral, unaffected epidermis.115

Fabry’s Disease Fabry’s disease is an X-linked recessive disorder caused by deficiency of -galactosidase A activity. Hemizygotes develop deposits of neutral glycosphingolipids, principally ceramide trihexoside, throughout the nervous system but predominantly in vascular endothelial cells. A painful small fiber neuropathy develops that is difficult to detect and quantitate by conventional methods. The neuropathy can develop in children as young as 5 years of age, with characteristic episodes of acral burning pain. Twenty Fabry’s disease patients (hemizygotes, ages 19 to 56 years) with preserved renal function were found to have normal nerve conduction studies and large fiber quantitation by sural nerve biopsy. By contrast, involvement of small cutaneous fibers in these patients was easily demonstrated and quantified by punch skin biopsy. All patients showed severe loss of intraepidermal innervation at the distal part of the leg, with a density of 0 to 2.4 fibers/mm versus a density for control subjects at this site of 4.7 to 6.5 fibers/mm. Fiber loss at the distal thigh was proportionately less severe. Fabry’s disease patients underwent biopsy at 6-month intervals for periods ranging from 1 to 3 years with no significant longitudinal changes in density except in two patients who demonstrated a rapid decrement in innervation density following an increase in spontaneous pain.92

Sensory Ganglionopathies Ganglionopathy patients tend not to show the normal gradient in intraepidermal nerve fiber density between the proximal thigh and the distal part of the leg—that is, they show a non–length-dependent process. One cohort presented with predominant gait and limb ataxia and proprioceptive sensory loss; 8 of 16 patients had positive sensory symptoms. Causes of the ganglionopathy were paraneoplastic in six and idiopathic in seven, and one had a hereditary sensory and autonomic neuropathy. In ganglionopathies, the mean intraepidermal nerve fiber density did not differ between the proximal thigh (10.37/mm) and the distal leg (10.41/mm). Compared to other sensory neuropathies and controls, the intraepidermal nerve fiber density was significantly lower at the proximal thigh site in patients with ganglionopathies.54

CADASIL Cerebral autosomal dominant arteriopathy with subcortical infarct and leukoencephalopathy (CADASIL) is one example of how skin biopsy can be used for analysis of nonneural substances in investigation of neurologic disease. CADASIL is an autosomal dominant disorder that presents as migrainous headaches, cerebral ischemia, and multi-infarct dementia. An abnormality on chromosome 19q12 produces a missense point mutation in the notch3 gene. Punch skin biopsy has been used to identify characteristic granular, electron-dense, osmiophilic material attached to vascular smooth muscle cells.109 However, a commercial laboratory test for the genetic abnormality is now available from Athena Labs (Worcester, MA), which should provide improved specificity and sensitivity.30

Chronic Inflammatory Demyelinating Polyneuropathy Patients with chronic inflammatory demyelinating polyneuropathy (CIDP) were recently shown to have about 50% reduction in intraepidermal nerve fiber density compared to age- and gender-matched controls, suggesting that smalldiameter sensory nerves as well as large-caliber nerve fibers are affected in CIDP.15

Postural Tachycardia Syndrome A collaborative study by Johns Hopkins University and the Mayo Clinic examined the relationship between the QSART and epidermal nerve fiber densities in patients with postural tachycardia syndrome (POTS).96 Inclusion criteria for POTS in the study included orthostatic heart

888

Pathology of the Peripheral Nervous System

rate increment (from 30 beats/min to an absolute heart rate of 120 beats/min within 5 minutes of head-up tilt), symptoms of orthostatic intolerance, distal abnormal QSART, and abnormalities or loss of vasoconstrictor response. In POTS, distal autonomic fibers are affected in isolation, sparing somatic fibers, which have consistently normal nerve conduction studies.91 Intraepidermal nerve fiber density correlated with peripheral postganglionic sudomotor status as assessed by the Composite Autonomic Severity Score (CASS)102 both in the category of the sudomotor subscores and in the overall total score, which is a composite index of combined sudomotor, adrenergic, and cardiovagal function. In POTS the mean intraepidermal nerve fiber density was normal, while the CASS was abnormal. Although three of the eight patients showed some minor morphologic abnormalities, the normalcy of the epidermal nerve fiber densities suggests that distal involvement of autonomic fibers occurs in isolation in POTS.

Pediatric Neurologic Disorders Skin biopsies have been less widely studied in pediatric neurologic disorders, but there are some conditions in which examination of the cutaneous innervation has characteristic changes. Two examples are provided. Giant Axonal Neuropathy Giant axonal neuropathy is an autosomal recessive neurologic disorder characterized by the development of a severe polyneuropathy, central nervous system abnormalities, and characteristic tightly curled hair. Mutations in the gigaxonin gene have been identified as the underlying genetic defect.51 Ultrastructurally, accumulations of neurofilaments and osmiophilic aggregates are found in giant axons. Intermediate filaments accumulate in Schwann cells, perineural cells, fibroblasts, and endothelial and epithelial cells in both nerve and skin biopsies.104 We have observed giant swellings within dermal nerve fibers, although the density of epidermal nerves appears to be normal. Neuroaxonal Dystrophies Neuronal intranuclear inclusion disease has no defined metabolic abnormalities, but it affects the peripheral nervous system, which could therefore be used to identify morphologic abnormalities.24 In infantile neuroaxonal dystrophy, there is also an accumulation of intermediate neurofilaments, as well as mitochondria and organelles with vesiculotubular profiles.59 There is limited experience with the use of skin biopsies in this condition; however, some children have shown dystrophic changes in the subepidermal neural plexus.

Other Conditions Depletion of epidermal nerve fibers has been demonstrated in a variety of conditions with dermatologic manifestations. Image analysis detected a significant decrease in PGP 9.5–immunoreactive nerves in the epidermis and subepidermal layers and in CGRP-immunoreactive nerves around the capillaries within dermal papillae of patients with Raynaud’s phenomenon, although the changes were not always apparent by visual screening.105 Patients with systemic sclerosis were found to have decreased CGRP- and PGP 9.5–immunoreactive nerves in all skin areas and vasoactive intestinal polypeptide–immunoreactive nerves in the sweat glands.105 An increased number of PGP 9.5–immunoreactive fibers has been reported in the dermis and dermal-epidermal junction in atopic dermatitis, a condition in which the sensation of itching is prominent.108 It has been suggested that increased dermal innervation could explain the pruritus, burning pain, and hyperalgesia in notalgia paresthetica,100 and that an increase of sensory neuropeptides SP and CGRP might be involved in the pathogenesis of nodular prurigo.64 No changes of innervation were found in patients with lichenified eczema.64

RESEARCH USES OF SKIN BIOPSY The existence of unmyelinated sensory nerves within a few microns of the skin’s surface has provided opportunities to study the reactions of peripheral nerves to mechanical, thermal, and chemical stimuli under safe, minimally invasive circumstances. Early experiments, using methylene blue staining, localized the itching sensation caused by cowphage spicules and injected proteases to the subepidermal neural plexus.93 Use of immunohistochemical staining has allowed expansion of research into the correlates of sensory testing, effects of mechanical and chemical trauma to cutaneous nerves, and regeneration of sensory nerves.

Human Models of Nerve Regeneration Skin Biopsy Variations We developed two models to study reinnervation of the epidermis.83 One model uses a circular incision that transects the subepidermal plexus, resulting in Wallerian degeneration of the nerve fibers that enter the incised cylinder, leaving a defined zone of denervated dermis and epidermis. The earliest reinnervation of epidermis occurred by collateral sprouting from the terminals of epidermal axons from just outside the incision line. These terminals extended horizontally across the incision line and through the superficial layers of the epidermis, beneath the stratum corneum. By 13 days, numerous regenerating axons appeared in the deeper dermis derived

889

Pathology and Quantitation of Cutaneous Innervation

After Capsaicin Treatment

During Capsaicin Treatment % Deviation from normal where normal = 100 (treated/control area)

from transected axons. The latter regenerating axons grew toward and ultimately into the epidermis, so that epidermal axonal density had normalized by 30 to 75 days. The invasion of these axons was associated with regression of the horizontally growing collateral sprouts. The second model utilizes an identical incision followed by removal of the incised cylinder of skin, leaving a denervated area in which Schwann cells are absent. New fibers arose by terminal elongation of the epidermal axons outside the incision line, as in the incision model, and especially by collateral branching of epidermal fibers at the incision margins. These collaterals reached the epidermal surface of the basal lamina at the dermal-epidermal junction, and then grew slowly toward the center of the denervated circle. In contrast to the first model, complete reinnervation was not achieved even after 23 months. These and other models might be used to study reinnervation of denervated skin in humans in different injury models and have relevance for exploring the stimuli for axonal growth and remodeling.

700 600 500 400 300 200 100 0

Heat (Latency) Cold (Latency)

100 80 60 40 20 0 time 0

D3 W1 W2 Heat Pain (VAS) Cold Pain (VAS) ENF Density

W3

W1

W2 W3 Touch Pain (%) Pain (VAS)

W6

Skin Blister: A Model of Mechanical Denervation Mechanical denervation of the terminal 40 to 60 m of epidermal sensory nerve fibers can be produced to study wound healing and sensory reinnervation by removing the epidermis with a skin blister.49 Regeneration of new epidermis occurs within 3 days after a 3-mm blister wound. The new epidermis is initially innervated by collateral sprouts arising from epidermal nerve fibers in the normal skin surrounding the blistered area. Slightly later, the proximal stumps of epidermal nerves at the base of the blister begin to enter the new epidermis, first near the edge of the lesion and later at the center. Eventually, regenerated nerves from the proximal stumps in the base reinnervate the epidermis in near-normal numbers and the collateral branches disappear.112 Reinnervation is accompanied by return of sensation within 30 days. Similar reinnervation occurs in patients with diabetic neuropathy, even in areas of reduced sensation, but the return of epidermal nerve fibers is only to the level of the prelesion innervation.42

FIGURE 34–18 Capsaicin-induced denervation and reinnervation. The mean ( standard error) change in sensation and number of nerve fibers expressed as percent deviation from normal during and following the course of topical capsaicin application. Note that scale for latencies (top) is different from that for other sensitivities and nerve fiber numbers (bottom). Touch was measured using Semmes Weinstein monofilaments. Intensity of pain was measured using a visual analogue scale (V.A.S.) for mechanical pain, heat pain, and cold pain. Mechanical pain (pain % and pain V.A.S.) was evaluated with a spring-mounted pin. Heat pain (V.A.S.) and cold pain (V.A.S) were tested with a 2-mm probe. Epidermal nerve fiber (ENF) number per area was calculated from analysis of nerves in blister roofs. (From Nolano, M., Simone, D. A., Wendelschafer-Crabb, G., et al.: Topical capsaicin in humans: parallel loss of epidermal nerve fibers and pain sensation. Pain 81:135, 1999, with permission.)

Capsaicin: A Model for Chemical Denervation When capsaicin, a pungent ingredient in hot chili peppers, is applied topically, it causes warm, painful sensations. These are replaced after several hours by a period of hyposensitivity. Sensation returns toward normal in 2 to 3 weeks.95 This sequence was initially believed to result from desensitization of nociceptive receptors but was later shown to be induced by the influx of calcium ions with osmotic changes as well as activation of calcium-sensitive proteases resulting in axonal degeneration.113 More recently, skin biopsy showed that capsaicin administered intradermally is soon followed by depletion of epidermal nerves and superficial dermal nerves within 24 hours (Fig. 34–18). Reinnervation of the epidermis

reached approximately 50% of normal, and heat pain sensation reached 60% of normal after 4 weeks.94 Capsaicin administered topically also causes loss of epidermal and superficial dermal nerves and decreased sensitivity to noxious heat and mechanical (pin) stimulation. Serial biopsies reveal that nerve regeneration after topical capsaicin is faster than after intradermal delivery. The return of epidermal nerves and sensation is nearly complete in 30 days (Fig. 34–19).72 These findings strengthen the hypothesis that epidermal nerves are polymodal nociceptors. The detection of a loss and subsequent recovery of sensitivity to noxious heat was possible when stimuli were delivered by a small probe (2- to 3-mm diameter) but not by the larger 30  30-mm probe of the type common on

890

Pathology of the Peripheral Nervous System

Diabetic subjects

100 μm

Healthy control subjects

150

100

% Baseline ENFD 50

0 0

25

50

Days

100

0

25

50

75

100

Days

FIGURE 34–20 ENFD regeneration expressed as a percent of baseline values in diabetic and healthy, nondiabetic control subjects. Healthy volunteers recovered their ENF density at a rate of 0.177  0.075 fibers/mm/day. The rate of nerve fiber regeneration was significantly decreased among those with diabetes (0.074  64 fibers/mm/day; P .001) and further reduced among diabetic subjects with neuropathy as compared to those without neuropathy (P  .003). (From Polydefkis, M., Hauer, P., Sheth, S., et al.: The time course of epidermal nerve fiber regeneration: Studies in normal controls and in people with diabetes, with and without neuropathy, Brain 127:1606, 2004, with permission.)

Control

Day 6

75

Kennedy & Wendelschafer- Crabb. University of Minnesota

FIGURE 34–19 Capsaicin-treated human epidermis. Nerves degenerate following topical capsaicin treatment, as can be seen by comparing control tissue at time 0 (top) with a sample taken after 1 week of capsaicin application (bottom). Epidermal nerve fibers have degenerated, while fibers in the subepidermal neural plexus are still plentiful. This provides a good model for the peripheral neuropathy seen in diabetic subjects, in whom epidermal nerve fiber loss precedes more proximal loss. We will use this model to study basement membrane during nerve regeneration. This model provides denervation without remodeling basement membrane. Bar: 100 m. See Color Plate

commercial instruments.50 Both of our groups have developed reproducible systems for applying capsaicin and examining the patterns and rate of regrowth of epidermal nerve fibers. In normal volunteers treated with topical capsaicin, epidermal nerve density returns at an average rate of 0.177 fibers/day (range, 0.023 to 0.362 fibers/day). The rate of nerve fiber regeneration is significantly decreased in diabetics (0.074 fibers/day; range, 0.002 to 0.258 fibers/day, P .001), and there was a trend for a lower rate of regeneration with increasing severity of neuropathy (Fig. 34–20).81a These models have potential as outcome measures in trials of regenerative agents and for improving understanding of trophic factor regulation after nerve injury.

Comparison of Nerve Function with Structure Comparisons have been made between the morphology of cutaneous nerve receptors and the results of electrophysiologic stimulation. Highly significant correlations were found between the density of normal Meissner’s corpuscles and proximally recorded sensory nerve action potential areas whether generated by electrical stimulation of the digital nerve or by tactile stimulation of the fingerpad. The density of abnormal Meissner’s corpuscles correlated significantly with the sensory action potential area but not with tactile potential area, whereas the density of papillary myelinated nerve fibers correlated with tactile but not with electrical stimuli. The natural stimuli appeared to be transmitted centrally more effectively by normal receptors than by atrophic receptors.71

Animal Models Using Skin Biopsy Several groups have now used skin biopsy techniques in animal models to examine denervation and regeneration after various surgical models of nerve injury.67,68,36,56,115 Early experiments showed that rat sciatic nerve transection resulted in denervation, thinning of the epidermis, and increased expression of PGP 9.5 by Langerhans cells.36 A mouse model has been devised to correlate the reinnervation time of muscles, sweat glands, and skin with the time required for return of function to these organs (Fig. 34–21). After sciatic nerve crush, sequential biopsies of foot muscles and footpads from the hind paw of the same mouse can be examined to detect the time of reappearance of alpha motor, sudomotor, and epidermal nociceptor nerves. Reappearance

891

Pathology and Quantitation of Cutaneous Innervation

of end-organ function can be tested at the same time intervals by sciatic nerve stimulation recording muscle, pilocarpine (intraperitoneal) stimulation to detect reappearance of sweat droplets, and pinprick to skin to detect return of nociception.68 The model has potential for testing the efficacy of therapeutic agents to facilitate nerve regeneration. Reinnervation of epidermal nerves in pigskin was tested after removal of a 3-mm skin biopsy and immediate replacement (autotransplantation). The first nerves that entered the newly regenerated epidermis over the wound were collateral branches of epidermal nerves in the normal uninjured epidermis surrounding the wound. These nerves grew on the epidermal surface of the basement membrane toward the center of the wound, then turned at an acute angle toward the stratum corneum. Nerve regeneration from the severed subepidermal plexus nerve trunks along the sides of the lesion began much later (weeks) and extended slowly toward the center of the lesion. Reinnervation never approached normal.29 The same experiment on one human subject gave identical results (M. Nolano, personal observation, 1997). An important experimental study demonstrated that intraepidermal nerve fiber density is significantly reduced after dorsal root ganglionectomy or sciatic nerve transection in rats, but not after dorsal and ventral radiculotomy, sympathectomy, or spinal motor neuron lesion.56

ACKNOWLEDGMENT A portion of this research is supported by grant NS44807 (J.M.).

REFERENCES FIGURE 34–21 Confocal micrographs of PGP 9.5 immunofluorescence in control (A) and reinnervating (B and C) mouse footpads following sciatic nerve crush. A, Nerves in control footpad form a dense network. Sweat glands are very heavily innervated. Large trunks extend through the central sweat gland area to form the subepidermal nerve plexus and Meissner’s corpuscles in the tufts of papillary dermis. The epidermis has many fine nerve endings. B, At day 14 postcrush, PGP 9.5–immunoreactive fibers returned in large nerve trunks from the base of the pad to the papillary dermis. C, By day 46 postcrush, the pattern of innervation resembles the control, but regenerated nerve fibers to the nerve trunks, subepidermal neural plexus, and sweat glands are less well compartmentalized, and the epidermis has shorter nerve fibers. (From Navarro, X., Verdu, E., Wendelschafer-Crabb, G., and Kennedy, W. R.: Immunohistochemical study of skin reinnervation by regenerative axons. J. Comp. Neurol. 380:164, 1997, with permission.)

1. Allen, R., LaBuda, M., Becker, P., and Earley, C.: Family history study of RLS patients from two clinical populations. Sleep Res. 26:309, 1997. 2. Allen, R. P., and Earley, C. J.: Defining the phenotype of the restless legs syndrome (RLS) using age-of-symptom-onset. Sleep Med. 1:11, 2000. 3. Axelrod, F. B., and Hilz, M. J.: Familial dysautonomia. In Appenzeller, O. (ed.): Handbook of Clinical Neurology: The Autonomic Nervous System, Part II. Dysfunctions. Amsterdam, Elsevier Science, p. 144, 2000. 4. Bajrovic, F., Kovacic, U., Pavcnik, M., and Sketelj, J.: Interneuronal signaling is involved in induction of collateral sprouting of nociceptive axons. Neuroscience 111:587, 2002. 5. Baker, K. W., and Habrowsky, J. E.: EDTA separation and ATPase Langerhans staining in the mouse epidermis. J. Invest. Dermatol. 80:625, 1983. 6. Blumenfeld, A., Slaugenhaupt, S. A., Liebert, C. B., et al.: Precise genetic mapping and haplotype analysis of the familial dysautonomia gene on human chromosome 9q31. Am. J. Hum. Genet. 64:1110, 1999.

892

Pathology of the Peripheral Nervous System

7. Bolton, C. F., Winkelmann, R. K., and Dyck, P. J.: A quantitative study of Meissner’s corpuscles in man. Neurology 16:1, 1966. 8. Brown, M. J., Martin, J. R., and Asbury, A. K.: Painful diabetic neuropathy: a morphometric study. Arch. Neurol. 33:164, 1976. 9. Campuzano, V., Montermini, L., Molto, M. D., et al.: Friedreich’s ataxia: autosomal recessive disease caused by an intronic GAA triplet repeat expansion. Science 271:1423, 1996. 10. Caterina, M. J., Leffler, A., Malmberg, A. B., et al.: Impaired nociception and pain sensation in mice lacking the capsaicin receptor. Science 288:306, 2000. 11. Caterina, M. J., Rosen, T. A., Tominaga, M., et al.: A capsaicin-receptor homologue with a high threshold for noxious heat. Nature 398:436, 1999. 12. Cauna, N.: The free penicillate nerve endings of the human hairy skin. J. Anat. 115:277, 1973. 13. Chen, C. H., Vazquez-Padua, M., and Cheng, Y. C.: Effect of anti-human immunodeficiency virus nucleoside analogs on mitochondrial DNA and its implication for delayed toxicity. Mol. Pharmacol. 39:625, 1991. 14. Cherry, C. L., Gahan, M. E., McArthur, J. C., et al.: Exposure to dideoxynucleosides is reflected in lowered mitochondrial DNA in subcutaneous fat. J. Acquir. Immune Defic. Syndr. 30:271, 2002. 15. Chiang, M. C., Lin, Y. H., Pan, C. L., et al.: Cutaneous innervation in chronic inflammatory demyelinating polyneuropathy. Neurology 59:1094, 2002. 16. Chien, H. F., Tseng, T. J., Lin, W. M., et al.: Quantitative pathology of cutaneous nerve terminal degeneration in the human skin. Acta Neuropathol. (Berl.) 102:455, 2001. 17. Cornblath, D. R., and McArthur, J. C.: Predominantly sensory neuropathy in patients with AIDS and AIDS-related complex. Neurology 38:794, 1988. 18. Dalsgaard, C. J., Rydh, M., and Haegerstrand, A.: Cutaneous innervation in man visualized with protein gene product 9.5 (PGP 9.5) antibodies. Histochemistry 92:385, 1989. 19. Dyck, P. J.: Neuronal atrophy and degeneration predominantly affecting peripheral sensory and autonomic neurons. In Dyck, P. J., Thomas, P. K., Griffin, J. W., et al. (eds.): Peripheral Neuropathy, 3rd ed. Philadelphia, W. B. Saunders, p. 1065, 1993. 20. Dyck, P. J., Winkelmann, R. K., and Bolton, C. F.: Quantitation of Meissner’s corpuscles in hereditary neurologic disorders: Charcot-Marie-Tooth disease, Roussy-Levy syndrome, Dejerine-Sottas disease, hereditary sensory neuropathy, spinocerebellar degenerations, and hereditary spastic paraplegia. Neurology 16:10, 1966. 21. Ekbom, K.: Restless legs. Acta Med. Scand. Suppl. 158:1, 1945. 22. Fundin, B. T., Arvidsson, J., Aldskogius, H., et al.: Comprehensive immunofluorescence and lectin binding analysis of intervibrissal fur innervation in the mystacial pad of the rat. J. Comp. Neurol. 385:185, 1997. 23. Gahan, M. E., Miller, F., Lewin, S. R., et al.: Quantification of mitochondrial DNA in peripheral blood mononuclear cells and subcutaneous fat using real-time polymerase chain reaction. J. Clin. Virol. 22:241, 2001.

24. Goebel, H. H.: Extracerebral biopsies in neurodegenerative diseases of childhood. Brain Dev. 21:435, 1999. 25. Goebel, H. H., Veit, S., and Dyck, P. J.: Confirmation of virtual unmyelinated fiber absence in hereditary sensory neuropathy type IV. J. Neuropathol. Exp. Neurol. 39:670, 1980. 26. Gold, M. S., Reichling, D. B., Shuster, M. J., and Levine, J. D.: Hyperalgesic agents increase a tetrodotoxin-resistant Na current in nociceptors. Proc. Natl. Acad. Sci. U. S. A. 93:1108, 1996. 27. Griffin, J. W., McArthur, J. C., and Polydefkis, M.: Assessment of cutaneous innervation by skin biopsies. Curr. Opin. Neurol. 14:655, 2001. 28. Halata, Z.: Sensory innervation of the hairy skin (light- and electronmicroscopic study). J. Invest. Dermatol. 101:75S, 1993. 29. He, C. L., Wendelschafer-Crabb, G., and Kennedy, W. R.: Reinnervation of epidermal nerve fibers in pig skin. Soc. Neurosci. Abstr. 22:760, 1996. 30. Helm, T., and Helm, K. F.: CADASIL: blood test versus skin biopsy? J. Am. Acad. Dermatol. 46:798, 2002. 31. Herrmann, D. N., Griffin, J. W., Hauer, P., et al.: Epidermal nerve fiber density and sural nerve morphometry in peripheral neuropathies. Neurology 53:1634, 1999. 32. Hilliges, M., and Johansson, O.: Comparative analysis of numerical estimation methods of epithelial nerve fibers using tissue sections. J. Peripher. Nerv. Syst. 4:53, 1999. 33. Hilz, M. J., Kennedy, W. R., Stemper, B., et al.: Cutaneous innervation in familial dysautonomia. Clin. Auton. Res. 11:182, 2001. 34. Holland, N. R., Crawford, T. O., Hauer, P., et al.: Smallfiber sensory neuropathies: clinical course and neuropathology of idiopathic cases. Ann. Neurol. 44:47, 1998. 35. Holland, N. R., Stocks, A., Hauer, P., et al.: Intraepidermal nerve fiber density in patients with painful sensory neuropathy. Neurology 48:708, 1997. 36. Hsieh, S. T., Choi, S., Lin, W. M., et al.: Epidermal denervation and its effects on keratinocytes and Langerhans cells. J. Neurocytol. 25:513, 1996. 37. Iannaccone, S., Zucconi, M., Marchettini, P., et al.: Evidence of peripheral axonal neuropathy in primary restless legs syndrome. Mov. Disord. 10:2, 1995. 38. Indo, Y., Tsuruta, M., Hayashida, Y., et al.: Mutations in the TRKA/NGF receptor gene in patients with congenital insensitivity to pain with anhidrosis. Nat. Genet. 13:485, 1996. 39. Johansson, O., Han, S. W., and Enhamre, A.: Altered cutaneous innervation in psoriatic skin as revealed by PGP 9.5 immunohistochemistry. Arch. Dermatol. Res. 283:519, 1991. 40. Johansson, O., Wang, L., Hilliges, M., and Liang, Y.: Intraepidermal nerves in human skin: PGP 9.5 immunohistochemistry with special reference to the nerve density in skin from different body regions. J. Peripher. Nerv. Syst. 4:43, 1999. 41. Karanth, S. S., Springall, D. R., Lucas, S., et al.: Changes in nerves and neuropeptides in skin from 100 leprosy patients investigated by immunocytochemistry. J. Pathol. 157:15, 1989.

Pathology and Quantitation of Cutaneous Innervation 42. Kennedy, W. R., Brown, J., and Wendelschafer-Crabb, G.: Sensory reinnervation in diabetic neuropathy. J. Peripher. Nerv. Syst. 6:150, 2001. 43. Kennedy, W. R., Nolano, M., Wendelschafer-Crabb, G., et al.: A skin blister method to study epidermal nerves in peripheral nerve disease. Muscle Nerve 22:360, 1999. 44. Kennedy, W. R., and Sakuta, M.: Collateral reinnervation of sweat glands. Ann. Neurol. 15:73, 1984. 45. Kennedy, W. R., and Wendelschafer-Crabb, G.: The innervation of human epidermis. J. Neurol. Sci. 115:184, 1993. 46. Kennedy, W. R., and Wendelschafer-Crabb, G.: Utility of skin biopsy in diabetic neuropathy. Semin. Neurol. 16:163, 1996. 47. Kennedy, W.R., Wendelschafer-Crabb, G. and Johnson, T.: Quantitation of epidermal nerves in diabetic neuropathy. Neurology 47:1042, 1996. 48. Kennedy, W. R., Wendelschafer-Crabb, G., and Walk, D.: Use of skin biopsy and skin blister in neurological practice. J. Clin. Neuromuscul. Dis. 1:196, 2000. 49. Kennedy, W .R., Wendelschafer-Crabb, G., and Lindall, A. W.: Quantitative epidermal nerve fiber analysis in skin blisters and skin biopsies. Ann. Neurol. 14:573, 1999. 50. Khalili, N., Wendelschafer-Crabb, G., Kennedy, W. R., and Simone, D. A.: Influence of thermode size for detecting heat pain dysfunction in a capsaicin model of epidermal nerve fiber loss. Pain 91:241, 2001. 51. Kuhlenbaumer, G., Young, P., Oberwittler, C., et al.: Giant axonal neuropathy (GAN): case report and two novel mutations in the gigaxonin gene. Neurology 58:1273, 2002. 52. Lauria, G.: Innervation of the human epidermis: a historical review. Ital. J. Neurol. Sci. 20:63, 1999. 53. Lauria, G., McArthur, J. C., Hauer, P. E., et al.: Neuropathological alterations in diabetic truncal neuropathy: evaluation by skin biopsy. J. Neurol. Neurosurg. Psychiatry 65:762, 1998. 54. Lauria, G., Sghirlanzoni, A., Lombardi, R., and Pareyson, D.: Epidermal nerve fiber density in sensory ganglionopathies: clinical and neurophysiologic correlations. Muscle Nerve 24:1034, 2001. 55. Levy, D. M., Terenghi, G., Gu, X. H., et al.: Immunohistochemical measurements of nerves and neuropeptides in diabetic skin: relationship to tests of neurological function. Diabetologia 35:889, 1992. 56. Li, Y., Hsieh, S. T., Chien, H. F., et al.: Sensory and motor denervation influence epidermal thickness in rat foot glabrous skin. Exp. Neurol. 147:452, 1997. 57. Lichtman, J. W., Magrassi, L., and Purves, D.: Visualization of neuromuscular junctions over periods of several months in living mice. J. Neurosci. 7:1215, 1987. 58. Low, P. A., Caskey, P. E., Tuck, R. R., et al.: Quantitative sudomotor axon reflex test in normal and neuropathic subjects. Ann. Neurol. 14:573, 1983. 59. Mahadevan, A., Santosh, V., Gayatri, N., et al.: Infantile neuroaxonal dystrophy and giant axonal neuropathy—overlap diseases of neuronal cytoskeletal elements in childhood? Clin. Neuropathol. 19:221, 2000. 60. Martin, J. L., Brown, C. E., Matthews-Davis, N., and Reardon, J. E.: Effects of antiviral nucleoside analogs on

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73. 74.

75.

76.

77.

893

human DNA polymerases and mitochondrial DNA synthesis. Antimicrob. Agents Chemother. 38:2743, 1994. McArthur, J. C., Stocks, E. A., Hauer, P., et al.: Epidermal nerve fiber density: normative reference range and diagnostic efficiency. Arch. Neurol. 55:1513, 1998. McArthur, J. C., Yiannoutsos, C., Simpson, D. M., et al.: A Phase II trial of nerve growth factor for sensory neuropathy associated with HIV infection. AIDS Clinical Trials Group Team 291. Neurology 54:1080, 2000. McCarthy, B. G., Hsieh, S. T., Stocks, A., et al.: Cutaneous innervation in sensory neuropathies: evaluation by skin biopsy. Neurology 45:1848, 1995. Molina, F. A., Burrows, N. P., Jones, R. R., et al.: Increased sensory neuropeptides in nodular prurigo: a quantitative immunohistochemical analysis. Br. J. Dermatol. 127:344, 1992. Molliver, D. C., Wright, D. E., Leitner, M. L., et al.: IB4binding DRG neurons switch from NGF to GDNF dependence in early postnatal life. Neuron 19:849, 1997. Moore, R. D., Wong, W. M., Keruly, J. C., and McArthur, J. C.: Incidence of neuropathy in HIV-infected patients on monotherapy versus those on combination therapy with didanosine, stavudine and hydroxyurea. AIDS 14:273, 2000. Navarro, X., Verdu, E., Wendelschafer-Crabb, G., and Kennedy, W. R.: Innervation of cutaneous structures in the mouse hind paw: a confocal microscopy immunohistochemical study. J. Neurosci. Res. 41:111, 1995. Navarro, X., Verdu, E., Wendelschafer-Crabb, G., and Kennedy, W. R.: Immunohistochemical study of skin reinnervation by regenerative axons. J. Comp. Neurol. 380:164, 1997. Nolano, M., Crisci, C., Santoro, L., et al.: Absent innervation of skin and sweat glands in congenital insensitivity to pain with anhidrosis. Clin. Neurophysiol. 111:1596, 2000. Nolano, M., Provitera, V., Crisci, C., et al.: Small fibers involvement in Friedreich’s ataxia. Ann. Neurol. 50:17, 2001. Nolano, M., Provitera, V., Lullo, F., et al.: Tactile stimulation and mechanoreceptors in sensory neuropathies. Neurol. Sci. 22(Suppl. 1):31, 2001. Nolano, M., Simone, D. A., Wendelschafer-Crabb, G., et al.: Topical capsaicin in humans: parallel loss of epidermal nerve fibers and pain sensation. Pain 81:135, 1999. Novak, V., Freimer, M. L., Kissel, J. T., et al.: Autonomic impairment in painful neuropathy. Neurology 56:861, 2001. Novella, S. P., Inzucchi, S. E. and Goldstein, J. M.: The frequency of undiagnosed diabetes and impaired glucose tolerance in patients with idiopathic sensory neuropathy. Muscle Nerve 24:1229, 2001. Oaklander, A. L.: The density of remaining nerve endings in human skin with and without postherpetic neuralgia after shingles. Pain 92:139, 2001. Oaklander, A. L., Romans, K., Horasek, S., et al.: Unilateral postherpetic neuralgia is associated with bilateral sensory neuron damage. Ann. Neurol. 44:789, 1998. Ochoa, J. L., and Yarnitsky, D.: The triple cold syndrome: cold hyperalgesia, cold hypoaesthesia and cold skin in peripheral nerve disease. Brain 117(Pt. 1):185, 1994.

894

Pathology of the Peripheral Nervous System

78. Periquet, M. I., Novak, V., Collins, M. P., et al.: Painful sensory neuropathy: prospective evaluation using skin biopsy. Neurology 53:1641, 1999. 79. Polydefkis, M., Allen, R. P., Hauer, P., et al.: Subclinical sensory neuropathy in late-onset restless legs syndrome. Neurology 55:1115, 2000. 80. Polydefkis, M., Hauer, P., Abraham, B., et al.: Novel measures of human axonal regeneration. J. Peripher. Nerv. Syst. 6:170, 2001. 81. Polydefkis, M., Hauer, P., Griffin, J. W., and McArthur, J. C.: Skin biopsy as a tool to assess distal small fiber innervation in diabetic neuropathy. Diabetes Technol. Ther. 3:23, 2001. 81a. Polydefkis, M., Hauer, P., Sheth, S., et al.: The time course of epidermal nerve fiber regeneration: Studies in normal controls and in people with diabetes, with and without neuropathy. Brain 127:1606, 2004. 82. Polydefkis, M., Yiannoutsos, C. T., Cohen, B. A., et al.: Reduced intraepidermal nerve fiber density in HIV-associated sensory neuropathy. Neurology 58:115, 2002. 83. Rajan, B., Polydefkis, M., Hauer, P., et al.: Epidermal reinnervation after intracutaneous axotomy in man. J. Comp. Neurol. 457:24, 2003. 84. Ren, K., and Dubner, R.: Descending modulation in persistent pain: an update. Pain 100:1, 2002. 85. Rice, F. L., Fundin, B. T., Arvidsson, J., et al.: Comprehensive immunofluorescence and lectin binding analysis of vibrissal follicle sinus complex innervation in the mystacial pad of the rat. J. Comp. Neurol. 385:149, 1997. 86. Riley, C. M., Day, R. L., Greely, D., and Langford, W. S.: Central autonomic dysfunction with defective lacrimation. Pediatrics 3:468, 1949. 87. Rosenberg, S., Nagahashi Marie, S. K., and Kliemann, S.: Congenital insensitivity to pain with anhidrosis (hereditary sensory and autonomic neuropathy type IV). Pediatr. Neurol. 11:50, 1994. 88. Rowbotham, M. C., Yosipovitch, G., Connolly, M. K., et al.: Cutaneous innervation density in the allodynic form of postherpetic neuralgia. Neurobiol. Dis. 3:205, 1996. 89. Rydh, M., Malm, M., Jernbeck, J., and Dalsgaard, C. J.: Ectatic blood vessels in port-wine stains lack innervation: possible role in pathogenesis. Plast. Reconstr. Surg. 87:419, 1991. 90. Schifitto, G., McDermott, M. P., McArthur, J. C., et al.: Incidence of and risk factors for HIV-associated distal sensory polyneuropathy. Neurology 58:1764, 2002. 91. Schondorf, R., and Low, P. A.: Idiopathic postural orthostatic tachycardia syndrome: an attenuated form of acute pandysautonomia? Neurology 43:132, 1993. 92. Scott, L. J., Griffin, J. W., Luciano, C., et al.: Quantitative analysis of epidermal innervation in Fabry disease. Neurology 52:1249, 1999. 93. Shelley, W. B., and Arthur, R. P.: The neurohistology and neurophysiology of the itch sensation in man. Arch. Dermatol. Res. 76:296, 1957. 94. Simone, D. A., Nolano, M., Johnson, T., et al.: Intradermal injection of capsaicin in humans produces degeneration and subsequent reinnervation of epidermal nerve fibers: correlation with sensory function. J. Neurosci. 18:8947, 1998. 95. Simone, D. A., and Ochoa, J.: Early and late effects of prolonged topical capsaicin on cutaneous sensibility and neurogenic vasodilatation in humans. Pain 47:285, 1991.

96. Singer, W., Spies, J. M., Hauer, P., et al.: Epidermal nerve fiber density in the neuropathic type of the postural tachycardia syndrome [S35.001]. Presented at the annual meeting of the American Academy of Neurology, San Diego, 2000. 97. Singleton, J. R., Smith, A. G., and Bromberg, M. B.: Increased prevalence of impaired glucose tolerance in patients with painful sensory neuropathy. Diabetes Care 24:1448, 2001. 98. Smith, A. G., Kroll, R., Ramachandran, P., et al.: The reliability of skin biopsy with measurement of intraepidermal nerve fiber density [A500]. Neurology 58(Suppl. 3):S66, 2002. 99. Smith, A. G., Ramachandran, P., Tripp, S., and Singleton, J. R.: Epidermal nerve innervation in impaired glucose tolerance and diabetes-associated neuropathy. Neurology 57:1701, 2001. 100. Springall, D. R., Karanth, S. S., Kirkham, N., et al.: Symptoms of notalgia paresthetica may be explained by increased dermal innervation. J. Invest. Dermatol. 97:555, 1991. 101. Stocks, E. A., McArthur, J. C., Griffen, J. W., and Mouton, P. R.: An unbiased method for estimation of total epidermal nerve fibre length. J. Neurocytol. 25:637, 1996. 102. Suarez, G. A., Opfer-Gehrking, T. L., Offord, K. P., et al.: The Autonomic Symptom Profile: a new instrument to assess autonomic symptoms. Neurology 52:523, 1999. 103. Sumner, C. J., Sheth, S., Griffin, J. W., et al.: The spectrum of neuropathy in diabetes and impaired glucose tolerance. Neurology 60:108, 2003. 104. Taratuto, A. L., Sevlever, G., Saccoliti, M., et al.: Giant axonal neuropathy (GAN): an immunohistochemical and ultrastructural study report of a Latin American case. Acta Neuropathol. (Berl.) 80:680, 1990. 105. Terenghi, G., Bunker, C. B., Liu, Y. F., et al.: Image analysis quantification of peptide-immunoreactive nerves in the skin of patients with Raynaud’s phenomenon and systemic sclerosis. J. Pathol. 164:245, 1991. 106. Thompson, R. J., Doran, J. F., Jackson, P., et al.: PGP 9.5—a new marker for vertebrate neurons and neuroendocrine cells. Brain Res. 278:224, 1983. 107. Underwood, R. A., Gibran, N. S., Muffley, L. A., et al.: Color subtractive-computer-assisted image analysis for quantification of cutaneous nerves in a diabetic mouse model. J. Histochem. Cytochem. 49:1285, 2001. 108. Urashima, R., and Mihara, M.: Cutaneous nerves in atopic dermatitis: a histological, immunohistochemical and electron microscopic study. Virchows Arch. 432:363, 1998. 109. Walsh, J. S., Perniciaro, C., and Meschia, J. F.: CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy): diagnostic skin biopsy changes determined by electron microscopy. J. Am. Acad. Dermatol. 43:1125, 2000. 110. Wang, L., Hilliges, M., Jernberg, T., et al.: Protein gene product 9.5-immunoreactive nerve fibres and cells in human skin. Cell Tissue Res. 261:25, 1990. 111. Watkins, L. R., and Maier, S. F.: Beyond neurons: evidence that immune and glial cells contribute to pathological pain states. Physiol. Rev. 82:981, 2002. 112. Wendelschafer-Crabb, G., Nolano, M., and Kennedy, W. R.: Sensory reinnervation of a skin blister wound. J. Peripher. Nerv. Syst. 6:150, 2001.

Pathology and Quantitation of Cutaneous Innervation 113. Wood, J. N., Winter, J., James, I. F., et al.: Capsaicininduced ion fluxes in dorsal root ganglion cells in culture. J. Neurosci. 8:3208, 1988. 114. Woolf, C. J., Shortland, P., and Coggeshall, R. E.: Peripheral nerve injury triggers central sprouting of myelinated afferents. Nature 355:75, 1992. 115. Wu, G., Ringkamp, M., Hartke, T. V., et al.: Early onset of spontaneous activity in uninjured C-fiber nociceptors after injury to neighboring nerve fibers. J. Neurosci. 21:RC140, 2001.

895

116. Wu, G., Ringkamp, M., Murinson, B. B., et al.: Degeneration of myelinated efferent fibers induces spontaneous activity in uninjured C-fiber afferents. J. Neurosci. 22:7746, 2002. 117. Zelickson, B. D., Kilmer, S. L., Bernstein, E., et al.: Pulsed dye laser therapy for sun damaged skin. Lasers Surg. Med. 25:229, 1999. 118. Zelickson, B. D., Mehregan, D. A., Wendelschafer-Crabb, G., et al.: Clinical and histologic evaluation of psoriatic plaques treated with a flashlamp pulsed dye laser. J. Am. Acad. Dermatol. 35:64, 1996.

35 Nerve Conduction and Needle Electromyography JUN KIMURA

Nerve Conduction Studies Fundamentals of Nerve Conduction Studies Basic Principles Motor Nerve Conduction Studies Sensory Nerve Conduction Studies Nerve Conduction in the Clinical Domain Commonly Assessed Nerves Cranial Nerves Major Nerves in the Upper Limb Nerves of the Shoulder Girdle Cutaneous Nerves in the Forearm Major Nerves in the Lower Limb Nerves of the Pelvic Girdle Human Reflexes and Late Responses Blink Reflex H and T Reflexes Masseter Reflex F Waves A Wave Clinical Applications Common Sources of Error

Collision Technique Anomalous Innervations Temporal Dispersion and Phase Cancellation Practical Assessment of Conduction Block: Criteria for Conduction Block Long and Short Segmental Nerve Conduction Studies

Needle Electromyography Potentials Associated with Needle Insertion Injury Potentials End-Plate Activities Myotonic Discharge Spontaneous Activity Fibrillation Potentials Positive Sharp Waves Single-Fiber Discharge and Denervation Complex Repetitive Discharges Fasciculation Potentials and Myokymic Discharges

Nerve Conduction Studies FUNDAMENTALS OF NERVE CONDUCTION STUDIES Basic Principles Nerve conduction studies play an important role in the assessment of peripheral nerve function. They are based on the simple principle, that electrical stimulation initiates nerve impulses, which propagate along motor or sensory nerve fibers. Standardization of both stimulating and

Continuous Muscle Fiber Activity Cramp Motor Unit Action Potential Amplitude and Area Rise Time Duration Phases Abnormalities of Motor Unit Potentials Lower Motor Neuron Versus Myopathic Disorders Discharge Pattern of Motor Units Recruitment Measurements of Turns and Amplitude Lower and Upper Motor Neuron Disorders Myopathy Involuntary Movement Other Aspects of Electromyography Principles of Localization Sequence of Abnormalities Summary

recording events are needed to get accurate and reproducible results and to avoid pitfalls. The study of the motor fibers depends on the analysis of compound evoked potentials recorded from the muscle. The function of the sensory fibers can be tested by recording sensory action potentials from the nerve itself. Nerve conduction studies have become a reliable test to assess peripheral nerve dysfunction and serve as a natural extension of the clinical evaluation. If conducted in the proper clinical context, the method provides a means of confirming the clinical diagnosis, localizing and characterizing the lesion, and quantifying the degree of involvement. 899

900

Nerve Conduction and Electromyography

Stimulus Intensity and Possible Risk To fully activate a healthy nerve, sufficient stimulation must be provided. Generally, it takes surface stimulation of 0.1 ms duration and 100 to 300 V or, assuming a tissue resistance of 10 k, 10 to 30 mA. However, as much as 400 to 500 V, or 40 to 50 mA, may be necessary to study diseased nerves with decreased excitability. Electrical stimulation within this intensity range ordinarily poses no particular risk to the patient. Routine nerve conduction studies, however, may be contraindicated in certain situations. For example, the entire current may reach the cardiac tissue in patients with indwelling catheters or central venous pressure lines inserted directly into the heart. Similarly, electric stimulation near the heart poses safety hazards in patients with cardioverters and defibrillators. Placing the stimulator at least 6 inches away reduces the chance of externally triggering the sensors.272 If necessary, a cardiologist with special expertise in electrophysiology can provide advice regarding the feasibility of a nerve conduction study in these patients, and the possible need to turn off such medical instruments during the procedure. Surface and Needle Electrodes Surface electrodes, in general, are better than needle electrodes for recording a compound muscle action potential, because they register the total contributions from all discharging motor units. The onset latency of an evoked potential recorded with surface electrodes refers to the conduction time of the fastest motor fibers, and the amplitude provides a measure of the number of axons innervating the muscle. In contrast, a needle electrode registers activity from only a small portion of the muscle, but may record action potentials from atrophic muscles when surface recording fails. The use of a needle electrode also helps to selectively record an action potential solely from a target muscle when proximal stimulation excites many other muscles simultaneously. An averaging technique, though usually unnecessary for recording muscle potentials, may help to identify markedly reduced responses from small atrophic muscles.14 Most electromyographers use surface electrodes for recording sensory and mixed nerve action potentials. A pair of ring electrodes placed over the proximal and distal interphalangeal joints is used to record antidromic sensory potentials. Single stimuli give rise to sensory potentials of sufficient amplitude for studying commonly tested nerves, thereby eliminating the need for averaging techniques. Some electromyographers prefer to use needle electrodes to increase the amplitude of the recorded sensory nerve potential and to decrease the noise from surrounding tissue. The combined effect of increased amplitude and decreased noise may enhance the signal-to-noise ratio by as much as five times,312 and, when averaging, markedly reduce the number of trials required to obtain the same resolution.

Averaging Technique Very small signals fall within the expected noise level. Interposing a step-up transformer between the recording electrodes and the amplifier improves the signal-to-noise ratio. Averaging techniques also help to identify small signals that may otherwise escape detection. For this purpose, the electronic devices now in use sum consecutive samples of waveforms in each sweep digitally and divide this by the number of trials. The voltage from noise that has no temporal relationship to stimulation will average close to zero with this technique, whereas signals timelocked to a stimulus will appear at a constant latency, distinct from the background noise. In one study testing the limit of an averaging technique, a sensory nerve action potential could be recorded against the background noise up to 50 times, but not 100 times, the signal.273

Motor Nerve Conduction Studies Waveform, Amplitude, and Duration A single shock applied to the nerve activates a group of motor units that summate slightly asynchronously because individual nerve axons vary in conduction velocity and in terminal length. Thus a compound muscle action potential consists of the temporally dispersed sum of many motor unit action potentials located within the recording radius of the electrode, a distance of approximately 20 mm.15 Motor units closer to the surface of the skin contribute more to the compound action potential than distant units, although their spatial orientation to the recording electrode also plays an important role.29,252,373 In belly-tendon recording (Fig. 35–1) with an active lead (E1) placed on the belly of the muscle and an indifferent lead (E2) on the tendon, muscle action potentials show an initially negative biphasic waveform. The usual measurements taken include (1) latency, the time from the stimulus to the onset of the negative response; (2) amplitude, the distance from the baseline to the negative peak or between negative and positive peaks; and (3) duration, the time from onset to the negative or positive peak or to the final return to the baseline. The area under the waveform measured by electronic integration correlates linearly with the product of the amplitude and duration. In general, onset latency provides a measure of the fast-conducting motor fibers. The shortest axons, however, may give rise to the initial potential, although they are not necessarily the fastest fibers. Latency and Conduction Velocity The latency of an action potential consists of three components: (1) nerve activation, the time from a stimulus to the generation of an action potential; (2) nerve conduction, the time from the activation point to the nerve terminal; and (3) neuromuscular transmission, the time from the axon terminal to the motor end plate, including the time required for generation of muscle action potential. The

Nerve Conduction and Needle Electromyography

901

FIGURE 35–1 A, Motor and sensory conduction studies of the median nerve. The photo shows stimulation at the wrist, 3 cm proximal to the distal crease, and recording over the belly (E1) and tendon (E2) of the abductor pollicis brevis for motor conduction, and around the proximal (E1) and distal (E2) interphalangeal joints of the second digit for antidromic sensory conduction. The ground electrode is located in the palm. B, Alternative recording sites for sensory conduction study of the median nerve with the ring electrodes placed around the proximal (E1) and distal (E2) interphalangeal joints of the third digit, or the base (E1) and the interphalangeal joint (E2) of the first digit. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

latency difference between two stimulation points, in effect, excludes components common to both stimuli, yielding the conduction time from one point of stimulation to the next (Fig. 35–2). Dividing the distance between these two stimuli by the latency difference, the motor nerve conduction velocity (MNCV) can be calculated by MNCV(m/s) 

D (mm) Lp  Ld(ms)

where D is the distance between the two stimulus points in millimeters, and Lp and Ld are the proximal and distal latencies in milliseconds, respectively. The reliability of

nerve conduction studies depends on how accurately the latencies are measured. Additionally, the length of nerve segment used in the calculation is only an estimate based on surface distance. In segmental conduction studies, separation of successive points of stimulation by at least 10 cm improves the accuracy of surface measurement. However, for localized lesions such as a compressive neuropathy, the inclusion of longer unaffected segments dilutes the effect of focal slowing and lowers the sensitivity. In contrast, the use of incremental stimulation across multiple shorter segments in the setting of local compression may reveal abrupt changes in latency and waveform. This may help localize an abnormality that might otherwise be difficult to identify with conventional stimulation lengths.

902

Nerve Conduction and Electromyography

FIGURE 35–2 Compound muscle action potential recorded from the thenar eminence following the stimulation of the median nerve at the elbow. The nerve conduction time from the elbow to the wrist equals the latency difference between the two responses elicited by the distal and proximal sites of stimulation. The motor nerve conduction velocity (MNCV), calculated by dividing the surface distance between the stimulus points by the subtracted times, concerns the fastest fibers. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

Conduction velocity cannot be calculated for the most distal segment because the terminal latency includes the time needed for neuromuscular transmission. The use of a fixed distance or anatomic landmarks for electrode placement standardizes the recording technique for improved accuracy and reproducibility. Because of distal slowing of nerve conduction, the terminal latency (Ld) estimated from the terminal distance (D) and proximal conduction velocity (CV) as the ratio D/CV is slightly less than the measured latency (Ld). The difference between measured and estimated latencies (Ld  Ld) is referred to as the residual latency. It represents a measure of the conduction delay at the nerve terminal and at the neuromuscular junction.215 The ratio between the estimated and measured latency (Ld/Ld), the terminal latency index, also relates to distal conduction delay.329 In a study of the median nerve, for example, a patient with a measured distal latency (Ld) of 4.0 ms, a terminal distance (D) of 8 cm, and a forearm conduction velocity (CV) of 50 m/s would have a calculated terminal latency (Ld) of 1.6 ms (8 cm/50 m/s), a residual latency of 2.4 ms (4.0 ms  1.6 ms) and a terminal latency index of 0.4 (1.6 ms/4.0 ms). Types of Abnormalities Nerve conduction studies help distinguish between two types of abnormalities: axonal degeneration and demyelination. This distinction is sometimes less clear for the

assessment of a nerve as a whole, as opposed to individual nerve fibers. Depending on the type of pathology, stimulation proximal to the site of presumed lesion gives rise to three basic types of conduction abnormalities (Fig. 35–3): (1) reduced amplitude with a relatively normal latency, (2) increased latency with relatively normal amplitude, and (3) absent response. Reduced amplitude with normal latency indicates an incomplete nerve lesion, either neurapraxia or axonotmesis involving a portion of the axons. In both situations, a shock applied distal to the lesion within the first few days of injury elicits a normal response showing a larger amplitude than stimulation proximal to the lesion. Within several days of the injury, however, distal stimulation may distinguish the two (Fig. 35–4). Regarding axonotmesis, stimulation below the lesion several days after its occurrence elicits an equally small response as stimulation above the lesion; this is secondary to wallerian degeneration. Regarding neurapraxia, often in the setting of demyelinative lesions, stimulation distal to the lesion continues to give rise to a larger compound muscle action potential than does stimulation proximal to the lesion, indicating a conduction block (Fig. 35–5). Increased latency combined with relatively normal amplitude generally implies that a majority of the nerve fibers have segmental demyelination without conduction block. As shown in rabbits, incomplete proximal compressive lesions may also give rise to slowed conduction with a reduction in external fiber diameter distal to the site of constriction. In this situation, the time course of recovery

FIGURE 35–3 Three basic types of alteration in the compound muscle action potential occur after a presumed nerve injury distal to the site of stimulation: reduced amplitude with nearly normal latency (top), normal amplitude with substantially increased latency (middle), or absent response even with a shock of supramaximal intensity (bottom). (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

Nerve Conduction and Needle Electromyography

903

a loss of large anterior horn cells in myelopathies may also cause slowed motor conduction. Here, the motor conduction velocity may decrease to 70% of the mean normal value with diminution of amplitude to less than 10% of normal. Despite a loss of the amplitude, however, a conduction velocity reduced to less than 60% of the mean normal value suggests peripheral nerve disease, rather than myelopathy.220

Sensory Nerve Conduction Studies

FIGURE 35–4 Nerve excitability distal to the lesion in neurotmesis or substantial axonotmesis. Distal stimulation elicits a normal compound muscle action potential during the first few days after injury, even with a complete separation of the nerve. Unlike neurapraxia, wallerian degeneration subsequent to transection will render the distal nerve segment inexcitable in 3 or 4 days. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

suggests distal paranodal demyelination as the cause of conduction slowing along the distal nerve segment.13 Absent responses indicate a complete failure of nerve conduction across the presumed site of the lesion. This may occur after either nerve transection with total axonotmesis, or neurapraxia involving all the motor fibers. Regardless of etiology, nerve stimulation distal to the lesion during the first week after injury elicits an entirely normal muscle action potential. By the second week, however, distal stimulation can distinguish neurapraxic changes from axonal degeneration as described above: normal response in neurapraxia and absent response in neurotmesis (see Fig. 35–4). In cases of nerve fiber degeneration, serial electrophysiologic studies help delineate progressive recovery of the amplitude of evoked potentials concomitant with nerve regeneration (Fig. 35–6). Additionally, a prolonged latency or slowing of conduction may also result from axonal neuropathy with a loss of the fast-conducting fibers.103 A major reduction in amplitude to below 40% to 50% of the mean normal value usually accompanies this type of slowing. If the amplitude remains above 80% of the control value in the absence of demyelination, the conduction velocity should not fall below 80% of the lower limit of normal. With further diminution of the amplitude to less than half the mean normal value, the conduction velocity may fall to 70% of the lower limit solely on the basis of axonal loss. Similarly,

Waveform, Amplitude, and Duration During sensory conduction studies in the upper limbs, stimulation of the digital nerves elicits an orthodromic sensory potential, which can be recorded proximally from the nerve trunk. Alternatively, stimulation of the nerve trunk evokes an antidromic potential distally in the digit (see Fig. 35–1) and a mixed nerve potential proximally. For example, a shock applied to the ulnar or median nerve at the wrist gives rise to an antidromic potential in the corresponding digits and a mixed nerve potential along the nerve trunk at the elbow. For routine clinical recordings, noninvasive recording with surface electrodes provide reliable and reproducible results.5,104 The use of needle recording, however, may improve the signal-to-noise ratio, especially when assessing temporal dispersion.217,277 Here again, signal averaging may help elucidate small late components originating from demyelinated, remyelinated, or regenerated fibers.40,131 These potentials correspond to fibers approximately 4 m in diameter with an average conduction velocity of 15 m/s.332 A reduction in the minimum conduction velocity may constitute the sole abnormality in several neuropathies that otherwise remain diagnostically elusive.156 Recording from surface electrodes generally result in higher amplitudes for an antidromic potential from the digits than an orthodromic response from the nerve trunk. This is in part because the digital nerves lie nearer to the surface.43 A change in the position of the recording electrodes may alter the waveform of an orthodromic sensory nerve action potential.304,385 Placing an active electrode (E1) on the nerve and a reference electrode (E2) at a remote site generally produces an initially positive triphasic potential. A separate late phase may also appear in a temporally dispersed response recorded at a more proximal site. In addition, the placement of E2 near the nerve at a distance greater than 3 cm from E1 results in a tetraphasic potential with the addition of the final negative component.43 The amplitude of an initially negative diphasic antidromic sensory potential is measured either from the baseline to its negative peak or between the negative and positive peaks. The amplitude of the initially positive triphasic orthodromic sensory potential may also be measured either from the baseline to the negative peak or

904

Nerve Conduction and Electromyography

FIGURE 35–5 A 67-year-old man with an acute onset of footdrop following chemotherapy and radiation treatment of prostate cancer. Although epidural metastasis was suspected clinically because of backache, the nerve conduction studies revealed the evidence of a conduction block at the knee, indicating a compressive neuropathy. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

between the positive and negative peaks. Amplitudes vary substantially among subjects and to a lesser extent between the two sides of the same individual, whether recorded with surface or needle electrodes.43 Factors that influence the amplitude of a nerve action potential include the density of sensory innervation and the subject’s body mass index, the latter reflecting the depth of the nerve from the skin surface.44 Women tend to have greater sensory nerve action potentials than men,157,230 possibly because the nerves lie more superficially. Left-handed individuals often have greater median sensory potentials at the wrist on their right side, whereas right-handed individuals often have greater amplitudes on their left side.248 The duration of the orthodromic or antidromic sensory potential is usually measured from the initial deflection from the baseline to the intersection between the descending phase and the baseline. Alternative points of reference occasionally used include the initial positive or negative peak, the subsequent positive peak, and, though less definable, the point of return to the baseline. Latency and Conduction Velocity Stimulation of the nerve at a single site generally suffices for calculation of sensory conduction velocity because, unlike a motor latency, a sensory latency consists solely of

the conduction time to the recording point. The latency of activation, a fixed delay of about 0.15 ms at the stimulus site,216 produces a measured latency slightly larger and the calculated conduction velocity slightly slower than actuality. The use of a second stimulus flanking the nerve segment corrects for this error because the latency difference between the two responses excludes the nerve activation time common to both stimulations. The latency of an antidromic potential is measured to the onset of the negative peak, whereas the latency of an orthodromic sensory potential is measured to the peak of the initial positive potential. The measurement to the negative peak may also be used if technical problems make identification of the preceding smaller positive peak difficult, as may occur in diseased or damaged nerves. Types of Abnormalities In general, similar types of abnormalities are found in both motor and sensory nerves. Demyelination of sensory fibers typically causes substantial slowing of conduction. Axonal degeneration typically results in a reduction in amplitude of the compound nerve action potentials elicited by distal stimulation. The sural nerve potential serves as a sensitive measure for identifying a lengthdependent distal axonal polyneuropathy.4 In patients with

Nerve Conduction and Needle Electromyography

905

FIGURE 35–6 A, A 21⁄2-year-old boy with hypothermia-induced axonal polyneuropathy following prolonged exposure to severe freezing weather on a frigid winter night in Iowa. Compound muscle action potentials recorded over the thenar eminence after stimulation of the median nerve at the wrist (W), elbow (E), or axilla (A). The initial study on March 17, 1986, revealed no response on either side followed by progressive return in amplitude and latency, with full recovery by January 8, 1987. B, Compound muscle action potential recorded from the abductor hallucis after stimulation of the tibial nerve at the ankle (A) or knee (K). The studies on March 17 and May 7, 1986, revealed no response on either side, with full recovery by January 8, 1987. Figure continued on following page

neuropathy, the ratio of the sensory potential of the sural nerve to that of the radial nerve often falls below 0.40, with a normal ratio of 0.71.320 Sensory fibers generally degenerate only after a lesion involving the sensory ganglion or postganglionic axon (see Fig. 35–6). Thus nerve conduction studies of the distal sensory potential can often differentiate preganglionic root avulsion from plexopathy.134 Radicular lesions within the spinal canal, however, may also involve the ganglion or postganglionic portion of the root, affecting the digital nerve potential.176,234 Localization of the lesion therefore depends on the distribution of the sensory potential abnormalities. In addition, plexopathy tends to affect multiple digits, whereas radiculopathy

usually reveals a selective abnormality of specific digits related to the root involved; the first digit by C6, the second and third digits by C7, and the fourth and fifth digits by C8.105

Nerve Conduction in the Clinical Domain Physiologic Variation among Different Nerve Segments Both motor and sensory fibers conduct slower in the legs than in the arms, with differences ranging from 7 to 10 m/s.205,358 Available data also confirm a length-dependent change in the conduction velocity for peroneal and sural nerves, but

906

Nerve Conduction and Electromyography

FIGURE 35–6 Continued C, Motor conduction studies of the tibial nerve on May 17, 1986. Stimulation at the knee elicited no response in the intrinsic foot muscle on either side (top three tracings), but a small compound action potential was elicited in the gastrocnemius bilaterally (bottom tracing) as the result of early reinnervation. D, Antidromic sensory nerve action potential recorded from the second digit after stimulation of the median nerve at the wrist (W) or elbow (E). The studies on March 17 and May 7, 1986, showed no response on either side, with full recovery by January 8, 1987. (From Afifi, A. K., Kimura, J., and Bell, W. E.: Hypothermiainduced reversible polyneuropathy: electrophysiologic evidence of axonopathy. Pediatr. Neurol. 4:49, 1988, with permission.)

not for motor or sensory fibers of the median nerve.338 Factors that may account for this velocity gradient in the lower limbs include, in the absence of histologic proof, abrupt axonal tapering, progressive reduction in axonal diameter, a shorter internodal distance, and lower temperatures. Statistical analyses of conduction velocities show no difference between median and ulnar nerves or between tibial and peroneal nerves, with a high degree of symmetry between the two sides.28 The nerve impulse propagates faster in the proximal than in the distal nerve segments,135 as evidenced by the F-wave conduction velocity between the spinal cord and axilla, which clearly exceeds the conventionally derived conduction velocity between the elbow and wrist.91,181,196 Statistical analyses show no significant difference between the two proximal segments, spinal cord to axilla and axilla to elbow, probably indicating a nonlinear slowing over the most distal segment.181 Effects of Temperature The effects of lower temperatures include a slowdown of impulse propagation and, at the same time, augmentation in amplitude of nerve and muscle action potentials, as

demonstrated in the squid axon152 and in human studies.73,300 For example, distal latencies increase by 0.3 ms/degree centigrade for both median and ulnar nerves upon cooling the hand.49 These principles apply to normal as well as demyelinated fibers as a straightforward consequence of the temperature coefficients governing voltage-sensitive sodium (Na) and potassium (K) conductance. In particular, cold-induced slowing of Na channel opening accounts for the slowing of conduction, and a delay of its inactivation, for a given increase in amplitude. In demyelinated nerve fibers, conduction fails at the lesion site when the rising temperature reduces the already marginal action potential below the critical level by fast inactivation of Na channels.116,374 A temperature rise, however, also quickens activation of Na channels, leading to a faster conduction over a length of a fiber that is normal.321 Thus the latency and amplitude measures follow two completely separate effects induced by change in temperature. Studies conducted in a warm room with ambient temperature maintained between 21° and 23° C reduce this type of variability. In practice, the skin temperature is measured with a plate thermistor to maintain it above 32° C, if necessary, using an infrared heat lamp or prior

Nerve Conduction and Needle Electromyography

immersion of the limb in warm water for 30 minutes.115 Alternatively, addition of 5% of the calculated conduction velocity for each degree below 32° C theoretically normalizes the result. Such conversion factors established in a normal population, however, may not necessarily apply in patients with diseases of the peripheral nerve.11,79,275 Maturation and Aging The process of maturational myelination accompanies a rapid increase in nerve conduction velocities from roughly half the adult value in full-term infants357 to the adult range at age 3 to 5 years (Fig. 35–7). Conduction velocities of physiologically slower fibers also show a similar time course of maturation.145 One series (Table 35–1) showed a steep increase in conduction of the peroneal nerve in infancy as compared to a slower maturation of the median nerve during early childhood.125 Conduction velocities in premature infants had an even slower range, from 17 to 25 m/s for the ulnar nerve and from 14 to 28 m/s for the peroneal nerve.51 The values reported at 23 to 24 weeks of fetal life averaged roughly one-third those of newborns of normal gestational age.255,335 Studies based on the expected date of birth of premature infant showed a different time course of maturation for motor and proprioceptive conduction.23 Fetal nutrition may alter peripheral nerve function by influencing myelin formation.309,310 In later childhood and adolescence, from age 3 to 19 years, both motor and sensory conduction velocities change as a function of age and growth in length, showing a slight increase in the upper limb and slight decrease in the lower limb.223 Conduction velocities begin to decline after 30 to 40 years of age, but not more than 10 m/s by

FIGURE 35–7 Relation of age to conduction velocity of motor fibers in the ulnar nerve between the elbow and wrist. Velocities in normal young adults range from 47 to 73 m/s, with most values between 50 and 70 m/s. Ages plotted indicate month after expected birth date based on calculation from the first day of last menstruation. (From Thomas, J. E., and Lambert, E. H.: Ulnar nerve conduction velocity and H-reflex in infants and children. J. Appl. Physiol. 15:1, 1960, with permission.)

907

60 years353 or even by 80 years of age.274 The most distal branches, such as the interdigital nerves, may show agerelated degeneration earlier.229 In one study,250 a reduction in the mean conduction rate averaged 10% at 60 years of age (Table 35–2). Aging also causes a diminution in amplitude and changes in the shape of the evoked potential,101 especially when recorded across the common sites of compression.64,65 Preferential loss of the largest and fastest conducting motor units375 probably results in a gradual increase in latencies of the F wave and somatosensory evoked potentials with advancing age. Height and Other Factors Various anthropometric characteristics also influence nerve conduction measures.322,348 Sural, peroneal, and tibial nerve conduction velocities all have inverse correlation with height in normal subjects308 and in patients with diabetic neuropathy,124 although in one study366 the height-related changes of sural nerve conduction velocity fell within the experimental error of 2.3% expected from the method. Women have faster conduction velocity and greater amplitude for both motor and sensory studies than men.309 Most gender-related velocity differences resolve when adjusted by height, whereas amplitude differences persist despite such correction and may, at least in part, relate to volume conductor characteristics of the body mass. Clinical Values and Limitations Besides a broad classification, the pattern of nerve conduction abnormalities can often characterize the general nature of some specific clinical disorders. In hereditary demyelinating neuropathies, for example, typical features

908

Nerve Conduction and Electromyography

Table 35–1. Normal Motor Nerve Conduction Velocities (m/s) in Different Age Groups Age 0–1 wk 1 wk to 4 mo 4 mo to 1 yr 1–3 yr 3–8 yr 8–16 yr Adults

Ulnar

Median

Peroneal

32 (21–39) 42 (27–53) 49 (40–63) 59 (47–73) 66 (51–76) 68 (58–78) 63 (52–75)

29 (21–38) 34 (22–42) 40 (26–58) 50 (41–62) 58 (47–72) 64 (54–72) 63 (51–75)

29 (19–31) 36 (23–53) 48 (31–61) 54 (44–74) 57 (46–70) 57 (45–74) 56 (47–63)

From Gamstorp, I.: Normal conduction velocity of ulnar, median and peroneal nerves in infancy, childhood and adolescence. Acta Paediatr. Suppl. 146:68, 1963, with permission.

include diffuse abnormalities with little difference from one nerve to another in the same patient and among different members in the same family.236 The studies also show approximately equal involvement of different nerve fibers with a limited degree of temporal dispersion despite a considerably increased latency. This stands in contrast to acquired demyelination, wherein disproportionate involvement of certain segments of the nerve185,206 often gives rise to more asymmetrical abnormalities with temporal dispersion. The reverse, however, does not always hold, because acquired demyelination can also produce diffuse abnormalities reminiscent of hereditary induced patterns.

Demyelinating and axonal polyneuropathies show a characteristic pattern of distribution in sensory nerve conduction abnormalities. Thus a reduced median amplitude compared with the sural amplitude supports the diagnosis of a primary demyelination,27 whereas a reduced sural amplitude compared with the radial amplitude tends to suggest an axonal polyneuropathy.320

COMMONLY ASSESSED NERVES Cranial Nerves Facial Nerve A quantitative assessment of nerve excitability depends on the size of muscle action recorded after stimulation of the facial nerve just below the ear and anterior to the mastoid process378 or directly over the stylomastoid foramen.352 Table 35–3 summarizes the normal onset latencies in 78 subjects divided into different age groups.377 Other normal values in adults (mean  standard deviation [SD]) range from 3.4  0.8 ms to 4.0  0.5 ms.352 In assessing a proximal lesion, as in Bell’s palsy, the latency of the muscle action potential rarely reveals a clear-cut abnormality even with substantial axonal degeneration because the remaining axons conduct normally. In contrast, the amplitude of the direct response serves importantly in determining the degree of axonal loss, which in turn dictates the prognosis.

Table 35–2. Normal Sensory and Motor Nerve Conduction Velocities (m/s) in Different Age Groups* Age 10–35 Yr (30 Cases) Nerve Median nerve Digit–wrist Wrist–muscle Wrist–elbow Elbow–axilla Ulnar nerve Digit–wrist Wrist–muscle Wrist–elbow Elbow–axilla Common peroneal nerve Ankle–muscle Ankle–knee Posterior tibial nerve Ankle–muscle Ankle–knee H reflex, popliteal fossa

Sensory

Motor

67.5  4.7 67.7  4.4 70.4  4.8

Sensory

Motor

65.8  5.7 3.2  0.3† 59.3  3.5 65.9  5.0

64.7  3.9

65.8  3.1 70.4  3.4

64.8  3.8 69.1  4.3

53.0  5.9

4.3  0.9† 49.5  5.6 5.9  1.3† 45.5  3.8 71.0  4.0 27.9  2.2†

Sensory

Motor

62.8  5.4 66.2  3.6

3.5  0.2† 54.5  4.0 63.6  4.4

57.5  6.6

67.1  4.7 70.6  2.4

2.7  0.3† 57.8  2.1 63.3  2.0

50.4  1.0

4.8  0.5† 43.6  5.1

49.0  3.8

Age 51–80 Yr (18 Cases)

59.4  4.9 3.7  0.3† 55.9  2.6 65.1  4.2

66.5  3.4 2.7  0.3† 58.9  2.2 64.4  2.6

56.9  4.4

Age 36–50 Yr (16 Cases)

7.3  1.7† 42.9  4.9 64.0  2.1 28.2  1.5†

*Values are means  1 standard deviation. † Latency in milliseconds. From Mayer, R. F.: Nerve conduction studies in man. Neurology (Minneap.) 13:1021, 1963, with permission.

56.7  3.7 64.4  3.0

3.0  0.35† 53.3  3.2 59.9  0.7

46.1  4.0

4.6  0.6† 43.9  4.3

48.9  2.6

6.0  1.2† 41.8  5.1 60.4  5.0 32.0  2.1†

Nerve Conduction and Needle Electromyography

Table 35–3. Facial Nerve Latency in 78 Subjects Divided into Different Age Groups Age 0–1 mo 1–12 mo 1–2 yr 2–3 yr 3–4 yr 4–5 yr 5–7 yr 7–16 yr

Mean (ms)

909

series, normal latencies to the upper trapezius in 25 subjects, 10 to 60 years of age, varied from 1.8 to 3.0 ms.54

Range (ms)

10.1 7.0 5.1 3.9 3.7 4.1 3.9 4.0

6.4–12.0 5.0–10.0 3.5–6.3 3.8–4.5 3.4–4.0 3.5–5.0 3.2–5.0 3.0–5.0

From Waylonis, G. W., and Johnson, E. W.: Facial nerve conduction delay. Arch. Phys. Med. Rehabil. 45:539, 1964, with permission.

Accessory Nerve Surface or needle stimulation along the posterior border of the sternocleidomastoid muscle activates the accessory nerve, eliciting a compound muscle action potential of the trapezius. For recording, an active electrode (E1) is placed at the angle of the neck and shoulder, and a reference electrode (E2) over the tendon near the acromion process.297,298 In one

FIGURE 35–8 A, Motor nerve conduction study of the median nerve. The sites of stimulation include Erb’s point (A), axilla (B), elbow (C), wrist (D), and palm (E). Compound muscle action potentials are recorded with surface electrodes placed on the thenar eminence. B, Sensory nerve conduction study of the median nerve. The sites of stimulation include axilla (A), elbow (B), wrist (C), and palm (D). Antidromic sensory potentials are recorded with a pair of ring electrodes placed around the second digit. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

Major Nerves in the Upper Limb Median Nerve The usual sites of stimulation of the median nerve include Erb’s point and the axilla, elbow, wrist, and palm along the relatively superficial course of the nerve (Fig. 35–8). Its stimulation at Erb’s point, the axilla, and the palm tends to coactivate other nerves in close proximity.186,288,314 Table 35–4 summarizes normal values in our laboratory. Separate stimulation of the median and ulnar nerves at the wrist evokes corresponding muscle action potentials from the second lumbricalis and the first volar interosseous at nearly identical latency if the distance between the stimulating and recording electrodes is kept the same for both nerves.74,112,303,322,330,365 A latency difference greater than 0.4 to 0.5 ms suggests an abnormal delay over the distal segment as might be seen in carpal tunnel syndrome. Table 35–4 summarizes normal values for the digital potentials recorded with ring electrodes placed 2 cm apart

910

Nerve Conduction and Electromyography

Table 35–4. Normal Values for Median Nerve* Site of Stimulation

Amplitude†: Motor (mV), Sensory (V)

Motor Fibers Palm Wrist Elbow Axilla Sensory Fibers Digit Palm Wrist Elbow

Latency‡ to Recording Site (ms)

Difference between Right and Left (ms)

6.9  3.2 (3.5)§ 7.0  3.0 (3.5) 7.0  2.7 (3.5) 7.2  2.9 (3.5)

1.86  0.28 (2.4)储 3.49  0.34 (4.2) 7.39  0.69 (8.8) 9.81  0.89 (11.6)

0.19  0.17 (0.5)储 0.24  0.22 (0.7) 0.31  0.24 (0.8) 0.42  0.33 (1.1)

39.0  16.8 (20) 38.5  15.6 (19) 32.0  15.5 (16)

1.37  0.24 (1.9) 2.84  0.34 (3.5) 6.46  0.71 (7.9)

0.15  0.11 (0.4) 0.18  0.14 (0.5) 0.29  0.21 (0.7)

Conduction Time between Two Points (ms)

Conduction Velocity (ms)

1.65  0.25 (2.2)储 3.92  0.49 (4.9) 2.42  0.39 (3.2)

48.8  5.3 (38)¶ 57.7  4.9 (48) 63.5  6.2 (51)

1.37  0.24 (1.9) 1.48  0.18 (1.8) 3.61  0.48 (4.6)

58.8  5.8 (47) 56.2  5.8 (44) 61.9  4.2 (53)

*Mean  standard deviation (SD) in 61 patients, 11 to 74 years of age (average, 40), with no apparent disease of the peripheral nerves. † Amplitude measured from the baseline to the negative peak. ‡ Latency measured to the onset. § Lower limits of normal, based on the distribution of the normative data. 储 Upper limits of normal, calculated as the mean  2 SD. ¶ Lower limits of normal, calculated as the mean  2 SD. From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.

around the proximal (E1) and distal (E2) interphalangeal joints of the second digit. Serial stimulation from the midpalm to the distal forearm in 1-cm increments normally shows a predictable latency change of 0.16 to 0.20 ms/cm (Fig. 35–9). Focal abnormalities of the median nerve usually accompany a sharply localized latency increase across a 1-cm segment (Fig. 35–10) associated with an abrupt change in waveform from abnormal temporal dispersion. The use of multichannel electrodes allows simultaneous recording of nerve potentials at several sites across the wrist after stimulation of the median nerve at the digit189,327 or at the elbow.161 This technique offers instantaneous comparison of latencies but suffers from a major drawback in the assessment of amplitudes, which vary so much depending on the depth of the nerve at the site of recording.189 Separate stimulation of the median and ulnar nerves at the wrist evokes a corresponding sensory potential of the fourth digit at nearly the identical latency for the same conduction distance. In our series (Table 35–5), normal values for the median nerve consisted of an onset latency of 2.88  0.35 ms (mean  SD) after wrist stimulation and a distal amplitude of 37.6  17.2 V after palmar stimulation. The corresponding values for the ulnar nerve were 2.86  0.37 ms and 46.1  24.3 V. The latency difference between the two nerves ranged from 0.01  0.17 ms to 0.02  0.17 ms, with an upper limit of normal of 0.4 ms (mean  2 SD). Ulnar Nerve Routine motor conduction studies consist of stimulating the ulnar nerve at multiple sites along the superficial course of the nerve (Fig. 35–11) and recording the muscle

potential from the hypothenar muscle with a pair of surface electrodes placed over the belly of the abductor digiti minimi (E1) and its tendon (E2) 3 cm distally (Tables 35–6 and 35–7).7,8 For accurate calculation of conduction velocities, the distance between sites of stimulation above and below the elbow should exceed 10 cm to minimize measurement error. Conventional studies, however, may fail to uncover focal abnormalities in tardy ulnar palsy or a cubital tunnel syndrome, which induces an insignificant delay when assessed over a 10-cm segment. In contrast, an “inching” study across the elbow with stimulation in 1- to 2-cm increments often detects an abrupt change in latency and waveform at the site of localized compression.47,194 The conduction along the deep palmar branch of the ulnar nerve can be tested by stimulating the ulnar nerve at the wrist and recording the muscle potential from the first dorsal interosseous or adductor pollicis. In one series of 373 patients, the upper limit of the normal range included 4.5 ms for the distal latency to the first dorsal interosseous, 2.0 ms for the latency difference between this muscle and the abductor digiti minimi, and 1.3 ms for the latency difference between the two sides.280 Stimulation of the deep palmar branch distal to the site of lesion also evokes muscle responses as a good measure of surviving motor axons. Lumbrical-interosseous comparison as described for median nerve study is also useful in assessing a distal ulnar nerve lesion, which typically causes a latency difference greater than 0.4 ms in the reverse direction.212,331 The common sites of stimulation for an antidromic sensory study include above and below the elbow, 3 cm proximal to the distal crease at the wrist, and 5 cm distal

Nerve Conduction and Needle Electromyography

911

FIGURE 35–9 A, Twelve sites of stimulation in 1-cm increments along the length of the median nerve. The “0” level at the distal crease of the wrist corresponds to the origin of the transverse carpal ligament. The photo shows a recording arrangement for sensory nerve potentials from the second digit and muscle action potentials from the abductor pollicis brevis. B, Sensory nerve potentials in a normal subject recorded after stimulation of the median nerve at multiple points across the wrist. The numbers on the left indicate the site of each stimulus (compare to A). The latency increased linearly with stepwise shifts of stimulus site proximally in 1-cm increments. (From Kimura, J.: The carpal tunnel syndrome: localization of conduction abnormalities within the distal segment of the median nerve. Brain 102:619, 1979, with permission.)

to the crease in the palm (see Fig. 35–11) to make the studies comparable to those of the median nerve (see Fig. 35–8). Recording from either the fourth or fifth digit tests the integrity of the C8 and T1 roots, lower trunk, and medial cord. Alternative studies consists of stimulation of the digital nerve with ring electrodes placed around the interphalangeal joints of the fifth digit, cathode proximally, and recording orthodromic sensory potentials at various sites along the course of the nerve. The dorsal ulnar cutaneous nerve, like the ulnar nerve proper, derives from the C8 to T1 roots, the lower trunk, and the medial cord. It leaves the common trunk of the ulnar nerve 5 to 8 cm proximal to the ulnar styloid between the tendon of the flexor carpi ulnaris and the ulna.22,168 Surface stimulation at this point selectively stimulates the dorsal ulnar cutaneous nerve, which subserves the sensation over the dorsum of the hand. Its abnormality helps localize the lesion in the segment proximal to the takeoff of this branch, although anatomic variations may alter cutaneous innervation.296

Radial Nerve The radial nerve becomes accessible to surface stimulation at (1) Erb’s point in the supraclavicular fossa; (2) the axilla near the spinal groove between the coracobrachialis and medial edge of the triceps, about 18 cm proximal to the medial epicondyle; (3) above the elbow between the brachioradialis and the tendon of the biceps, 6 cm proximal to the lateral epicondyle; and (4) in the forearm between the extensor carpi ulnaris and extensor digiti minimi on the dorsal aspect of the ulna, 8 to 10 cm proximal to the styloid process (Fig. 35–12). Either needle or surface electrodes may be used to record muscle action potentials from the extensor digitorum communis387 or the extensor indicis. The sensory branch gives off the posterior antebrachial cutaneous nerve, which innervates the dorsolateral aspect of the forearm at the level of the elbow before emerging near the surface about 10 cm above the lateral styloid process.52 Surface stimulation at any site along the lateral edge of the radius up to 10 to 14 cm proximal to the base of the thumb activates the sensory fibers selectively.

912

Nerve Conduction and Electromyography

FIGURE 35–10 A, Sensory nerve potentials in a patient with the carpal tunnel syndrome. Both hands showed a sharply localized slowing from 2 to 1, with the calculated segmental conduction velocity of 14 m/s on the left (top) and 9 m/s on the right (bottom). Note a distinct change in waveform of the sensory potential at the point of localized conduction delay. The double-humped appearance at 2 on the left suggests sparing of some sensory axons at this level. B, Sensory nerve potential in a patient with the carpal tunnel syndrome. Temporally dispersed responses on the right at 1 and beyond had greater negative and positive peaks in this area compared to normal, more distal responses, presumably because of loss of physiologic phase cancellation. Both hands show a sharply localized slowing from 3 to 2, with a segmental conduction velocity of 10 m/s on the left (top) and 7 m/s on the right (bottom). An abrupt change in waveform of the sensory potential also indicates the point of localized conduction delay. C, Sensory nerve potential in a patient with the carpal tunnel syndrome before (top) and after surgery (bottom). Preoperative study showed a localized slowing from 4 to 3, with a calculated segmental conduction velocity of 8 m/s, which normalized in a repeat study 6 months postoperatively. (From Kimura, J.: The carpal tunnel syndrome: localization of conduction abnormalities within the distal segment of the median nerve. Brain 102:619, 1979, with permission.)

913

Nerve Conduction and Needle Electromyography

Table 35–5. Distal Sensory Conduction Study Comparing Median and Ulnar Nerves* Measurement of Antidromic Sensory Potential Recording

Stimulation

Amplitude† (V)

Latency‡ (ms)

Median nerve, 2nd digit Median nerve, 4th digit Ulnar nerve, 4th digit Median & ulnar difference

Palm Wrist Palm Wrist Palm Wrist Palm Wrist

49.8  21.5 (25)§ 38.4  15.6 (19) 37.6  17.2 (19) 22.3  8.2 (11) 46.1  24.3 (23) 29.0  14.8 (25) 8.5  20.7 5.9  10.1

1.43  0.16 (1.7)储 2.87  0.31 (3.5) 1.45  0.20 (1.9) 2.88  0.35 (3.6) 1.48  0.26 (2.0) 2.86  0.37 (3.6) 0.02  0.17 (0.3) 0.01  0.17 (0.4)

Calculated Values for Wrist to Palm Segment Conduction Time (ms)

Conduction Velocity (m/s)

1.44  0.20 (1.9)储

57.1  8.3 (40)¶

1.43  0.22 (1.9)

57.4  8.9 (40)

1.38  0.30 (1.8)

59.1  8.3 (43)

0.04  0.20 (0.4)

*Mean  standard deviation (SD) in 31 healthy subjects, 16 to 64 years of age (average, 38), with no apparent disease of the peripheral nerve. † Amplitude measured from the baseline in the negative peak. ‡ Latency measured to the onset with a standard distance of 8 cm between the stimulus sites at the wrist and palm. § Lower limits of normal, based on the distribution of the normative data. 储 Upper limits of normal, calculated as the mean  2 SD. ¶ Lower limits of normal, calculated as the mean  2 SD. From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.

FIGURE 35–11 A, Motor nerve conduction study of the ulnar nerve. The sites of stimulation include Erb’s point (A), axilla (B), above the elbow (C), elbow (D), below the elbow (E), and wrist (F). Compound muscle action potentials are recorded with surface electrodes placed on the hypothenar eminence. B, Sensory nerve conduction study of the ulnar nerve. The sites of stimulation include axilla (A), above the elbow (B), elbow (C), below the elbow (D), wrist (E), and palm (F). The tracings show antidromic sensory potentials recorded with the ring electrodes placed around the fifth digit. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

914

Nerve Conduction and Electromyography

Table 35–6. Normal Values for Ulnar Nerve* Site of Stimulation

Amplitude†: Motor (mV), Sensory (V)

Latency‡ to Recording Site (ms)

Difference between Right and Left (ms)

Conduction Time between Two Points (ms)

Conduction Velocity (m/s)

Motor Fibers Wrist Below elbow Above elbow Axilla

5.7  2.0 (2.8)§ 5.5  2.0 (2.7) 5.5  1.9 (2.7) 5.6  2.1 (2.7)

2.59  0.39 (3.4)储 6.10  0.69 (7.5) 8.04  0.76 (9.6) 9.90  0.91 (11.7)

0.28  0.27 (0.8)储 0.29  0.27 (0.8) 0.34  0.28 (0.9) 0.45  0.39 (1.2)

3.51  0.51 (4.5)储 1.94  0.37 (2.7) 1.88  0.35 (2.6)

58.7  5.1 (49)¶ 61.0  5.5 (50) 66.5  6.3 (54)

35.0  14.7 (18) 28.8  12.2 (15) 28.3  11.8 (14)

2.54  0.29 (3.1) 5.67  0.59 (6.9) 7.46  0.64 (8.7)

0.18  0.13 (0.4) 0.26  0.21 (0.5) 0.28  0.27 (0.8)

2.54  0.29 (3.1) 3.22  0.42 (4.1) 1.79  0.30 (2.4)

54.8  5.3 (44) 64.7  5.4 (53) 66.7  6.4 (54)

Sensory Fibers Digit Wrist Below elbow Above elbow

*Mean  standard deviation (SD) in 65 patients, 13 to 74 years of age (average, 39), with no apparent disease of the peripheral nerves. † Amplitude measured from baseline to the negative peak. ‡ Latency measured to the onset after stimulation of the ulnar with the cathode 3 cm above the distal crease in the wrist. § Lower limits of normal, based on the distribution of the normative data. 储 Upper limits of normal, calculated as the mean  2 SD. ¶ Lower limits of normal, calculated as the mean  2 SD. From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.

Recording arrangements for an antidromic sensory potential include a pair of ring electrodes placed around the thumb, with the disc electrode (E1) over the first web space or slightly more proximally in the snuffbox, and the reference electrode (E2) near the first dorsal interosseous or between the second and third metacarpals. The radial nerve subserves the sensation of the thumb by fibers originating from the C6 and C7 roots, which traverse the upper and middle trunk and enter the posterior cord.

Nerves of the Shoulder Girdle Phrenic Nerve Optimal placement of the needle along the posterior edge of the sternocleidomastoid muscle reduces the shock intensity necessary for maximal stimulation of the phrenic nerve. The contraction of the diaphragm, inducing hiccup

or interrupting voluntarily sustained vocalization, serves as a clinical sign of effective stimulation. Excessive stimulation voids selective recording by coactivating the brachial plexus, located posteriorly behind the anterior scalene muscle. The diaphragmatic action potential results in the seventh or eighth intercostal space near the costochondral junction being strongly positive and the xiphoid process mildly negative.245 The largest amplitude is recorded between the electrodes placed at these sites as the difference of the two potentials of opposite polarity. Brachial Plexus The brachial plexus comprises the anterior rami of the spinal nerves derived from the C5 through C8 and T1 roots. Thus stimulation at Erb’s point activates the proximal muscles of the shoulder girdle as well as the distal muscles, such as those of the thenar and hypothenar eminences. Table 35–8

Table 35–7. Latency Comparison between Median and Ulnar Nerves in the Same Limb* Site of Stimulation Motor Fibers Wrist Elbow Sensory Fibers Palm Wrist

Median Nerve (ms)

Ulnar Nerve (ms)

Difference (ms)

3.34  0.32 (4.0)† 7.39  0.72 (8.8)

2.56  0.37 (3.3)† 7.06  0.79 (8.6)

0.79  0.31 (1.4)† 0.59  0.60 (1.8)

1.33  0.21 (1.8) 2.80  0.32 (3.4)

1.19  0.22 (1.6) 2.55  0.30 (3.2)

0.22  0.17 (0.6) 0.29  0.21 (0.7)

*Mean  standard deviation (SD) in 35 patients, 14 to 74 years of age (average, 37) with no apparent disease of the peripheral nerve. † Upper limits of normal, calculated as mean  2 SD. From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.

915

Nerve Conduction and Needle Electromyography

FIGURE 35–12 A, Motor nerve conduction study of the radial nerve. The sites of stimulation include Erb’s point (A), axilla (B), above the elbow (C), and mid-forearm (D). Compound muscle action potentials are recorded from the extensor indicis with a pair of surface electrodes. B, Sensory nerve conduction study of the radial nerve. The sites of stimulation include elbow (A) and distal forearm (B). Antidromic sensory potentials are recorded using the ring electrodes placed around the first digit. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

summarizes the conduction time across the brachial plexus calculated by subtracting the distal latency of the ulnar nerve.242 For a unilateral lesion, the side-to-side difference, which should not exceed 0.6 ms, serves as a more sensitive index than the absolute latency. The latency criteria, however, rarely provide useful information in the evaluation of axonal degeneration because remaining motor fibers tend to show a relatively normal value.305 In contrast, the amplitude of the recorded response determines the degree of axonal loss. Despite its considerable variability between the two sides in the same individual, amplitude preservation in the

injured side above one half of the normal side suggests limited distal degeneration and a good prognosis. Musculocutaneous Nerves Optimal sites of stimulation for motor conduction studies361 include the posterior cervical triangle 3 to 6 cm above the clavicle just behind the sternocleidomastoid muscle and the axilla between the axillary artery medially and the coracobrachialis muscle laterally.305 Either surface or needle electrodes may be used to stimulate the nerve and to record the muscle action potentials from the biceps brachii.

Table 35–8. Brachial Plexus Latency with Nerve Root Stimulation Latency across Plexus (ms)

Plexus

Site of Stimulation

Recording Site

Range

Mean

SD

Brachial (upper trunk and lateral cord) Brachial (posterior cord) Brachial (lower trunk and medial cord)

C5 and C6 C6, C7, and C8 C8 and T1, ulnar nerve

Biceps brachii Triceps brachii Abductor digiti quinti

4.8–6.2 4.4–6.1 3.7–5.5

5.3 5.4 4.7

0.4 0.4 0.5

From MacLean, I. C.: Nerve root stimulation to evaluate conduction across the brachial and lumbosacral plexuses. In Third Annual Continuing Education Course, American Association of Electromyography and Electrodiagnosis, Philadelphia, September 25, 1980, with permission.

916

Nerve Conduction and Electromyography

Cutaneous Nerves in the Forearm Lateral Antebrachial Cutaneous Nerves Stimulation applied between the tendon of the biceps medially and the brachioradialis laterally activates the sensory branch of the musculocutaneous nerve that runs superficially at the level of the elbow. Studies of this sensory potential detect abnormalities of the C6 root, upper trunk, or lateral cord better than the median sensory potentials recorded from the second digit, which, more often than not, receives sensory fibers from the C7 root and middle trunk.105 Medial Antebrachial Cutaneous Nerve The medial antebrachial cutaneous nerve subserves the sensation over the medial aspect of the forearm, the area not affected by lesions of the ulnar nerve. Like the ulnar nerve, it originates from the C8 and T1 roots, traverses the lower trunk and medial cord, and pierces the deep fascia in the lower end of the arm, thus bypassing the ulnar glove and cubital tunnel.211

Major Nerves in the Lower Limb Tibial Nerve Motor conduction studies consist of stimulating the tibial nerve at the popliteal fossa and at the ankle lateral to the medial malleolus and recording the muscle response from the abductor hallucis medially or abductor digiti quinti laterally (Fig. 35–13). Reported normal values (mean  SD) include distal latencies of 4.9  0.6 ms for the medial and 6.0  0.7 ms for the lateral plantar nerves over a 12-cm segment,165 and nerve conduction times, calculated as the latency difference between proximal and distal stimulation, of 3.8  0.5 ms for the medial and 3.9  0.5 ms for the lateral plantar nerves over a 10-cm segment across the medial malleolus.119 Tables 35–9 and 35–10 summarize the normal values in our laboratory. Common and Deep Peroneal Nerve Motor conduction studies consist of stimulation of the common peroneal nerve above and below the head of the fibula, and just above the ankle, and recording of muscle action

potentials from the extensor digitorum brevis (Fig. 35–14). Increasing the distance flanked by two sites of stimulation across the knee to 10 cm or longer improves the accuracy in the determination of conduction velocity. A series of shocks applied in short increments, however, delineates a focal conduction abnormality better.174,188 The extensor digitorum longus59 or tibialis anterior77 can usefully substitute for the atrophic extensor digitorum brevis, which may give rise to a small or no response, especially when assessing an advanced neuropathy. The use of needle electrodes and an averaging technique improves resolution in recording small sensory potentials of the deep peroneal nerve over the web between the first and second toes228 or mixed nerve potentials at the fibular head133 after stimulation of the peroneal nerve at the ankle. Tables 35–10 and 35–11 summarize the normal values in our laboratory. Superficial Peroneal Nerve This mixed nerve, originating from the L5 root, begins at the common peroneal bifurcation at the fibular head and gives rise to two sensory nerves in the lower third of the leg. The medial and intermediate dorsal cutaneous nerves subserve the skin of the dorsum of the foot, and the anterior and lateral aspects of the leg. Stimulation of the intermediate dorsal cutaneous branch with the cathode placed against the anterior edge of the fibula elicits the antidromic sensory potential at the ankles just medial to the lateral malleolus. The nerve may be stimulated at two points, 12 to 14 cm from the recording electrode and 8 to 9 cm further proximally, to differentiate the distal and proximal segments. The study helps distinguish a distal lesion from an L5 radiculopathy, which usually spares sensory potentials.169 Sural Nerve The sural nerve, primarily derived from the S1 root, originates from the tibial nerve in the popliteal fossa. It receives a communicating branch from the common peroneal nerve at the junction of the middle and lower third of the leg, where it becomes superficial. Antidromic sensory studies

FIGURE 35–13 Motor nerve conduction study of the tibial nerve. The sites of stimulation include knee (A), above the medial malleolus (B), and below the medial malleolus (C). Compound muscle action potentials are recorded with surface electrodes placed over the abductor hallucis. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

917

Nerve Conduction and Needle Electromyography

Table 35–9. Normal Values for Tibial Nerve*

Site of Stimulation Ankle Knee

Amplitude† (mV)

Latency‡ to Recording Site (ms)

Difference between Two Sides (ms)

Conduction Time between Two Points (ms)

Conduction Velocity (m/s)

5.8  1.9 (2.9)§ 5.1  2.2 (2.5)

3.96  1.00 (6.0)储 12.05  1.53 (15.1)

0.66  0.57 (1.8)储 0.79  0.61 (2.0)

8.09  1.09 (10.3)储

48.5  3.6 (41)¶

*Mean  standard deviation (SD) in 59 patients, 11 to 78 years of age (average, 39), with no apparent disease of the peripheral nerve. † Amplitude measured from the baseline to the negative peak. ‡ Latency measured to the onset with a standard distance of 10 cm between the cathode and the recording electrode. § Lower limits of normal, based on the distribution of the normative data. 储 Upper limits of normal, calculated as the mean  2 SD. ¶ Lower limits of normal, calculated as the mean  2 SD. From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.

Table 35–10. Latency Comparison between Peroneal and Tibial Nerves in the Same Limb* Site of Stimulation Ankle Knee

Peroneal Nerve

Tibial Nerve

3.89  0.87 (5.6) 12.46  1.38 (15.2)

4.12  1.06 (6.2) 12.13  1.48 (15.1)

†

Difference †

0.77  0.65 (2.1)† 0.88  0.71 (2.3)

*Mean  standard deviation (SD) in 52 patients, 17 to 86 years of age (average, 41), with no apparent disease of the peripheral nerve. † Upper limits of normal, calculated as the mean  2 SD. From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.

consist of stimulation of the nerve in the lower third of the leg over the posterior aspect slightly lateral to the midline and recording sensory potentials either around the lateral malleolus (Fig. 35–15) or more distally from the lateral dorsal cutaneous branch.228 In one study comparing three contiguous portions of 7 cm each, the most distal segment revealed a smaller mean velocity than the middle or proximal segment.367 Sural nerve study serves as one of the most sensitive measures for detecting electrophysiologic abnormalities in various types of neuropathies.366 In equivocal cases, the sural-to-radial amplitude ratio may help document abnormalities not apparent based on the absolute values. Preganglionic lesions such as S1 or S2 radiculopathy or a cauda equina involvement spare the sensory action potentials despite clinical sensory loss. This stands in contrast to postganglionic lesions, which consistently cause reduction in amplitude of sensory action potentials. Physiologic studies of the sural nerve provide a unique opportunity for direct histologic confirmation of pathology found in biopsied specimens.88

Nerves of the Pelvic Girdle Lumbosacral Plexus The lumbosacral plexus comprises the lumbar plexus, derived from the L2, L3, and L4 roots, and the sacral plexus, arising from the L5, S1, and S2 roots jointly. Their inaccessibility

to ordinary surface electrical stimulation prevents the use of conventional conduction studies to evaluate their integrity. The recording of the F wave and H reflex serves as an indirect measure of nerve conduction across this region. An alternative method involves needle stimulation99,100,242,254 or high-voltage surface stimulation162 of the L4, L5, or S1 spinal nerve just proximal to the plexus combined with distal stimulation of the plexus. The latency difference between the two stimulus sites equals the conduction time through the plexus. Femoral Nerve Shocks delivered to the femoral nerve above or below the inguinal ligament elicit a muscle potential of the rectus femoris at various distances from the point of stimulation. The latency of the recorded response increases progressively with the distance, reflecting the vertically oriented end-plate region in this muscle.127 The femoral nerve conduction averages 70 m/s, based on the latency difference between the two responses recorded at 14 and 30 cm from the point of stimulation (Table 35–12). This calculation holds only if the nerve branches supplying proximal and distal parts of the muscle have similar and directly comparable electrophysiologic characteristics. Saphenous Nerve The saphenous nerve, derived from the L3 and L4 roots, originates from the femoral nerve as its largest and longest

918

Nerve Conduction and Electromyography

FIGURE 35–14 Motor nerve conduction study of the common peroneal nerve. The sites of stimulation include above the knee (A), below the knee (B), and ankle (C). Compound muscle action potentials are recorded with surface electrodes over the extensor digitorum brevis. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

sensory branch, lying deep along the medial border of the tibialis anterior tendon. The nerve is stimulated with the surface electrodes pressed firmly between the tibia and medial gastrocnemius muscle, usually 12 to 14 cm above the ankle. Signal averaging facilitates the recording of small antidromic sensory potentials with a pair of electrodes placed just anterior to the highest prominence of the medial malleolus. Lateral Femoral Cutaneous Nerve The lateral femoral cutaneous nerve divides into large anterior and small lateral branches about 10 to 12 cm below the anterior superior iliac spine. Studies consist of surface stimulation of the nerve at this point to record orthodromic sensory potentials with a needle electrode inserted 1 cm medial to the lateral end of the inguinal ligament.240 An alternative technique depends on stimulation with a needle or surface electrode placed at the inguinal ligament, and recording of antidromic sensory potentials from the thigh.

HUMAN REFLEXES AND LATE RESPONSES Blink Reflex The blink reflex, elicited by stimulation of the trigeminal nerve, tests the integrity of both the trigeminal and facial nerves, including the proximal segment. A single shock to the supraorbital nerve evokes two separate contractile responses of the orbicularis oculi, R1 and R2. The latency of R1 represents the conduction time along the reflex pathway, including pontine relay, whereas the latency of R2 primarily reflects the excitability of interneurons and a delay of multiple synaptic transmissions. The upper limits of normal (mean  3 SD) include 13.0 ms for electrically elicited R1 and 16.7 ms for mechanically evoked R1. These compare to the distal latency of 4.1 ms after direct stimulation of the facial nerve. The latency difference between the two sides serves as the most sensitive criterion for unilateral lesions, with upper limits of 0.6 ms for distal latency and 1.2 ms and 1.6 ms for the latencies of electrically and mechanically evoked R1, respectively.

Table 35–11. Normal Values for Common and Deep Peroneal Nerves*

Site of Stimulation

Amplitude† (mV)

Latency‡ to Recording Site (ms)

Ankle Below knee Above knee

5.1  2.3 (2.5)§ 5.1  2.0 (2.5) 5.1  1.9 (2.5)

3.77  0.86 (5.5)储 10.79  1.06 (12.9) 12.51  1.17 (14.9)

Difference between Right and Left (ms)

Conduction Time between Two Points (ms)

Velocity (m/s)

0.62  0.61 (1.8)储 0.65  0.65 (2.0) 0.65  0.60 (1.9)

7.01  0.89 (8.8)储 1.72  0.40 (2.5)

48.3  3.9 (40)¶ 52.0  6.2 (40)

*Mean  standard deviation (SD) in 60 patients, 16 to 86 years of age (average, 41), with no apparent disease of the peripheral nerves. † Amplitude measured from the baseline to the negative peak. ‡ Latency measured to the onset with a standard distance of 7 cm between the cathode and the recording electrode. § Lower limits of normal, based on the distribution of the normative data. 储 Upper limits of normal, calculated as the mean  2 SD. ¶ Lower limits of normal, calculated as the mean  2 SD. From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.

919

Nerve Conduction and Needle Electromyography

FIGURE 35–15 Antidromic sensory nerve conduction study of the sural nerve. The diagram shows stimulation on the calf slightly lateral to the midline in the lower third of the leg, and recording with surface electrodes placed behind the lateral malleolus. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

Tables 35–13 and 35–14 summarize a 10-year experience with the blink reflex in our laboratory. A substantial increase in latency of R1 implies demyelination of the central reflex pathway either in the pons,155,177,180,182 the trigeminal151,203,282 or facial nerve,198,202,292 or both.90,179,199,239 Posterior fossa tumors may increase the latency of R1 by compressing the trigeminal or facial nerve either extra-axially or intra-axially.58,178,199

H and T Reflexes Compared to clinical examination of the muscle stretch reflex, electrophysiologic recordings have the advantage of quantitating the response by objectively evaluating its briskness, velocity, or symmetry. The commonly used techniques to measure motor neuron excitability in spasticity and other related conditions include a mechanical tap to the Achilles tendon and electrical stimulation of the tibial nerve. In healthy adults, H reflexes are elicited only in the soleus muscle after stimulation of the tibial nerve and, less consistently, the flexor carpi radialis after stimulation of the median nerve, whereas mechanical stretch of any muscle evokes T reflexes. For example, stretch reflexes elicited by tapping the voluntarily contracted erector spinae comprise two components: the short latency R1, considered segmental in origin, and the long latency R2, induced by a suprasegmental pathway.351 In one study,218 mechanical

stimuli to the ankle and patella elicited T reflexes consistently in healthy subjects, suggesting their clinical value. For evaluation of localized lesions such as radiculopathy, studies of a longer segment including the normal portions of the reflex pathway tend to dilute the effect of focal slowing. Magnetic or electrical stimulation of the S1 nerve root evokes an M response via direct activation of the motor axons and, a few milliseconds later, an H reflex by reflexive excitation of the motor neurons through the group Ia afferent fibers. The peak latency difference between the simultaneously recorded M response and H reflex, or H-M interval, provides a reliable measure of conduction across a short segment within the spinal canal.98,241,289,390 In one study of 100 healthy subjects (Fig. 35–16),390 the H-M interval ranged from 2.6  0.7 ms (mean  SD) with stimulation at the T12 to L1, to 4.2  0.6 ms at the L2 to L3, 5.5  0.3 ms at the L4 to L5, and 6.8  0.5 ms at the S1 spinal process. Table 35–15 summarizes the normal values for the H reflex in healthy adults in our laboratory. In evaluating a unilateral lesion, the latency difference between the two sides provides the most sensitive measure of the T or H reflex.25 Unilateral absence or a right-left latency difference greater than 2.0 ms supports the diagnosis of S1 radiculopathy in the proper clinical context but does not by itself constitute sufficient evidence of a herniated disc. Preterm neonates have slower H reflex conduction velocity

Table 35–12. Normal Values for Femoral Nerve

No.

Age

Onset Latency (ms)

14 cm from stimulus point

42

8–79

3.7  0.45

30 cm from stimulus point

42

8–79

6.0  0.60

Stimulation Point

Recording Site

Just below inguinal ligament

Modified from Gassel, M. M.: A study of femoral nerve conduction time. Arch. Neurol. 9:57, 1963.

Conduction Velocity (m/s) 70  5.5 between the two recording sites

12 4

0

9 0 2 0

14

4

62 17 86

62

0

83 (glabellar tap, 21)* 90

0

88 0 20

0

13

63

0

Abs Delay

124

27 34 150

8

11

105

166

Nl

1

0 1 1

0

7

20

0

44

105 0 17

1

13

78

0

Abs Delay

79

19 33 154

7

8

82

166

Nl

R1 Right and Left Combined

10.5  0.8 (12.5  1.4)* 15.1  5.9 16.4  6.4 10.7  0.8

17.0  3.7 10.1  0.6 11.4  1.2 12.3  2.7

4.2  2.1 5.8  2.6 2.7  0.2

6.7  2.7 2.9  0.4 3.4  0.6 2.9  0.5

R1 (ms)

2.9  0.4

Direct Response (ms)

4.3  0.9

2.8  0.9 3.6  0.6 3.4  0.5

3.9  0.4

3.1  0.5

3.9  1.3

3.6  0.5

R/D Ratio

Abs  absent response; Nl  normal. *R1 elicited bilaterally by a midline glabellar tap in another group of 21 healthy subjects. From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.

Guillain-Barré syndrome Chronic inflammatory polyneuropathy Fisher syndrome Hereditary sensorimotor neuropathy Type I Type II Diabetic polyneuropathy Multiple sclerosis

Normal

Category

Number of Patients

Direct Response Right and Left Combined

35.8  8.4

39.5  5.7 30.1  3.8 33.7  4.6

31.8  1.3

39.5  9.4

37.4  8.9

30.5  3.4

Ipsilateral R2 (ms)

37.7  8.0

39.3  6.4 30.1  3.7 34.8  5.3

31.4  1.9

42.0  10.3

37.7  8.4

30.5  4.4

Contralateral R2 (ms)

Table 35–13. Blink Reflex Elicited by Electrical Stimulation of Supraorbital Nerve in Normal Subjects and Patients with Bilateral Neurologic Diseases (Mean  SD)

921

Nerve Conduction and Needle Electromyography

Table 35–14. Blink Reflex Elicited by Electrical Stimulation of Supraorbital Nerve on the Affected and Normal Sides in Patients with Unilateral Neurologic Diseases (Mean  SD) Category and Side of Stimulation Trigeminal neuralgia Affected side Normal side Compressive lesion of the trigeminal nerve Affected side Normal side Bell’s palsy Affected side Normal side Acoustic neuroma Affected side Normal side Wallenberg’s syndrome Affected side Normal side

Number of Patients

Direct Response (ms)

R1 (ms)

R/D Ratio

Ipsilateral R2 (ms)

89 89

2.9  0.4 2.9  0.5

10.6  1.0 10.5  0.9

3.7  0.6 3.7  0.6

30.4  4.4 30.5  4.2

31.6  4.5 31.1  4.7

17 17

3.1  0.5 3.2  0.6

11.9  1.8 10.3  1.1

3.9  1.0 3.4  0.6

36.0  5.5 33.7  3.5

37.2  5.7 34.8  4.1

100 100

2.9  0.6 2.8  0.4

12.8  1.6 10.2  1.0

4.4  0.9 3.7  0.6

33.9  4.9 30.5  4.3

30.5  4.9 34.0  5.4

26 26

3.2  0.7 2.9  0.4

14.0  2.7 10.9  0.9

4.6  1.7 3.8  0.5

38.2  8.2 33.1  3.5

36.6  8.2 35.3  4.5

23 23

3.2  0.6 3.2  0.4

10.9  0.7 10.7  0.5

3.6  0.6 3.4  0.4

40.7  4.6 34.0  5.7

38.4  7.1 35.1  5.8

Category R2 (ms)

From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.

compared to full-term babies.259 Normal values (mean  SD) for the soleus H reflex established in 83 preterm and term infants include a latency of 19.2  2.16 ms for conceptional ages of 31 to 34 weeks, 16.7  1.5 ms for 35 to 39 weeks, and 15.9  1.5 ms for 40 to 45 weeks.31

body (intra-axial mesencephalic nucleus and extra-axial craniospinal ganglia). Similarly, ganglionopathy, but not axonal sensory neuropathy, tends to spare the masseter reflex in patients with pure sensory symptoms.12

F Waves Masseter Reflex A sharp tap to the mandible stretches the muscle spindles of the masseter muscle, causing the jaw reflex, or masseteric T reflex.155 The jaw reflex shows physiologic characteristics different from those of the spinal monosynaptic reflex. For example, muscle vibration that inhibits the soleus T and H reflexes potentiates the masseteric T and H reflexes.136,137 The criteria for abnormality established in one study, with the use of a needle recording electrode,283 comprised unilateral absence of the reflex, a difference of more than 0.5 ms in latency between the two sides, or bilateral absence of the reflex up to the age of 70 years. Table 35–16 summarizes normal values in our laboratory.195 Despite technical problems in standardizing the mechanical stimulus and regulating the tonus of the masseter muscle for optimal activation, an unequivocal unilateral delay or absent response suggests a lesion of the trigeminal nerve or the brainstem.203 Patients with Friedreich’s ataxia have hypoactive or no stretch reflexes in the upper and lower limbs, but masseter reflexes are normal or paradoxically hyperactive. This discrepancy may result from a different location of the afferent nerve cell

The F wave, which results from the backfiring of antidromically activated anterior horn cells, serves as a measure of motor conduction along the entire peripheral axon, including the most proximal segment.56,91,107,117,196,197,206,284,285,311,388 Recording Procedures For clinical studies, routine procedures include stimulation of the median and ulnar nerves at the wrist and elbow and of the tibial and peroneal nerves at the ankle and knee. In general, the sum of the F latency and M latency remains the same regardless of stimulus sites. Thus Fd  Md  Fp  Mp, where Fd and Fp are F latencies and Md and Mp are M latencies with distal and proximal stimulation, respectively. When necessary, this equation provides an estimated latency of the F wave from any proximal site by subtracting the proximal M latency from the sum of the distal F and M latencies. Stimulation of the facial nerve also elicits an F wave,324 although, in the short segment under study, superimposition of the M response usually makes its recognition difficult. Furthermore, inadvertent stimulation of neighboring trigeminal afferent fibers may simultaneously activate reflex responses that may mimic the late response.

922

Nerve Conduction and Electromyography

FIGURE 35–16 A, Schematic representation of soleus H reflexes from electrical stimulation of the S1 nerve root at the S1 foramen and of the tibial nerve at the popliteal fossa. B, The response complex of the H reflex and M response in one of the subjects elicited by magnetic (upper traces) and electrical (lower traces) stimulation of the S1 nerve root at the S1 foramen. The intensity of the nerve stimulus is to the right of the traces. (From Zhu, Y., Starr, A., Haldeman, S., et al.: Soleus H-reflex to S1 nerve root stimulation. Electroencephalogr. Clin. Neurophysiol. 109:10, 1998, with permission.)

Central Latency Central latency from the stimulus point to and from the spinal cord is derived as F  M, where F and M are latencies of the F wave and the M response (Fig. 35–17). Subtracting an estimated delay of 1.0 ms for the turnaround time at the cell and dividing by 2 ([F  M  1]/2) provides the conduction time along the proximal segment

from the stimulus site to the spinal cord. Although no human studies measured the central activation time at the anterior horn cells, animal data indicate a delay of nearly 1.0 ms.306 On one hand, the absolute refractory period of the fastest human motor fibers lasts about 1.0 ms or slightly less.184 Thus the recurrent discharge cannot propagate distally across the refractory period of

Table 35–15. Normal Values for H Reflex*

Amplitude† (mV) 2.4  1.4

Difference between Right and Left (mV)

Latency‡ to Recording Site (ms)

Difference between Right and Left (ms)

1.2  1.2

29.5  2.4 (35)§

0.6  0.4 (1.4)§

*Mean  standard deviation (SD) in the same 59 patients shown in Table 35–9. † Amplitude measured from the baseline to the negative peak. ‡ Latency measured to the onset. § Upper limits of normal calculated as mean  2 SD. From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.

923

Nerve Conduction and Needle Electromyography

Table 35–16. Latency and Amplitude of Masseter Reflex in 20 Normal Subjects Latency (ms) Mean right Mean left Total SD Mean  3 SD

Latency Difference (Large Value Minus Small Value)

Amplitude (mV)

0.27 0.15 0.8

0.23 0.21 0.22 0.24 Variable

7.10 7.06 7.08 0.62 9.0

Amplitude Ratio (Large Value over Small Value)

1.44 0.42 2.7

From Kimura, J.: Needle electromyography. In Bertorini, R. (ed.): Evaluation and Diagnostic Tests in Neuromuscular Disorders. Woburn, MA, Butterworth-Heinemann, p. 331, 2002, with permission.

the axon hillock for 1.0 ms after the passage of an antidromic impulse. On the other hand, the somatodendrite spike would abate after Renshaw inhibition activated by an antidromic impulse with a synaptic delay of 1.0 ms. A narrow window based on these limiting factors allows backfiring to materialize in only 1% to 5% of the motor neuron population invaded by antidromic impulses.

FIGURE 35–17 The latency difference between the F wave and the M response represents the passage of a motor impulse to and from the cord through the proximal segment. Considering an estimated minimal delay of 1.0 ms at the motor neuron pool, the proximal latency from the stimulus site to the cord equals (F  M  1)/2, where F and M are latencies of the F wave and M response, respectively. In the segment to and from the spinal cord, FWCV  (D 2)/(F  M  1), where D is the distance from the stimulus site to the cord, (F  M  1)/2 is the time required to cover the length D, and FWCV is F-wave conduction velocity. Dividing the conduction time in the proximal segment to the cord by that of the remaining distal segment to the muscle, the F ratio  (F  M  1)/2M, where (F  M  1)/2 and M are proximal and distal latencies, respectively. (From Kimura, J.: Proximal versus distal slowing of motor nerve conduction velocity in the Guillain-Barré syndrome. Ann. Neurol. 3:344, 1978, with permission.)

F-Wave Conduction Velocity F-wave latencies suffice in studying a limb of average length or in documenting sequential changes in the same subject. For unilateral lesions affecting one nerve, latency comparison between the right and left sides in the same subject or of one nerve with another in the same limb serves as a sensitive measure of abnormality (Tables 35–17 and 35–18).185 Otherwise, clinical assessment of F-wave latency calls for a surface determination of the limb length or the patient’s height with the use of a nomogram359 to adjust for individual differences.350 In the upper limbs, the surface measurement should follow from the stimulus point to the C7 spinous process via the axilla and midclavicular point to closely estimate the nerve length corresponding to the central latency.181 In the lower limb, surface measurement approximates the nerve course best if done from the stimulus site to the T12 spinous process by way of the knee and greater trochanter of the femur.196 The estimated length of a nerve segment by surface measurement correlates well with its F-wave latency. Observations in cadavers showed a good agreement between surface determinations and actual lengths of the nerves in the upper238 as well as lower limbs.196 F-wave conduction velocity (FWCV) in the proximal segment equals the estimated nerve length divided by the conduction from the stimulus point to the spinal cord. Thus FWCV  (2D)/(F  M  1) where D represents the distance from the stimulus site to the cord and (F  M  1)/2 is the central latency, or the time required to cover the length. Clinical Value and Limitations Neurologic disorders showing consistent F-wave abnormalities include hereditary motor sensory neuropathy,181,284 acute or chronic demyelinating neuropathy,197,206 diabetic neuropathy,60,205 uremic neuropathy,1,286 alcoholic neuropathy,231 and a variety of other neuropathies. F waves show less consistent changes in such diverse categories of disorders as entrapment neuropathies,91,386 amyotrophic lateral sclerosis,10 radiculopathies,110 and cervical syringomyelia.291

Wrist Elbow Axilla储 Wrist Above elbow Axilla储 Ankle Above knee Ankle Knee

Site of Stimulation

0.73  0.54 (1.8) 1.42  1.03 (3.5) 1.28  0.91 (3.1) 1.40  1.04 (3.5) 1.25  0.92 (3.1)

20.3  1.6 (24) 48.4  4.0 (56) 39.9  3.2 (46) 47.7  5.0 (58) 39.6  4.4 (48)

43.8  4.5 (53) 27.6  3.2 (34)

10.4  1.1 (13) 44.7  3.8 (52) 27.3  2.4 (32)

23.0  2.1 (27)¶ 15.4  1.4 (18) 10.6  1.5 (14) 25.0  2.1 (29) 16.0  1.2 (18)

0.95  0.67 (2.3)¶ 0.76  0.56 (1.9) 0.85  0.61 (2.1) 1.0  0.83 (2.7) 0.68  0.48 (1.6)

26.6  2.2 (31)¶ 22.8  1.9 (27) 20.4  1.9 (24) 27.6  2.2 (32) 23.1  1.7 (27)

1.52  1.02 (3.6) 1.23  0.88 (3.0)

0.76  0.52 (1.8) 1.28  0.90 (3.1) 1.18  0.89 (3.0)

0.93  0.62 (2.2)¶ 0.71  0.52 (1.8) 0.85  0.58 (2.0) 0.84  0.59 (2.0) 0.73  0.52 (1.8)

Difference between Right and Left (ms)

52.6  4.3 (44) 53.7  4.8 (44)

1.11  0.11 (0.89–1.33)

1.05  0.09 (0.87–1.23)

1.05  0.09 (0.87–1.23)

65.3  4.8 (55) 65.7  5.3 (55) 49.8  3.6 (43) 55.1  4.6 (46)

0.98  0.08 (0.82–1.14)¶**

F Ratio§ between Proximal and Distal Segments

65.3  4.7 (56)** 67.8  5.8 (56)

Conduction Velocity‡ to and from the Spinal Cord (m/s)

*Mean  standard deviation (SD) in the same patients shown in Tables 35–4, 35–6, 35–9, and 35–11. † Central latency  F  M, where F and M are latencies of the F wave and M response, respectively. ‡ Conduction velocity  2D/(F  M  1), where D is the distance from the stimulus point to C7 or T12 spinous process. § F ratio  (F  M  1)/2M with stimulation with the cathode on the volar crease at the elbow (median), 3 cm above the medial epicondyle (ulnar), just above the head of fibula (peroneal), and in the popliteal fossa (tibial). 储 F(A)  F(E)  M(E)  M(A), where F(A) and F(E) are latencies of the F wave with stimulation at the axilla and elbow, respectively, and M(A) and M(E) are latencies of the corresponding M response. ¶ Upper limits of normal calculated as mean  2 SD. **Lower limits of normal calculated as mean  2 SD. From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.

120 peroneal nerves from 60 subjects 118 tibial nerves from 59 subjects

122 median nerves from 61 subjects 130 ulnar nerves from 65 subjects

Number of Nerves Tested

Central Latency† to and from the Spinal Cord (ms)

Difference between Right and Left (ms)

F-Wave Latency to Recording Site (ms)

Table 35–17. F Waves in Normal Subjects*

Wrist Elbow

Ankle Knee

70 nerves from 35 patients

104 nerves from 52 patients

Tibial Nerve 48.1  4.2 (57) 40.1  3.7 (48)

Peroneal Nerve 47.7  4.0 (55) 39.6  3.7 (47)

‡

1.68  1.21 (4.1) 1.71  1.19 (4.1)

Difference

1.00  0.68 (2.4) 0.84  0.55 (1.9)

27.2  2.5 (32) 23.0  1.7 (26)

26.6  2.3 (31) 22.9  1.8 (26)

Difference

‡

‡

Ulnar Nerve

Median Nerve

43.6  4.0 (52) 27.1  2.9 (33)

Peroneal Nerve

23.3  2.2 (28) 15.5  1.4 (18)

‡

Median Nerve

‡

44.1  3.9 (52) 28.0  2.7 (33)

Tibial Nerve

24.5  2.4 (29) 16.0  1.2 (18)

Ulnar Nerve

Difference

1.79  1.20 (4.2) 1.75  1.07 (3.9)

Difference

1.24  0.75 (2.7)‡ 0.79  0.65 (2.1)

Central Latency† to and from the Spinal Cord

*Mean  standard deviation (SD) in the same patients shown in Tables 35–7 and 35–10. † Central latency  F  M, where F and M are latencies of the F wave and M response, respectively. ‡ Upper limits of normal calculated as mean  2 SD. From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.

Site of Stimulation

Number of Nerves Tested

F-Wave Latency to Recording Site

Table 35–18. Comparison between Two Nerves in the Same Limb*

926

Nerve Conduction and Electromyography

Normal Values Tables 35–17 and 35–18 summarize the normal ranges and the upper and lower limits of F-wave latency and other aspects of the F-wave values, defined as 2 SD around the mean, in the same control subjects whose data were presented in the preceding section on nerve conduction studies. Neonates and infants tend to have large F waves probably because the immaturity of physiologic inhibition renders the anterior horn cells more excitable than in adults.259 The F-wave latency remains relatively constant during childhood because a rapid change in conduction velocity compensates for the increase in arm length. In one study,257 the minimal latency of F waves elicited with stimulation of the median nerve at the wrist averaged 17 ms in neonates (1 to 28 days), 15 ms in infants (1 month to 1 year), and 16 ms in children (2 to 12 years). The F-wave latency then increases, reaching its maximal value seen at 20 years of age.219 An older group of subjects has longer F-wave latencies compared to young healthy subjects.268

A Wave With a submaximal stimulation, activating one branch of an axon but not the other, the antidromic impulse propagated up to the point of division may turn around to proceed distally along the second fiber not excited by the stimulus. This orthodromic impulse induces a constant late response, called an A wave, a term preferred over the traditional designation, “axon reflex,” which erroneously implies a reflexive origin. As suggested by its alternative description, “intermediate latency response,” the A wave usually, though not always, appears between the M response and F wave.122 A decrease in A-wave latencies with more proximal stimulation indicates that the response results from an initially antidromic passage of impulse analogous to the F wave (Fig. 35–18). Pathophysiologic mechanisms possibly involving a majority of A waves include, in addition to the less common collateral sprouting described above, ephaptic and ectopic discharges generated at a hyperexcitable site along the proximal portion of the nerve.19,243 Shocks of a higher intensity, activating both branches distally, can eliminate the A waves generated by collateral sprouting because two antidromic impulses now collide at the turning point. Thus supramaximal stimuli normally abolish the collateral A wave altogether, unless the structural changes of the nerve or surrounding tissues prevent the current from activating one of the branches. In contrast, an ephaptic A wave may persist even with the use of very highintensity stimuli because the antidromic impulse of the fastconducting axon may have already passed the site of ephaptic transmission induced by a neighboring slow-conducting axon, thus precluding the collision. An increase in shock intensity also fails to inhibit an ectopic A wave generated by antidromic passage of an impulse across a hyperexcitable segment of a nerve branch. In this case, paired shocks abolish the A wave because of the collision between the ectopically

generated orthodromic impulse and the second antidromic impulse. With repetitive stimuli, even numbered shocks cause collision for the same reason, producing ectopic A waves only in response to odd-numbered stimuli. The presence of A waves signals a heterogeneous group of neurogenic abnormalities encompassing acute and chronic neuropathies widely varying in pathophysiology, from nerve regeneration to demyelination. The disease entities commonly associated with the A wave include entrapment syndromes, brachial plexus lesions, diabetic neuropathy, hereditary motor sensory neuropathy, facial neuropathy, amyotrophic lateral sclerosis, Guillain-Barré syndrome, and cervical root lesions.19,122,210,315

CLINICAL APPLICATIONS Common Sources of Error Although simple principles dictate the nerve conduction studies, pitfalls abound in practice mostly because of technical problems that usually account for unexpected findings. Often overlooked sources of error include intermittent power failure in the stimulating or recording system, excessive spread of stimulation current, anomalous innervation, temporal dispersion, and inaccuracy of surface measurement. Stationary far-field peaks may result from a moving source not only in a referential but also in a bipolar derivation, usually used for recording near-field potentials. Lack of awareness of these possibilities can cause confusion in the interpretation of the data, suggesting an incorrect diagnosis, especially if the findings mimic common manifestations of a disease. Special care in regard to technical details leads to good quality control, which in turn improves the reproducibility of the results.

Collision Technique An appropriately placed cathode can selectively activate the median or ulnar nerve at the wrist or elbow, but usually not at the axilla, where the two nerves lie in close proximity. If the current intended for the median nerve spreads to the ulnar nerve when studying carpal tunnel syndrome, the thenar eminence potentials comprise not only a slowed median but also a normal ulnar component. The measured onset latency will then indicate the normal ulnar response, which precedes the median response (Fig. 35–19). In the same case, selective stimulation of the median nerve at the elbow reveals a prolonged latency. The calculated conduction time between the ulnar response from the axilla and the median response from the elbow would suggest an erroneously fast conduction velocity. In extreme cases, the median response from the elbow has a greater latency than the ulnar component from the axilla. The reverse discrepancy can occur in a study of tardy ulnar palsy, with spread of axillary stimulation to the median nerve.

FIGURE 35–18 A, A 51-year-old man with low back pain. Stimulation of the right tibial nerve at the ankle elicited a number of A waves. A series of eight tracings displayed with stepwise vertical shift of the baseline confirm the consistency. This type of display not only facilitates the selection of the F wave with minimal latency but also allows individual assessments of all the late responses. Of the three A waves (small arrows, 1, 2, and 3) elicited by weak shocks (S(1)), stronger shocks (S(2)) eliminated only the earliest response. B, Collateral sprouting in the proximal part of the nerve. A strong shock, activating both branches, can eliminate the A wave generated by weak stimulation by collision. (From Fullerton P. M., and Gilliatt, R. W.: Axon reflexes in human motor nerve fibres. J. Neurol. Neurosurg. Psychiatry 28:1, 1965, with permission.) C, A waves after stimulation of the left tibial nerve at the ankle or knee. Proximal stimulation eliminated the A wave (arrow) that followed the F wave with distal stimulation. D, A 50-year-old man with recurrent backaches following laminectomy. Stimulation of the tibial nerve at the ankle or knee elicited the A wave (arrow). Like the F wave, the latency of the A wave decreased with proximal site of stimulation. This indicates that the impulse first propagates in the centripetal direction. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

928

Nerve Conduction and Electromyography

FIGURE 35–19 A 39-year-old man with carpal tunnel syndrome. The stimulation of the median nerve at the wrist (S1) or elbow (S2) elicited a muscle action potential with increased latency in the thenar eminence. Spread of axillary stimulation (S3) to the ulnar nerve (third tracing from top) activated ulnar-innervated thenar muscles with shorter latency. Another stimulus (S4) applied to the ulnar nerve at the wrist (bottom tracing) blocked the proximal impulses by collision. The muscle action potential elicited by S4 occurred much earlier. The diagram on the left shows collision between the orthodromic (solid arrows) and antidromic (dotted arrows) impulses. (From Kimura, J.: Collision technique— physiological block of nerve impulses in studies of motor nerve conduction velocity. Neurology [Minneap.] 26:680, 1976, with permission.)

The addition of distal stimulation achieves a physiologic nerve block through collision, allowing selective recording of the median or ulnar component despite coactivation of both nerves proximally.183 In studying the median nerve, for example, a distal stimulus is delivered to the ulnar nerve at the wrist. The antidromic impulse generated here collides with the orthodromic ulnar impulse from the axilla, leaving only the median impulse to reach the muscle. Delaying the proximal stimulus a few milliseconds accomplishes an optimal separation between the ulnar response induced by the distal stimulus and the median response under study. If this delay exceeds the conduction time between the distal and proximal points of stimulation, the antidromic impulse from the wrist passes the stimulus site at the axilla without collision. By the same principles, the use of a distal stimulus can block the median nerve in selective recording of the ulnar response after coactivation of both nerves at the axilla.

Anomalous Innervations Martin-Gruber Anastomosis Martin (in 1763)247 and Gruber (in 1870)140 demonstrated frequent communication from the median to the ulnar nerve at the level of the forearm. This anastomosis, usually

seen a few centimeters distal to the medial humeral epicondyle,369 often originates from the anterior interosseous nerve and predominantly involves motor axons. Rare sensory contribution, when present, may follow a different distribution.57,372 The number of axons taking the anomalous course varies widely, usually, though not always,232 supplying ordinarily ulnar-innervated intrinsic hand muscles, most notably the first dorsal interosseous, adductor pollicis, and abductor digiti minimi.382 The possibility of selectively stimulating the anomalous fibers at the elbow without coactivating the median nerve proper (Fig. 35–20)191 suggests a grouping of the anastomotic fibers in a separate bundle, rather than intermingling with the median nerve proper. The anomaly is reported in 15% to 32% of an unselected population, often bilaterally.9,201 A higher incidence among congenitally abnormal fetuses in general and those with trisomy 21 in particular favors its phylogenetic origin.340 Other anomalies related to Martin-Gruber anastomosis include rare communication crossing from the ulnar to the median nerve in the forearm,138,278 occasionally involving only the sensory axons,154 and innervation of the ulnar aspect of the dorsum of the hand by the superficial radial sensory nerve.251 Careful waveform analysis of the muscle action potentials can confirm the presence of a Martin-Gruber anastomosis if suspected during nerve conduction studies. Thus

Nerve Conduction and Needle Electromyography

929

FIGURE 35–20 A 46-year-old woman with carpal tunnel syndrome and the Martin-Gruber anastomosis. Stimulation at the elbow (S2) activated not only the median nerve but also communicating fibers, giving rise to a complex compound muscle action potential. With proper adjustment of electrode position and shock intensity, another stimulus at the elbow (S2) excited the median nerve selectively without activating the anastomosis. Another stimulus (S3) applied to the ulnar nerve at the wrist (bottom tracing) achieved the same effect by blocking the unwanted impulse transmitted through the communicating fibers. (From Kimura, J.: Electromyography and Nerve Stimulation Techniques: Clinical Applications [in Japanese]. Tokyo, Igaku-Shoin, 1990, with permission.)

stimulation of the median nerve at the elbow coactivates the communicating ulnar nerve fibers, producing action potentials of not only the median-innervated thenar muscle but also anomalously innervated thenar and hypothenar muscles. In contrast, stimulation of the median nerve at the wrist evokes a smaller response lacking the ulnar component. Studies of the ulnar nerve in the presence of Martin-Gruber anastomosis also show confusing results.208 Stimulation at the elbow, sparing the communicating branch still attached to the median nerve, evokes only a partial response, whereas stimulation at the wrist activates the additional anomalous fibers, giving rise to a full response, mimicking a conduction block of the ulnar nerve at the elbow.246 The collision technique183,323 allows selective blocking of unwanted impulses transmitted via the communicating fibers. If a patient with this anomaly develops carpal tunnel syndrome, stimulation of the median nerve at the elbow reveals two temporally dispersed potentials, a normal ulnar and a delayed median response. The initial ulnar component may give a mistaken notion of normalcy of the median nerve unless the anomaly is detected. In the same patient, stimulation at the wrist only elicits a delayed median component without an ulnar response.183,220 The short latency difference incorrectly calculated between the ulnar and median responses would lead to an unreasonably fast conduction velocity from the elbow to the wrist.41,181,220

The anomalously innervated ulnar muscles located at some distance from the recording site on the thenar eminence often, though not always, display an initial positive deflection.143 In most cases, a collision technique can block impulses in the anomalous fibers without affecting those transmitted along the median nerve proper to clarify the ambiguity. Severance or substantial injury of the ulnar nerve at the elbow ordinarily leads to inexcitability of the distal segment concomitant with evolving wallerian degeneration. In the presence of this anomaly, the communicating fibers bypass the lesion, maintaining a normal supply of anomalously innervated muscle fibers. Stimulation of the ulnar nerve at the wrist, therefore, evokes a small but otherwise normal muscle action potential. As an extreme example, an injury that would usually separate the ulnar nerve completely will not appreciably affect the intrinsic hand muscles if all or nearly all ulnar fibers are attached to the median nerve at the elbow. In this rare condition, called all-median hand, the intrinsic hand muscles usually supplied by the ulnar nerve receive innervation via the communicating fibers.244 Normal motor unit potentials in the ulnar-innervated muscles, despite severe damage to the ulnar nerve at the elbow, may signal the presence of this anomaly. Conversely, studies will reveal spontaneous discharges in the ulnar-innervated intrinsic hand muscles after an injury to the median nerve at the elbow.

930

Nerve Conduction and Electromyography

Accessory Deep Peroneal Nerve The extensor digitorum brevis, the muscle commonly used in peroneal nerve conduction studies, usually derives its supply from the deep peroneal nerve, a major branch of the common peroneal nerve. As the most frequent anomaly of the lower limb, seen in 20% to 28% of unselected subjects, this muscle may receive innervation from the superficial peroneal nerve via a communicating branch, called the accessory deep peroneal nerve (Fig. 35–21). The anomalous fibers descend on the lateral aspect of the leg after arising from the superficial peroneal nerve, then pass behind the lateral malleolus before proceeding anteriorly to innervate the lateral portion of the extensor digitorum brevis.163,221 The anomaly, inherited as dominant trait,63 occasionally supplies the extensor digitorum brevis

exclusively with no contribution from the common peroneal nerve.270 In the presence of this anastomosis, the compound muscle action potential evoked by stimulation of the common peroneal nerve at the knee equals the sum of responses elicited by stimulation of the deep peroneal nerve at the ankle and stimulation of the accessory deep peroneal nerve behind the lateral malleolus. Injury to the deep peroneal nerve causes weakness of the tibialis anterior, extensor digitorum longus, extensor hallucis longus, and medial but not lateral part of the extensor digitorum brevis supplied by the anastomosis.76,142 The collision technique181 may help identify isolated abnormalities and localize a lesion to various branches, for example, the accessory deep peroneal nerve.323

FIGURE 35–21 A, Compound muscle action potentials recorded from surface electrodes over the extensor digitorum brevis after a maximal stimulus to the common peroneal nerve at the knee (A), deep peroneal nerve on the dorsum of the ankle (B), accessory deep peroneal nerve posterior to the lateral malleolus (C and D), and tibial nerve posterior to the medial malleolus (E) at the ankle. Diagram of the foot indicates the site of block (X) of the accessory deep peroneal nerve with 2% lidocaine and the points of stimulation (B, C, and D) and recording (R). B, Course of the accessory deep peroneal nerve and action potentials recorded with coaxial needle electrode (R) in the lateral belly of the extensor digitorum brevis muscle following stimulation of the common peroneal nerve at the knee (A), just below the head of the fibula (B), superficial peroneal nerve (C), accessory deep peroneal nerve posterior to the lateral malleolus (D), and deep peroneal nerve on the dorsum of the ankle (E). The volume-conducted potential from the medial bellies of the extensor digitorum brevis (E) reduces the amplitude of the action potential of the lateral belly with simultaneous stimulation of the common peroneal nerve at A or B. (From Lambert, E. H.: The accessory deep peroneal nerve: a common variation in innervation of extensor digitorum brevis. Neurology [Minneap.] 19:1169, 1969, with permission.)

Nerve Conduction and Needle Electromyography

Temporal Dispersion and Phase Cancellation Physiologic Temporal Dispersion In the study of nerve function, not only calculation of the maximal velocities based on the onset latency, but also waveform analyses of the recorded response play an essential role. The latter aspect of the study provides a measure of greater importance, especially in evaluating peripheral neuropathies with segmental block, in which surviving axons conduct normally. In clinical tests of motor and sensory conduction, the size of the recorded response serves as a measure of the number of excitable nerve axons. Any discrepancy between responses to proximal and distal shocks, however, does not necessarily imply the presence of conduction block or other abnormalities. The impulses of physiologically slow- and fast-conducting nerve fibers become increasingly separated in latency over a long conduction path.43,66,220 Thus the longer the distance between stimulating and pickup electrodes, the greater the temporal dispersion among different nerve fibers. Because of this desynchronization, a response recorded at a point distant from the stimulation site becomes lower in amplitude and longer in duration and, contrary to the common belief, smaller in area under the waveform. This diminution of evoked responses associated with temporal dispersion is caused by phase cancellation, which affects nerve action potentials more than muscle responses, as described below. A slight latency difference could line up the negative and positive peaks of short-duration diphasic sensory spikes of the fast and slow fibers, causing them to cancel each other (Fig. 35–22). In contrast, the same slight shift in latency still superimposes longer duration motor unit potentials nearly in phase rather than out of phase, without much cancellation. Indeed, the same degree of temporal dispersion, if measured in percentage, gives rise to a much greater increase in duration of the sensory potential than that of the muscle response.200 As expected from the principle of duration-dependent phase cancellation, a physiologic temporal dispersion also substantially reduces the size of short-duration muscle action potentials such as those recorded from intrinsic foot muscles. A number of technical factors influence the degree of physiologic phase cancellation. Of particular importance is the interelectrode distance between a pair of recording electrodes that dictates the duration of the recorded response. Pathologic Temporal Dispersion If the latency increases for some but not other axons, as might be expected in demyelinating neuropathy, compound muscle action potentials also diminish dramatically based solely on phase cancellation between normally conducting and pathologically slow fibers. As predicted by our model,204 this type of phase cancellation reduces the amplitude of the muscle response well beyond the usual

931

physiologic limits. This finding could give rise to a false impression of motor conduction block, and probably accounts for an occasionally encountered discrepancy between diminished size of proximally evoked responses and relatively normal recruitment of the motor unit potentials with preserved strength. Model for Phase Cancellation A simple model tests the effects of desynchronized inputs204 by stimulating the median (S1) and ulnar (S2) nerve at varying interstimulus intervals at the wrist, and recording a sensory potential of the fourth digit and a muscle potential over the thenar eminence. As predicted, the sensory potential declines nearly 50%, with a 1.0- to 1.5-ms delay of S2 after S1, whereas the muscle response reaches a minimal size with a latency change of 5 to 6 ms. These findings indicate that a latency change on the order of one half the total duration of unit discharge leads to a maximal phase cancellation and consequently loss of amplitude and area. With further separation, the two units no longer overlap out of phase to cause cancellation, and in fact may paradoxically increase the size of the response because excessive desynchronization may now counter the physiologic phase cancellation. Because variables such as interelectrode distance between E1 and E2 influence the outcome substantially, the commonly held criteria based on percentage reduction are untenable except in entirely standardized studies.224 Thus comparison between distally and proximally elicited responses often fails to differentiate physiologic, as opposed to pathologic, temporal dispersion, not to mention conduction block. An alternative method relies on segmental stimulation at more than two sites to test a linear relationship between the latency and the size of the recorded responses seen in physiologic phase cancellation.193 This approach enjoys the distinct advantage of having a built-in internal control for all recording variables such as interelectrode spacing (Fig. 35–23). A nonlinear change in amplitude or waveform indicates either a conduction block or a pathologic temporal dispersion (Fig. 35–24). The distinction between the two possibilities must, in part, depend on clinical clues and needle electrode study of motor unit recruitment as described below.

Practical Assessment of Conduction Block: Criteria for Conduction Block Motor nerve conduction studies play a key role in the differential diagnosis of clinical weakness. A conduction block usually implies demyelination,102 although other conditions such as ischemia and channelopathies can cause similar reversible changes.132,153 Additional characteristics of demyelination include slowing of nerve conduction, pathologic temporal dispersion, and repetitive discharges. These abnormalities may also alter the waveform of the evoked

932

Nerve Conduction and Electromyography

FIGURE 35–22 A, Sensory action potentials. A model for phase cancellation between fast-conducting (F) and slow-conducting (S) sensory fibers. With distal stimulation, two unit discharges summate in phase to produce a sensory action potential twice as large. With proximal stimulation, a delay of the slow fiber causes phase cancellation between the negative peak of the fast fiber and positive peak of the slow fiber, resulting in a 50% reduction in size of the summated response. B, Compound muscle action potentials. Same arrangements as above to show the relationship between fast-conducting (F) and slow-conducting (S) motor fibers. With distal stimulation, two unit discharges representing motor unit potentials summate to produce a muscle action potential twice as large. With proximal stimulation, long-duration motor unit potentials still superimpose nearly in phase despite the same latency shift of the slow motor fiber as the sensory fiber. Thus a physiologic temporal dispersion alters the size of the muscle action potential only minimally, if at all. (From Kimura, J., Machida, M., Ishida, T., et al.: Relation between size of compound sensory or muscle action potentials and length of nerve segment. Neurology 36:647, 1986, with permission.)

action potential but, unlike conduction block, do not by themselves cause muscle weakness. Electromyography reveals little or no evidence of denervation after selective damage of the myelin sheath unless the patient develops secondary axonal degeneration. In partial conduction blocks, the surviving motor unit potentials, normal in amplitude and waveform, discharge at a high rate to com-

pensate for a decreased recruitment in proportion to the number of blocked nerve fibers. Axonal damage, after the first week of injury, reduces distal and proximal responses equally, maintaining a normal ratio because excitability fails at the neuromuscular junction and along the distal nerve segment. Thus the usual criteria for conduction block depend on amplitudes or area ratio of

Nerve Conduction and Needle Electromyography

933

FIGURE 35–23 Simultaneous recordings of compound muscle action potentials from the thenar eminence and sensory nerve action potentials from index and middle fingers after stimulation of the median nerve at palm, wrist, elbow, and axilla. Progressively more proximal series of stimuli elicited nearly the same muscle response but progressively smaller sensory response from the wrist to the axilla, showing a linear relationship to latency. (From Kimura, J., Machida, M., Ishida, T., et al.: Relation between size of compound sensory or muscle action potentials and length of nerve segment. Neurology 36:647, 1986, with permission.)

compound muscle action potentials elicited by proximal versus distal stimulation.8,279,349,370 For example, motor conduction block may be defined as a reduction in amplitude ratio greater than 20% to 50% with less than 15% increase in duration. These definitions, however, suffer from considerable

uncertainty because similar waveform changes may also result from pathologic temporal dispersion and phase cancellation in the absence of conduction block. The combination of clinical and electrophysiologic findings usually documents a motor conduction block better

FIGURE 35–24 A 34-year-old man with selective weakness of foot dorsiflexors and low back pain radiating to the opposite leg. The nerve conduction studies revealed a major conduction block between the two stimulation sites, b and b, at the knee. The weakness abated promptly when the patient refrained from habitual leg crossing. (From Kimura, J.: Electromyography and Nerve Stimulation Techniques: Clinical Applications [in Japanese]. Tokyo, Igaku-Shoin, 1990, with permission.)

934

Nerve Conduction and Electromyography

than the commonly used criteria based solely on waveform analysis.192 In typical situations, stimulation distal to the lesion elicits a large muscle potential associated with a vigorous twitch,190 despite clinical weakness and a paucity of voluntarily activated motor unit potentials.62 If distal stimulation elicits normal or nearly normal M response but fail to evoke F waves, this too suggests a conduction block in the proximal nerve segment.108,181 A sustained period of immobility, however, may render the motor neurons hypoexcitable, making it difficult to evoke a normal number of F waves. An unexpected absence of F wave, therefore, may be seen in some patients with a hysterical paralysis, even if the integrity of the peripheral nerve is preserved. Careful analysis of the waveform of an evoked potential improves the accuracy of electrophysiologic interpretation by elucidating technical problems that account for a majority of unusual findings. For example, a pattern of waveform changes mimicking a conduction block may result from the use of a submaximal stimulus at one point and a supramaximal stimulus at a second site. Unrecognized contribution from anomalous branches such as the MartinGruber anastomosis also may lead to a puzzling alteration in amplitude, as does unintended current spread to a neighboring nerve.183,184 Any of these circumstances gives rise to a confusing set of electrophysiologic findings, sometimes leading to an erroneous conclusion. In addition, dissimilar responses elicited by distal and proximal stimuli usually imply that the two onset latencies measured relate to fibers of different conduction characteristics. This finding, therefore, invalidates the value of calculation of conduction velocities based on the conventional formula. A proximal stimulation may fail despite the use of ordinarily adequate intensities if structural abnormalities render the nerve segment inexcitable. In some cases of multifocal motor neuropathy, for example, failure to maximally excite the involved segment sometimes gives a false impression of sustained conduction block after clinical recovery. This situation calls for the use of near-nerve needle stimulation to counter the increased threshold of the nerve fibers. Alternatively, stimulation at a more proximal, unaffected portion of the nerve segment will help: A muscle response, if elicited by a distant stimulation, proves the propagation of the nerve impulse across the lesion site despite its abnormally elevated threshold for local excitation (Fig. 35–25).173,192

Long and Short Segmental Nerve Conduction Studies Segmental Stimulation in Short Increments The inching technique, or short incremental stimulation along the affected segment,45,187,193 can isolate a lesion more precisely than ordinary conduction studies, which localize the approximate site of involvement. In studying a focal lesion such as compressive neuropathy, inclusion of

the unaffected segments lowers the sensitivity of the study, whereas short segmental stimulation helps localize an abnormality that may otherwise escape detection. With a nerve impulse conducting at a rate of 0.2 ms/cm (50 m/s) except for a 1-cm segment where it drops to 0.4 ms/cm, the conduction time over a 10-cm segment increases only 10%, from 2.0 ms to 2.2 ms, a change well within the normal range of variability. If measured over a 1-cm segment, the same 0.2-ms increase would constitute a 100% change, from 0.2 ms to 0.4 ms, signaling a clear abnormality. In this case, the large per-unit increase in latency more than compensates for the inherent measurement error associated with short incremental stimulation.46,129 Inching study helps not only to detect but also to precisely localize a focal pathology of the median nerve at the wrist,161,187,313,327 the ulnar nerve at the elbow,47,48,175 and the peroneal nerve at the knee.174 For example, stimulation of a normal median nerve with the advance of the cathode in 1-cm increments across the wrist causes latency changes ranging from 0.16 to 0.21 ms/cm from midpalm to distal forearm. In carpal tunnel syndrome, a sharply localized nonlinear latency increase averages 0.8 ms/cm at the compression site. This is nearly always accompanied by an abrupt change in waveform, which confirms a focal abnormality.187 Regardless of the nerve under consideration, waveform analysis can distinguish a pathologic latency change from apparent shifts that might have resulted from inaccurate advances of the stimulating electrodes or inadvertent spread of stimulus current. Even if technical difficulties prevent sequential stimulation near the site of lesion, successive responses above and below the affected zone can characterize local abnormality by forming two parallel lines rather than one (see Fig. 35–25). Together with abrupt waveform changes, this confirms a focal lesion within the short interval flanked by normal segments proximally and distally. Late Responses for Evaluation of Long Pathways Ordinary nerve conduction studies mainly assess the relatively short distal segments of the peripheral nerves. For a diffuse or multisegmental process, the longer the segment under study, the more evident is the conduction delay as a sum of all the segmental abnormalities. If a nerve impulse propagates at a rate of 0.2 ms/cm (50 m/s), for example, a 20% delay amounts only to 0.4 ms for a 10-cm segment. The same abnormality, if measured for a 100-cm segment, causes a 4.0-ms delay, an obvious difference. In addition, the same absolute error constitutes a smaller percentage in evaluating a longer as compared to shorter segment, improving the accuracy of latency and distance determination. Measuring the surface distance for a 10-cm segment, a 1-cm error constitutes 10% of the actual value, resulting in a calculated conduction velocity varying between 50 and 55 m/s. When measuring a 100-cm segment, the same 1-cm

Nerve Conduction and Needle Electromyography

FIGURE 35–25 A, Motor and sensory conduction studies of the left median nerve in a patient with multifocal motor neuropathy. The left diagram illustrates the consecutive magnetic resonance imaging sections in relation to the sites of stimulation at the wrist crease (A1) and at 2-cm increments more proximally. One horizontal division equals 5 ms (motor) or 2 ms (sensory), and one vertical division corresponds to the gain indicated at the end of each trace together with stimulus intensity. Note a complete and selective motor conduction block across the segment between A2 and A3, corresponding to the site of maximal nerve enlargement. (From Kaji, R., Oka, N., Tsuji, S., et al.: Pathological findings at the site of conduction block in multifocal motor neuropathy. Ann. Neurol. 33:152, 1993, with permission.) B, A repeat study in the same patient after return of strength of the median-innervated intrinsic hand muscles. High-intensity stimulation failed to excite the nerve along the affected segments (A5 through A8), mimicking a conduction block. More proximal stimulation at the elbow applied to the presumably normal nerve segments (A9 through A11), however, induced a series of temporally dispersed muscle responses associated with thumb abduction, indicating recovery of conduction. (From Kimura, J.: Facts, fallacies, and fancies of nerve conduction studies: twenty-first annual Edward H. Lambert lecture. Muscle Nerve 20:777, 1997, with permission.)

935

936

Nerve Conduction and Electromyography

error represents only 1% of the actual value, with a range of calculated conduction velocity between 50 and 50.5 m/s. The same argument holds in latency measurement, in which the same absolute error poses a smaller percent difference over a longer as compared to a shorter segment. In summary, the study of a longer path offers better sensitivity and accuracy and, as noted below, improves reproducibility in serial studies. Of a number of neurophysiologic methods that supplement assessment of longer pathways, those of general interest include the F wave and the H reflex. Reproducibility of Various Measures Nerve conduction studies offer a sensitive and objective indicator53 to document serial changes during the clinical course.205 Here the adherence to technical details plays an important role because any deviations from the standards result in inconsistencies of the results, especially in designing a multicenter clinical trial involving many participants of different training backgrounds.371 The assessment of reproducibility exploits two independent indices: the relative intertrial variation (RIV) and the intraclass correlation coefficient (ICC). Of the two, the RIV is derived as a variation of measurements expressed as a percentage of the difference between the two measures over the mean value of the two. Thus RIV (%)  100(V2  V1)/0.5(V1  V2) where V1 and V2 represent the values of the first and the second measurements of the pair, respectively. Ranges of RIV between 10% and 10% indicate an acceptably low variability. Measures having a larger among-subject difference show a greater intraindividual variability as well. Taking this into consideration, calculation of the ICC partially offsets the effects of a large interindividual variability as follows: ICC  s2/( s2  2) where s2 and 2 represent among-subject variance and experimental error. Values exceeding 0.9 indicate a reliable measure, although, as seen from the formula, this may result from a large among-subject variance rather than a small experimental error. Our own experience with a multicenter analysis dealt with a study of the intertrial variability of nerve conduction studies to determine the confidence limits of the variations in preparation for a drug trial.192,209 All measurements were repeated twice at a time interval of 1 to 4 weeks, using a standardized method. In all, 32 centers participated in the study of 132 healthy subjects (63 men) and 65 centers in the evaluation of 172 patients (99 men) with mild diabetic polyneuropathy. Motor nerve conduction studies included

the left median and tibial nerves, measuring amplitude, terminal latency, and minimal F-wave latency and calculating motor conduction velocity and F-wave conduction velocity. Sensory nerve conduction studies included the left median and sural nerves, measuring amplitude and latency of antidromic responses and calculating sensory conduction velocities over the distal segment. Figure 35–26 shows the ICC and the 5th to 95th percentiles of RIVs in both groups, and Figure 35–27 illustrates individual data from healthy subjects. Both the controls and the patient group showed the most variability in amplitude, followed by terminal latency and motor and sensory conduction velocity. The minimal F-wave latency showed the least variability, averaging only 10% for the median nerve and 11% for the tibial nerve in normals and 12% and 14%, respectively, in patients with mild diabetic polyneuropathy. The measures meeting the RIV criterion of less than 10% included F-wave latency and F-wave conduction velocity of the median and tibial nerves and sensory conduction velocity of the median nerve in both healthy subjects and patients. Similarly, the measures meeting the ICC criterion of over 0.9 included F-wave latency of the median and tibial nerves in both groups. In some amplitude measurements, a large among-subject variance concealed a large experimental error, leading to a high ICC despite a considerable variability as evidenced by a large RIV. These included median sensory nerve potential and median and tibial compound muscle action potentials. These findings suggest that a high ICC, indicating a statistical correlation between two measurements,87,384 does not necessarily imply a good reproducibility. Thus, to achieve an optimal comparison, sequential studies must exclude the use of any measurements with a wide RIV regardless of ICC values. Our data indicate that F-wave latencies of the median and tibial nerves meet the criteria for a reliable measure, showing a large ICC ( 0.9) with a small RIV ( 10%), provided, of course, that advanced stage of illness does not preclude adequate recording of F waves. When evaluating single patients against a normal range established in a group of subjects, however, F-wave conduction velocity suits better, minimizing the effect of limb length. Similarly effective is the use of a nomogram plotting the latency against height as an indirect measure of limb length. Clinical Considerations A question often posed in regard to the accuracy, sensitivity, and reliability of latency and velocity measurements relates to the length of the segment under study. Other factors being equal, should one study a shorter or longer segment for better results? The choice depends entirely on the pattern of the conduction change; short segmental approaches better suit a focal lesion than a longer distance, which tends to obscure the abnormality. In contrast, a longer segment reveals diffuse or multisegmental

Nerve Conduction and Needle Electromyography

937

FIGURE 35–26 Reproducibility of various measures in healthy volunteers (A) and patients with diabetic neuropathy (B). All studies were repeated twice at a time interval of 1 to 4 weeks to calculate relative intertrial variations as an index of comparison. (From Kimura, J.: Facts, fallacies, and fancies of nerve conduction studies: twenty-first annual Edward H. Lambert lecture. Muscle Nerve 20:777, 1997; data courtesy of Kohara et al.,209 from a multicenter reliability study sponsored by Fujisawa Pharmaceutical Co., Ltd.)

abnormalities better than a short distance, which tends to reduce sensitivity, increase measurement errors, and lower reproducibility. In summary, short distances magnify focal conduction abnormalities despite increased measurement error, and long distances, though insensitive to focal lesions, accumulate diffuse or multisegmental abnormalities for better sensitivity, accuracy, and reliability.

Needle Electromyography Electromyography tests the motor system as an extension of the physical examination, rather than a laboratory procedure.68,343 It helps to localize the site of a lesion that may cause muscle weakness into upper and lower motor neurons, neuromuscular junction, and muscle (Fig. 35–28). Each

muscle of interest is tested at rest and during degrees of voluntary contraction. The patient’s symptoms and signs guide the optimal selection of specific muscle groups.128,293 The findings in the initially tested muscles determine the subsequent course of exploration. This strategy precludes a routine approach with a predetermined list of muscles for a given condition. Denervated muscle fibers fire at rest independent of volitional neural control, giving rise to spontaneous singlefiber potentials, which constitute one of the most distinct findings in electromyography. The motor unit and its electrical counterpart, the motor unit potential, represent the smallest functional element of volitional contraction. Diseases of the nerve or muscle cause structural or functional disturbances leading to alterations in waveform and discharge patterns. Recorded signals also reflect other biologic and nonbiologic factors, such as the characteristics of field, which serve as the volume conductor. In particular,

938

Nerve Conduction and Electromyography

FIGURE 35–27 Comparison between the first and the second measures of median nerve motor conduction velocity (A) and F-wave latency (B). Individual values plotting the first study on the abscissa and second study on the ordinate show a greater reproducibility of the F-wave latency compared to the motor nerve conduction velocity (cf. Fig. 35–26). (From Kimura, J.: Facts, fallacies, and fancies of nerve conduction studies: twenty-first annual Edward H. Lambert lecture. Muscle Nerve 20:777, 1997; data courtesy of Kohara et al.,209 from a multicenter reliability study sponsored by Fujisawa Pharmaceutical Co., Ltd.)

the spatial relationship between the source and the tip of the needle electrode substantially impacts the size of an action potential. When recorded using an electrode with a small leadoff surface, the amplitude falls off sharply, often to less than 10%, at a distance of 1 mm from the generator.

A patient treated with anticoagulants should not undergo needle study unless appropriate laboratory tests exclude an intolerable bleeding tendency. Our criterion calls for a partial thromboplastin time less than 1.5 times the control value with heparin infusion, and an international normalized ratio of 2.0 or less with warfarin (Coumadin) therapy.

FIGURE 35–28 Schematic view of the motor system with seven anatomic levels. They include upper motor neuron from the cortex to the spinal cord (I and II), lower motor neuron with the anterior horn cell and nerve axon (III and IV), neuromuscular junction (V), and muscle membrane and contractile elements (VI and VII). (Adapted from Netter, F. H.: The Ciba Collection of Medical Illustrations. Vol. 1: Nervous System. Part 1: Anatomy and Physiology. Summit, NJ, Ciba Pharmaceutical Co., Medical Education Division, 1983.) The inset illustrates diagrammatically the four steps of electromyographic examination and normal findings. (Figure from Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

939

Nerve Conduction and Needle Electromyography

Patients with other coagulopathies, such as hemophilia, must have the same precautions.325 Local pressure can usually accomplish hemostasis with a platelet count above 20,000/L for thrombocytopenia.24 In questionable cases, we first test a superficial muscle, where external compression can counter bleeding adequately. This also helps determine the degree of bleeding tendency and the feasibility of further study of deeper muscles. Electromyographic needle explorations often render the muscle unsuitable for a subsequent biopsy examination. Prior needle studies also make it difficult to interpret an elevated serum level of creatine kinase, which may signal certain muscle diseases such as muscular dystrophy and polymyositis. Other conditions associated with this finding include cardiac ischemia, hypothyroidism, and sustained athletic participation. Needle examination by itself should not elevate the enzyme to a misleading level, although, when combined with diurnal variation and prolonged exercise,30,290 it may alter the level considerably even in normal muscles. Electromyographic examination comprises the following four steps: 1. Needle insertion to find the muscle and assess the magnitude of injury potential 2. Observation of spontaneous activity recorded with the needle stationary in a relaxed muscle 3. Study of individual motor unit potentials activated by mild voluntary contraction

4. Evaluation of recruitment during progressively stronger effort and interference patterns at maximum level of contraction Adequate sampling calls for frequent repositioning of the needle electrode in small steps because a needle electrode registers muscle action potentials only from a restricted area. A single puncture site allows exploration in four directions to minimize patient discomfort. In studying larger muscles, however, additional insertions should include proximal, central, and distal portions. Electromyography helps categorize motor dysfunction into upper and lower motor neuron disorders and myogenic lesions, as illustrated schematically in Figure 35–29, which emphasizes typical findings at the risk of oversimplification.

POTENTIALS ASSOCIATED WITH NEEDLE INSERTION Injury Potentials A brief burst of electrical discharges follows movement of a needle electrode inserted into the muscle, lasting on average a few hundred milliseconds. The magnitude of the response primarily depends on the extent and speed of needle movement. Semiquantitative analysis of this activity provides an important measure of muscle excitability, typically reduced in fibroses and exaggerated in denervated or

FIGURE 35–29 Typical findings in lower and upper motor neuron disorders and myogenic lesions. Myotonia shares many features common to myopathy in general, in addition to myotonic discharges triggered by insertion of the needle or with voluntary effort to contract the muscle. Polymyositis shows combined features of myopathy and neuropathy, including (1) prolonged insertional activity; (2) abundant spontaneous discharges; (3) small-amplitude, short-duration, polyphasic motor unit potentials; and (4) early recruitment leading to low-amplitude, full-interference pattern. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

940

Nerve Conduction and Electromyography

inflammatory muscles. A reduced insertional activity may also signal functional inexcitability of muscle fibers, for example, during attacks of familial periodic paralysis. An irritable muscle with instability of the muscle membrane may show an abnormally prolonged insertional activity outlasting the cessation of needle movement, as might be seen in denervation, myotonic disorders, or myositis. A sustained run of positive waves may follow insertional activity during early stages of denervation, 10 days to 2 weeks after nerve injury. These early abnormalities of denervation are seen before the appearance of spontaneous activity. A normal insertional activity may also take the form of a few repetitive positive sharp potentials, but denervated muscles often show reproducible trains lasting several seconds to minutes after cessation of the needle movement (Table 35–19).

End-Plate Activities In normal resting muscles, a needle placed at the end-plate region detects end-plate activities that consist of two components, occurring conjointly or independently: low-amplitude, undulating end-plate noise and high-amplitude intermittent spikes (Fig. 35–30). The patient typically reports a dull pain in association with end-plate activities, which dissipates with slight withdrawal of the needle. The end-plate noise represents extracellularly recorded miniature end-plate potentials, which frequently recur as irregular negative potentials, 10 to 50 V in amplitude and 1 to 2 ms in duration. Over the loudspeaker, it characteristically sounds much like a seashell held to the ear. If this depolarization reaches the threshold, single muscle fibers fire repetitively, giving rise to the end-plate

spikes.43,148 These spikes, 100 to 200 V in amplitude and 3 to 4 ms in duration, fire irregularly at 5 to 50 Hz and typically show a diphasic waveform with initial negativity because the discharges originate at the tip of the recording electrode. Fibrillation potentials, when recorded at the endplate region, also have initial negativity, although they fire more regularly at a slower rate.

Myotonic Discharge Myotonia and its electrical counterpart, myotonic discharge, represents a sustained muscle fiber discharge outlasting the external source of excitation (Table 35–20). The disorders associated with this clinical finding include myotonia congenita, myotonia dystrophica, paramyotonia congenita,379 and hyperkalemic periodic paralysis.38 Clinical myotonia usually results from myotonic discharges triggered by a minimal voluntary contraction or mechanical stimulation with a tap on the muscle belly. Electromyography, however, may uncover myotonic discharges without manifest myotonia in a variety of disorders, such as polymyositis, type II glycogen storage disease with acid maltase deficiency,159 cytoplasmic body myopathy resembling myotonic dystrophy,260 and other conditions with chronic denervation. Depending on the spatial relationship of the discharging muscle fibers to the tip of the needle electrode, myotonic discharge takes one of two forms: (1) negative spikes with a small initial positivity or (2) positive sharp waves, each followed by a low-amplitude negative component of much longer duration (Fig. 35–31). The negative and positive waveforms, like those of denervation, represent recurring

Table 35–19. Origin of Spontaneous and Triggered Discharges Muscle Fiber Insertional positive waves Myotonic discharge Fibrillation potential Positive sharp waves Complex repetitive discharge End-plate noise End-plate spikes

Briefly sustained single muscle fiber discharges triggered by needle movement Repetitive single muscle fiber discharges triggered by needle movement Spontaneous single muscle fiber discharges, negative type Spontaneous single muscle fiber discharges, positive type A group of ephaptically activated spontaneous single muscle fiber discharges Miniature end-plate potentials recorded extracellularly at motor point Single muscle fiber discharges triggered by needle movement at motor point

Lower Motor Neuron Fasciculation potential Myokymic discharge Neuromyotonic discharge Cramp discharges Hemifacial spasm Hemimasticatory spasm

Spontaneous motor unit discharges, involving a single unit, totally or fractionally Clusters of repetitive firing of the same motor unit, usually from demyelination Continuous high-frequency discharges, involving many motor units Briefly sustained high-frequency discharges, involving many motor units Intermittent unilateral contraction of facial muscles, either idiopathic or subsequent to Bell’s palsy Intermittent unilateral contraction of masseter muscle

Upper Motor Neuron Stiff-man syndrome Involuntary movement

Sustained contraction of motor units in many agonistic and antagonistic muscles Tremor, chorea, hemiballismus, athetosis, dystonia, myoclonus, epilepsia partialis continua

From Kimura, J.: Needle electromyography. In Bertorini, R. (ed.): Evaluation and Diagnostic Tests in Neuromuscular Disorders. Woburn, MA, Butterworth-Heinemann, p. 331, 2002, with permission.

Nerve Conduction and Needle Electromyography

941

FIGURE 35–30 End-plate activities recorded from the tibialis anterior in a healthy subject. Two types of potentials shown represent the initially negative, high-amplitude end-plate spikes (a, b, and c) and low-amplitude end-plate noise (g, h, and i). The spikes and end-plate noise usually, though not necessarily, appear together (d, e, and f). (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

single-fiber potentials recorded from an intact and an injured muscle membrane. Thus the negative spikes, resembling fibrillation potentials, tend to occur at the beginning of slight volitional contraction, whereas the positive sharp waves are usually initiated by needle insertion. In both forms of discharge, amplitude waxes and wanes over the range of 10 V to 1 mV, often, though not always, showing inversely varying firing rate within the range of 50 to 100 Hz. This fluctuating pattern of discharge gives the characteristic noise over the loudspeaker, reminiscent of an accelerating or decelerating motorcycle or chain saw.

SPONTANEOUS ACTIVITY Table 35–21 summarizes types of spontaneous activities commonly seen in electromyography. These include fibrillation potentials, positive sharp waves, complex repetitive

discharges, fasciculation potentials, and myokymic discharges. Complex repetitive and myokymic discharges and fasciculation potentials cause isolated or briefly sustained visible muscle twitches over a localized area, whereas fibrillation potentials and positive sharp waves do not. A spontaneously activated single muscle fiber produces fibrillation potentials and positive sharp waves.39,67,68,82,86,213,214 Rapid firing of a cluster of muscle fibers in sequence gives rise to complex repetitive discharges. These discharges are driven ephaptically at a point of lateral contact between neighboring units.93,346 One fiber serves as a pacemaker, regulating the frequency and pattern of discharge by the rate of its rhythmic depolarization and circus movements of currents among muscle fibers.167 Fasciculation potentials result from isolated spontaneous discharges of a motor unit, and myokymic discharges from repetitive firing of a motor unit, as the name grouped fasciculation potentials indicates. These discharges are associated

942

Nerve Conduction and Electromyography

Table 35–20. Disorders with Myotonic Discharges With Clinical Myotonia Myotonia dystrophica Myotonia congenita Paramyotonia congenita Hyperkalemic periodic paralysis Proximal myotonic myopathy Without Clinical Myotonia Myositis Acid maltase deficiency Cytoplasmic body myopathy Hyperthyroidism Hypothyroidism Familial granulovacuolar lobular myopathy Malignant hyperpyrexia Multicentric reticulohistiocytosis Myopathies induced by Glycyrrhizin Hypocholesterolemic agent Colchicine From Kimura, J.: Needle electromyography. In Bertorini, R. (ed.): Evaluation and Diagnostic Tests in Neuromuscular Disorders. Woburn, MA, Butterworth-Heinemann, p. 331, 2002, with permission.

with local muscle twitches, called fasciculation, and sustained segmental contractions seen in cramp syndromes, called myokymia. Similarly, continuous motor unit discharges give rise to generalized muscle spasms or the syndrome of neuromyotonia, representing peripheral nerve hyperexcitability. In contrast, involuntary muscle contractions seen in patients with the stiff-man syndrome result from spontaneous discharges originating in the central nervous system. Semiquantitating each of these spontaneous activities can provide a numeric grading as follows: 1—Rare spontaneous potentials recordable in one or two sites only after some search. This category includes insertional positive sharp waves or abnormally sustaining injury potentials elicited after moving the needle electrode. 2—Occasional spontaneous potentials registered in more than two different sites. 3—Frequent spontaneous potentials recordable regardless of the position of the needle electrode. 4—Abundant spontaneous potentials nearly filling the screen of the oscilloscope.

Fibrillation Potentials Fibrillation potentials have diphasic or triphasic waveforms with initial positivity (Fig. 35–32) ranging from 1 to 5 ms in duration and from 20 to 500 V in amplitude, when recorded with a concentric needle electrode.43 Fibrillation potentials remain the same in shape from

FIGURE 35–31 Myotonic discharges from the right anterior tibialis in a 39-year-old man with myotonic dystrophy. The tracings show two types of discharges: trains of positive sharp waves (a, b, and c) and negative spikes with initial positivity (d, e, and f). The discharges in a and d reveal a waxing and waning quality. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

the first to the last discharges of a train, and have the same shape and amplitude as voluntarily activated singlefiber potentials when recorded with single-fiber electromyography.342 All these findings support the view that fibrillation potentials represent single muscle fibers or the smallest unit recordable by the needle electrode. A crisp clicking noise that fibrillation potentials produce over the loudspeaker resembles the sound caused by wrinkling tissue paper. Spontaneous oscillations in the membrane potential probably regulate the firing pattern in the range of 1 to 30 Hz, with an average rate of 13 Hz.354 Occasional fibrillation potentials may discharge irregularly in the range of 0.1 to 25 Hz.37,256 In profound denervation, discharges from more than one fiber may give a false impression of a very irregular firing pattern.

Positive Sharp Waves Positive sharp waves also result from single-fiber activation despite their sawtooth appearance, much lower in amplitude and longer in duration compared to fibrillation potentials

Nerve Conduction and Needle Electromyography

Table 35–21. Common Types of Spontaneous Discharges Fibrillation Potentials and Positive Sharp Waves Neuropathic condition Muscular dystrophy Myositis Complex Repetitive Discharges Motor neuron disease Radiculopathy Chronic polyneuropathy Polymyositis Muscular dystrophy Myxedema Schwartz-Jampel syndrome Fasciculation Potentials Motor neuron disease Radiculopathy Entrapment neuropathy Muscular pain–fasciculation syndrome Healthy subjects Myokymic Discharges Guillain-Barré syndrome Radiation plexopathy Spinal stenosis Nerve root compression Bell’s palsy Multiple sclerosis Syringobulbia From Kimura, J.: Needle electromyography. In Bertorini, R. (ed.): Evaluation and Diagnostic Tests in Neuromuscular Disorders. Woburn, MA, Butterworth-Heinemann, p. 331, 2002, with permission.

(see Fig. 35–32). In general, the absence of a negative spike implies recording near the damaged part of the muscle fiber. If the inserted needle penetrates the membrane, sustained standing depolarization precludes the generation of a negative spike at this point. Therefore, a propagating action potential only approaches the site of injury, giving rise to a sharp positive discharge, followed by low-amplitude negative deflection in the absence of a negative spike. As discussed earlier, positive sharp waves may form part of myotonic discharges, usually triggered by insertion of the needle and less often by mild voluntary contraction. Although both represent single-fiber discharges, denervation potentials occur spontaneously, whereas myotonic discharges appear in response to mechanical stimuli or following voluntary contraction. This distinction, however, may often blur, because needle movements also enhance spontaneous discharges associated with denervation.

Single-Fiber Discharge and Denervation Spontaneous single-fiber activity, in the appropriate clinical setting, usually signals disorders of the lower motor neuron as one of the most useful signs of abnormality in

943

clinical electromyography. Its presence implies denervation, although it is otherwise nonspecific, being seen in such diverse diseases as anterior horn cell diseases, radiculopathies, plexopathies, and axonal polyneuropathies. Its absence, however, proves little during the latent period of 2 to 3 weeks after nerve injury. The distribution of spontaneous potentials helps localize lesions of the spinal cord, root, plexus, and peripheral nerve. Certain myopathic processes, such as muscular dystrophy,43 dermatomyositis, and polymyositis, also show spontaneous discharges, possibly because of muscle necrosis75 or focal degeneration,334 which separates a part of the muscle fiber from the end-plate region, thus denervating that portion. Alternatively, spontaneous activity may also result in polymyositis from increased membrane irritability or inflammation of intramuscular nerve terminals.307 Other conditions less consistently associated with fibrillation potentials include diseases of the neuromuscular junction,123 facioscapulohumeral dystrophy, limb-girdle dystrophy, oculopharyngeal dystrophy,149 myotubular (centronuclear) myopathy,339 and trichinosis.378

Complex Repetitive Discharges Complex repetitive discharges result from a group of muscle fibers that fire repetitively in the same sequence, producing the sound of a machine gun over the loudspeaker. The entire epoch, ranging from 50 V to 1 mV in amplitude and up to 50 to 100 ms in duration (Fig. 35–33), repeats itself at slow or fast rates, usually in the range of 5 to 100 Hz. These discharges typically have a sudden onset, a constant rate of firing for a short period, and an abrupt cessation. The polyphasic waveform remains identical from one burst to the next until a sudden shift to a new pattern of discharge intervenes. The unique repetitive features once prompted the use of the now discarded terms bizarre high frequency discharges and pseudomyotonia. The uniform firing pattern showing a design-like waveform makes these discharges distinct from myokymia, neuromyotonia, and cramp syndromes despite their superficial resemblance. This discharge usually indicates a wide variety of chronic denervating conditions. These include motor neuron disease, radiculopathy, chronic polyneuropathy, myxedema, and the Schwartz-Jampel syndrome sometimes associated with neurogenic muscle hypertrophy,319 as well as certain myopathies, such as muscular dystrophy and polymyositis. An overall analysis in a large series93 revealed the highest prevalence in Duchenne type muscular dystrophy, spinal muscular atrophy, and Charcot-Marie-Tooth disease. In women with urinary retention, the striated muscle of the urethral sphincter may show profuse activity of this type.113,114 Some apparently healthy subjects may have incidental findings of complex repetitive discharges, probably indicating the presence of clinically silent foci of an irritative process that tends to involve deeper muscles in general and the iliopsoas in particular.

944

Nerve Conduction and Electromyography

FIGURE 35–32 A, Spontaneous single-fiber discharges recorded from the denervated tibialis anterior in a 67-year-old man with acute onset of footdrop. Note gradual alteration of the waveform from triphasic spike with major negativity to paired positive potentials and finally to a single positive sharp wave over the time course of some 8 seconds without movement of the needle. This fortuitous recording provides direct evidence that the same single-fiber discharge can be recorded either as fibrillation potentials or positive sharp waves. Long-duration positive deflections seen in c, f, and g represent a pulse artifact. (From Kimura, J.: Electromyography and Nerve Stimulation Techniques: Clinical Applications [in Japanese]. Tokyo, Igaku-Shoin, 1990, with permission.) B, Spontaneous single-fiber activity of the anterior tibialis in a 68-year-old woman with amyotrophic lateral sclerosis. The tracings show two types of discharges: positive sharp waves (a, b, and c) and fibrillation potentials (d, e, and f). C, Spontaneous single-fiber activity of the paraspinal muscle in a 40-year-old man with radiculopathy, consisting of positive sharp waves (a, b, and c) and fibrillation potentials (d, e, and f). D, Spontaneous single-fiber activity of the deltoid (a, b, and c) and tibialis anterior (d, e, and f) in a 9-year-old boy with a 6-week history of dermatomyositis, with two types of discharges: positive sharp waves (a, b, and c) and fibrillation potentials (d, e, and f). E, Spontaneous single-fiber activity of the tibialis anterior in a 7-year-old boy with Duchenne type muscular dystrophy, showing positive sharp waves (a, b, and c) and fibrillation potentials (d, e, and f). (B–E from Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.) Figure continued on opposite page

Nerve Conduction and Needle Electromyography

945

FIGURE 35–32 Continued

Fasciculation Potentials and Myokymic Discharges Fasciculation potentials, or spontaneous discharges of a single motor unit either as a whole or possibly in part, usually, but not necessarily, induce visible twitches depending on the depth of discharge. In some instances, therefore, electromyography reveals fasciculation potentials in the absence of a clinical counterpart. Normal voluntary motor unit potentials maintain constant shape if the needle remains stationary, whereas fasciculation potentials undergo changes in amplitude and waveform from time to time. Discharge characteristics help discriminate semirhythmic voluntary units firing at a rate ranging from 5 to 10 Hz and irregular fasciculation potentials showing a much slower rate. In questionable cases, attempts to volitionally control the discharge pattern by mildly contracting agonistic or antagonistic muscles alters the firing rate of voluntary units but not fasciculation potentials. Although the neural discharge responsible for fasciculation potentials may originate in the spinal cord or anywhere along the length of the peripheral nerve,380 the existing evidence favors a very distal site of origin at or near the motor terminals.227 Fasciculation potentials are seen not only in diseases of anterior horn cells, but also in other neuropathic conditions such as radiculopathy, entrapment neuropathy, and the muscular pain–fasciculation syndrome,160 as well as in healthy subjects. Patients with cervical spondylotic myelopathy may have fasciculation potentials in the lower limbs. Possible underlying

mechanisms include loss of inhibition, vascular insufficiency, cord traction, and denervation, although none of these hypotheses has anatomic or physiologic evidence. Spontaneous discharges abate after cervical decompression.172 Fasciculation potentials abound in some metabolic derangements, such as tetany, thyrotoxicosis, and overdoses of anticholinesterase medication.67 Multiple units may discharge simultaneously in amyotrophic lateral sclerosis and progressive spinal muscular atrophy, but also in other degenerative diseases of the anterior horn cells, such as poliomyelitis and syringomyelia. Synchronous spontaneous discharges may simultaneously involve muscles supplied by different nerves or occur in homologous muscles on opposite sides. This pattern may possibly suggest an intraspinal mechanism, or a reflex origin via spindle afferent triggered by the arterial pulse.317 Healthy subjects may have either single or grouped spontaneous discharges in otherwise normal muscles,258 sometimes, but not always, associated with cramps. A number of investigators have sought to differentiate these benign fasciculations without ominous implication from those of progressive motor neuron disease. No single method has emerged based on waveform characteristics such as amplitude, duration, and number of phases.364 The frequency criteria, however, may possibly serve to separate the two categories: irregular firing at an average interval of 3.5 seconds in motor neuron disease versus 0.8 seconds in asymptomatic individuals.71,364 Also, fasciculation potentials characteristically arise proximally early and occur distally in the later stages of the disease in amyotrophic lateral

946

Nerve Conduction and Electromyography

Continuous Muscle Fiber Activity

FIGURE 35–33 Complex repetitive discharges of the left quadriceps in a 58-year-old man with herniated lumbar disc. The tracings show two types of discharges: trains of single- or double-peaked negative spikes (a, b, and c) and complex positive sharp waves (d, e, and f). In f, each sweep, triggered by a recurring motor unit potential, shows remarkable reproducibility of the waveform within a given train. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

sclerosis.71 In conclusion, fasciculation potentials by themselves cannot provide an absolute proof of abnormality, unless accompanied by either fibrillation potentials or positive sharp waves, indicating disease of the lower motor neuron. In contrast to isolated discharges, more complex bursts of repetitive discharges of one motor unit cause vermicular movements of the skin, called myokymia.61 Repetitive firing of the same motor units typically comprises 2 to 10 spikes discharging at 30 to 40 Hz in each burst (Fig. 35–34), which repeats at fairly regular intervals of 0.1 to 10 seconds. Myokymic discharges often involve facial muscles in patients with brainstem glioma or multiple sclerosis (see Table 35–21). When present in the limb muscles, they usually favor the diagnosis of certain chronic demyelinative neuropathic processes, such as multifocal motor neuropathy, GuillainBarré syndrome,249 and radiation plexopathies.3,6,70 Hyperventilation enhances myokymic bursts generated ectopically in a demyelinated segment of motor fibers because induced hypocalcemia amplifies axonal excitability.26

A heterogeneous group of central or peripheral disorders shows prolonged spontaneous motor unit activity, or continuous muscle fiber activity.89 In stiff-man syndrome, a rare, well-recognized but poorly characterized entity, sustained involuntary discharges of central origin produce a continuous interference pattern in the agonists as well as antagonists simultaneously. These motor unit potentials abate with peripheral nerve or neuromuscular block, after spinal or generalized anesthesia, or during sleep. The administration of diazepam, but not phenytoin or carbamazepine, also abolishes or attenuates the activity. Neuromyotonia implies continuous muscle fiber activity of peripheral origin126 seen in various, probably related, disease entities such as Isaacs’ syndrome, quantal squander, generalized myokymia, pseudomyotonia, and normocalcemic tetany.166,271 In these syndromes, clinical and electrophysiologic presentations vary, although they all show the feature of sustained spontaneous discharges generated at various sites of axons, from proximal segments to the intramuscular terminals.243,316,360 Involuntary motor activity continues during sleep and after general or spinal anesthesia but abates totally with neuromuscular blocking, confirming its peripheral origin. Procaine injection of the peripheral nerve will abolish the abnormal activity only if the discharges originate in the more proximal segment. Undulating movements of the overlying skin and a delay of relaxation after muscle contraction result from sustained firing of motor units characteristically discharging at frequencies as high as 300 Hz. The firing motor unit potentials produce a typical “pinging” sound over the loudspeaker. An increasing number of single muscle fibers fail to follow the high rate of repetitive firing pattern, leading to a decline in amplitude slowly or rapidly. Ischemia and electrical nerve stimulation, but not voluntary contraction, accentuate the high-frequency discharge. Phenytoin or carbamazepine effectively reduces involuntary movements in most patients with neuromyotonia.

Cramp Cramp may commonly develop in normal subjects but also occasionally as a sign of abnormality in patients with neuromuscular disorders. Following severe cramps, the pain may persist for days. This briefly sustained involuntary muscle contraction probably results from spontaneous impulses that originate in the peripheral nerve. Although the exact underlying mechanism remains unknown, mechanical excitation of motor nerve terminals during muscle shortening may serve as a trigger.226,227,276 Spinal or general anesthesia has no effects, but peripheral nerve block often abolishes the activity. The cramp discharges comprise repetitive discharges of normal motor unit potentials at a high rate, in the range of 200 to 300 Hz. They begin with single potentials or

Nerve Conduction and Needle Electromyography

947

FIGURE 35–34 A, Myokymic discharges in a 21-year-old woman with multiple sclerosis. The patient had visible undulating movement of the facial muscles on the right associated with characteristic bursts of spontaneous activity recorded from the orbicularis oris (a, b, c, and d) as well as orbicularis oculi (e, f, g, and h). In d, each sweep, triggered by a recurring spontaneous potential, shows a repetitive but not exactly time-locked pattern of the waveform. B, Myokymic discharges in a 57-year-old man with a 2-week history of Guillain-Barré syndrome and a nearly complete peripheral facial palsy. Despite the absence of visible undulating movement, rhythmically recurring spontaneous discharges appeared in the upper (a, b, and c) and lower (d, e, and f) portions of the left orbicularis oris. In c and f, each sweep triggered by a recurring spontaneous potential shows the repetitive pattern. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

doublets, and gradually spread to involve other areas of a muscle. The discharges may wax and wane for several minutes, involving several different sites that may be activated simultaneously or sequentially, and then abate spontaneously.

MOTOR UNIT ACTION POTENTIAL A group of muscle fibers innervated by a single anterior horn cell constitutes a motor unit, with anatomic and physiologic properties determined by the innervation ratio, fiber density, propagation velocity, and integrity of neuromuscular transmission. Motor unit potentials vary not only from one muscle group to another but also with age for a given muscle. According to principal components analysis, three elements determine 90% of the variance of the data

set: changes in the size of the motor units, variations in the arrival time at the recording electrode, and loss of muscle fibers within the motor unit territory.264 Other physiologic factors that influence the shape of motor unit potentials include the resistance and capacitance of the intervening tissue and intramuscular temperature.18,73 In addition, the spatial relationships between the needle and individual muscle fibers critically determine the waveform.33 The measures that define a motor unit potential include amplitude, rise time, duration, phases, stability, and territory (Fig. 35–35). A wide variety of neuromuscular disorders cause characteristic combinations of abnormalities, which help distinguish primary muscle diseases from disorders of the neuromuscular junction and lower motor neurons (Table 35–22). A random loss of individual fibers seen in myopathies reduces the spike duration and amplitude of motor unit potentials.42 In contrast, an increased

948

Nerve Conduction and Electromyography

Turn

Baselineto-peak amplitude

Phase

Late component

Peakto-peak amplitude

weakness in the diseases of nerve and muscle.93,146 This, taken together with abnormalities of insertional and spontaneous activities, and recruitment pattern, helps delineate a characteristic abnormality. In addition, serial assessment provides a good means to monitor the disease process based on the established correlation between physiologic and histologic alterations of the motor unit.341

Amplitude and Area 200μV 500μs Risetime Duration

FIGURE 35–35 Profile of motor unit potential determined by (1) peak-to-peak or baseline-to-peak amplitude; (2) rise time, or the time interval from the onset of a change to its peak; (3) turn, or point of change in direction; (4) phase, or portion of a waveform between departure from and the return to the baseline; and (5) late component, or satellite potential, separated from the main motor unit potential. (Courtesy of David Walker, M.S.E.E.; from Kimura, J.: Needle electromyography. In Bertorini, R. [ed.]: Evaluation and Diagnostic Tests in Neuromuscular Disorders. Woburn, MA, Butterworth-Heinemann, p. 331, 2002, with permission.)

fiber density with sprouting after a loss of axons results in a larger potential in neuropathies or anterior horn cell diseases. Thus changes in the size of the motor unit potential show characteristic dichotomy in the classification of

Table 35–22. Types of Motor Unit Potentials Brief Duration, Small Amplitude, Early Recruitment Muscular dystrophy Congenital myopathy Metabolic myopathy Endocrine myopathy Myositis Myasthenia gravis Myasthenic syndromes Long Duration, Large Amplitude, Late Recruitment Motor neuron disease Radiculopathies Plexopathies Polyneuropathies Entrapment syndromes From Kimura, J.: Needle electromyography. In Bertorini, R. (ed.): Evaluation and Diagnostic Tests in Neuromuscular Disorders. Woburn, MA, Butterworth-Heinemann, p. 331, 2002, with permission.

Of all the individual muscle fibers of a motor unit that fire in near-synchrony, only a limited number located within the recording radius of the electrode contribute to the amplitude of a motor unit potential. Thus the position of the needle electrode relative to the discharging unit alters the recorded amplitude greatly. With the use of a concentric needle, for example, the amplitude falls off to less than 50% at a distance of 200 to 300 m from the source and to less than 1% a few millimeters away.92,139 To avoid a distant unit, it is recommended to select a motor unit potential with a short rise time of 500 s or less, which guarantees its proximity to the recording surface (see Fig. 35–35). Histologic studies reveal that only a few single muscle fibers lie within the recording radius from the tip of the needle. Thus the high-voltage spike of the motor unit potential results from fewer than 5 to 10 muscle fibers lying within a 500-m radius of the electrode tip.342,355 Computer simulation indicates that the proximity of the closest muscle fiber to the recording electrode determines the amplitude.85,266,347 Therefore, the muscle fibers lying closer together within the recording surface of the needle give rise to a higher amplitude. These findings indicate that, in general, the measure of amplitude aids in determining the muscle fiber density, not the motor unit territory. Many different profiles of motor unit potentials appear from the same motor unit with repositioning of the recording electrode. Motor unit potentials normally vary in amplitude from several hundred microvolts to a few millivolts. Recording with a monopolar needle tends to enhance the average amplitude of motor unit potentials compared to a concentric needle, which has a smaller recording radius.207 The “thickness” of the potential or ratio between area and amplitude varies much less with changes in electrode position.262 The combination of amplitude and area/amplitude ratio further improves discrimination, detecting around 70% of neurogenic changes compared to only 15% to 30% by duration criteria alone.337

Rise Time The rise time, or time lag from the initial positive peak to the subsequent negative peak, serves as a measure of proximity of the signal to the recording tip of the electrode. A greater rise time indicates a distant unit because

Nerve Conduction and Needle Electromyography

949

the resistance and capacitance of the intervening tissue act as a high-frequency filter. A distant motor unit also has a dull sound, as an audio counterpart of a long rise time. Repositioning the electrode closer to the source shortens the rise time, adding a high-pitched, crisp sound over the loudspeaker. In general, a rise time less than 500 s ensures that the electrode tip lies within the motor unit territory,164 although some argue for less restrictive criteria.16 Most electromyographers depend on a sharp, crisp sound over the loudspeaker as an important clue for the proximity of the discharging unit to the electrode. The measurement of the rise time determines the suitability of the recorded potential for inclusion in quantitative assessment.

loss of fibers (Fig. 35–37). These polyphasic potentials are also found in a healthy muscle, constituting between 5% and 15% of the total population, when recorded with a concentric needle electrode. The percentage of polyphasic activities increases with the use of a monopolar needle, although the exact incidence is not known. Some action potentials show several “turns,” defined as directional changes, without crossing the baseline. These serrated or, less appropriately, complex or pseudopolyphasic potentials have the same implication as polyphasic units, signaling desynchronization among discharging muscle fibers. In one study,389 these irregular potentials appeared more commonly in acute processes.

Duration

Abnormalities of Motor Unit Potentials

The duration of a motor unit potential, measured from the initial takeoff to the return to the baseline (see Fig. 35–35), reflects the length, conduction velocity, and degree of synchrony of many individual muscle fibers of the discharging motor unit.83 Although only a very small number of muscle fibers near the electrode determine the spike amplitude, a greater number of muscle fibers contribute to the duration of a motor unit potential. A concentric needle has a recording surface with an uptake area extending about 2.5 mm from the core.266,344 This allows distant units, not contributing to the amplitude of the negative spike, to add to the time of the initial and terminal positivity of the motor unit, increasing its duration. Thus the motor unit duration serves as a measure of a larger part of the muscle fiber population than the amplitude, although it still falls short of covering the entire motor unit territory, which measures 1 to 2 cm. In contrast to the amplitude, the duration remains relatively unchanged with a slight shift or rotation of the needle,263 normally varying from 5 to 15 ms, depending on the age of the subject (Table 35–23).32 With the use of a wide-open amplifier bandpass combined with enhanced signal-to-noise ratio, the duration increases substantially, approaching 30 ms. Under this circumstance, either a single-fiber electrode or macroelectrode seems to register the total time of a single action potential from end-plate zone to musculotendinous junction as the overall duration of the motor unit action potential.84

The contrasting abnormalities of the motor unit potential in myopathies versus lower motor neuron disorders occur as a common feature in a number of diseases in the respective categories. Thus such abnormalities per se cannot confirm a specific diagnosis, although they provide supportive evidence for the clinical impression. A majority of motor unit potentials have two or three phases in normal muscles. The number of polyphasic units having four or more phases increases in myopathy, neuropathy, and motor neuron disease (see Fig. 35–37). In these motor units, abnormal temporal dispersion of muscle fiber potentials probably results from unequal conduction time along either the nerve terminal or muscle fibers. Satellite potentials or extrapotentials clearly separated from the main unit have the same implications as polyphasic activities,69,362 occurring five times more often in patients with neuropathy and myopathy as compared to normal subjects.106 Motor unit potentials, if recorded at all during neurapraxia or an acute stage of axonotmesis, have normal waveforms, which result from unaffected axons. Motor units activated by voluntary effort normally fire semirhythmically, showing nearly identical configuration for successive discharges. Fatigue causes irregularity and reduction in the firing rate of a motor unit, without altering its waveform. With defective neuromuscular transmission, a repetitively firing motor unit fluctuates in amplitude and waveform from intermittent blocking and increased jitter of the constituent single-fiber potentials.287 Waveform variability of the motor unit potential plays an important role in evaluating deficiency of neuromuscular transmission, especially in proximal muscles not readily accessible to repetitive nerve stimulation. Instability of a motor unit potential, termed jiggle,345 however, accompanies a large group of disorders, which include myasthenia gravis, myasthenic syndrome, botulism, motor neuron disease, poliomyelitis, and syringomyelia as well as early stages of reinnervation. Motor unit potentials in myotonia congenita typically show a progressive decline in amplitude with successive discharges, reflecting changes that

Phases The portion of a waveform between the departure from and return to the baseline is called a phase. The number of phases, determined by counting negative and positive peaks to and from the baseline, equals the number of baseline crossings plus one. In contrast to normal motor unit potentials, with four or fewer phases (Fig. 35–36), desynchronized units show polyphasic potentials with more than four phases, reflecting either fiber size variability or random

8.8 9.0 9.2 9.4 9.6 9.9 10.1 10.4 10.7 11.4 12.2 13.0 13.4 13.8 14.3 14.8 15.1 15.3 15.5 15.7

7.1 7.3 7.5 7.7 7.8 8.0 8.2 8.5 8.7 9.2 9.9 10.6 10.9 11.2 11.6 12.0 12.3 12.5 12.6 12.8

Biceps Deltoideus Brachii 8.1 8.3 8.5 8.6 8.7 9.0 9.2 9.6 9.9 10.4 11.2 12.0 12.4 12.7 13.2 13.6 13.9 14.1 14.3 14.4

Triceps Brachii 6.6 6.8 6.9 7.1 7.2 7.4 7.5 7.8 8.1 8.5 9.2 9.8 10.1 10.3 10.7 11.1 11.3 11.5 11.6 11.8

Extensor Digitorum Communis 7.9 8.1 8.3 8.5 8.6 8.9 9.1 9.4 9.7 10.2 11.0 11.7 12.1 12.5 12.9 13.3 13.6 13.9 14.0 14.2

Opponens Pollicis; Interosseus 9.2 9.5 9.7 9.9 10.0 10.3 10.5 10.9 11.2 11.9 12.8 13.6 14.1 14.5 15.0 15.5 15.8 16.1 16.3 16.5

8.0 8.2 8.4 8.6 8.7 9.0 9.2 9.5 9.8 10.3 11.1 11.8 12.2 12.5 13.0 13.4 13.7 14.0 14.1 14.3

7.1 7.3 7.5 7.7 7.8 8.0 8.2 8.5 8.7 9.2 9.9 10.6 10.9 11.2 11.6 12.0 12.3 12.5 12.6 12.8

8.9 9.2 9.4 9.6 9.7 10.0 10.2 10.5 10.8 11.5 12.3 13.2 13.6 13.9 14.4 14.9 15.2 15.5 15.7 15.9

6.5 6.7 6.8 6.9 7.0 7.2 7.4 7.6 7.8 8.3 8.9 9.5 9.8 10.1 10.5 10.8 11.0 11.2 11.4 11.5

Biceps Abductor Femoris; Tibialis Peroneus Digiti Quinti Quadriceps Gastrocnemius Anterior Longus

Leg Muscles

7.0 7.2 7.4 7.6 7.7 7.9 8.1 8.4 8.6 9.1 9.8 10.5 10.8 11.1 11.5 11.9 12.2 12.4 12.5 12.7

4.2 4.3 4.4 4.5 4.6 4.7 4.8 5.0 5.1 5.4 5.8 6.2 6.4 6.6 6.8 7.0 7.1 7.3 7.4 7.5

Orbicularis Oris Extensor Superior; Digitorum Triangularis; Brevis Frontalis

Facial Muscles

*The values given are mean values from different subjects without evidence of neuromuscular disease. The standard deviation of each value is 15% (20 potentials for each muscle). Therefore, deviations up to 20% are considered within the normal range when comparing measurements in a given muscle with the values of the table. From Buchthal, F.: An Introduction to Electromyography. Copenhagen, Scandinavian University Books, 1957, with permission.

0 3 5 8 10 13 15 18 20 25 30 35 40 45 50 55 60 65 70 75

Age in Years

Arm Muscles

Table 35–23. Mean Action Potential Duration (in Milliseconds) in Various Muscles at Different Ages (Concentric Electrodes)*

Nerve Conduction and Needle Electromyography

951

FIGURE 35–36 A, Normal motor unit potentials from minimally contracted biceps in a 40-year-old healthy man (a, b, and c) and maximally contracted tibialis anterior in a 31-year-old woman with hysterical weakness (d, e, and f). In both, low firing frequency indicates weak voluntary effort. B, Normal variations of motor unit potentials from the same motor unit in the biceps of the same healthy subject shown in a through c above. Tracings a through h represent eight slightly different sites of recording with the patient maintaining isolated discharges of a single motor unit. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

affect constituent single muscle fiber action potentials during continued contraction.

Lower Motor Neuron Versus Myopathic Disorders Various disorders of the lower motor neuron that share the same abnormalities in electromyography include motor neuron disease, poliomyelitis, syringomyelia, and diseases of the peripheral nerve.94,337 In these disorders, denervated muscle fibers undergo reinnervation, which leads to an increase in size of motor unit potentials (Fig. 35–38). The small recording radius of a monopolar or concentric needle is not suited to identify the territory of motor unit potentials, but a macro study can delineate the size of discharging units. Sprouting axon terminals usually remain within their own motor unit territory, without reaching out to the denervated muscle fibers located outside this boundary. Thus, after reinnervation, the number of muscle fibers increases with incorporation of denervated fibers, but the territory of the surviving motor unit remains the same. More specifically, a larger amplitude reflects a greater

muscle fiber density, whereas an increased duration results from a greater range in length and conduction time of regenerating axon terminals, as predicted by computer simulation.233 With abnormal synchronization at the cord level or ephaptic activation at the root level or near the axon terminal,318 two or more motor units may discharge simultaneously showing a complex waveform.55 Primary myopathic disorders, characterized by reduction in amplitude and duration of the motor unit potential (Fig. 35–39), include muscular dystrophy, congenital myopathies, periodic paralysis, myositis, and disorders of neuromuscular transmission such as myasthenia gravis and myasthenic syndromes. All of these entities have in common a random loss of functional muscle fibers from each motor unit as the result of muscle degeneration, inflammation, metabolic changes, or failure of neuromuscular activation. A lower fiber density from a decreased number of muscle fibers, in turn, causes reduction in amplitude and duration of motor unit potentials. In extreme cases, only a single muscle fiber represents a motor unit potential, showing voluntary activities indistinguishable from a fibrillation potential. Thus a high-frequency sound produced

952

Nerve Conduction and Electromyography

or increased duration associated with myopathic changes probably reflect an abnormally increased range of muscle fiber diameter.264 Additionally, a greater fiber density from muscle fiber regeneration results in much larger amplitude than might ordinarily be expected for a myopathy. In analyzing abnormalities of motor unit potentials, therefore, an oversimplified clinical dichotomy between myopathy and neuropathy does not necessarily hold.96,376 Nonetheless, it is reassuring that electromyography and histochemical findings from muscle biopsies have an overall concordance of 90% or greater.20,35,36,146

DISCHARGE PATTERN OF MOTOR UNITS Recruitment

FIGURE 35–37 Polyphasic motor unit potentials from the anterior tibialis in a 52-year-old man with amyotrophic lateral sclerosis. Temporal variability of repetitive discharges in waveform suggests intermittent blocking of some axon terminals. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

by short spikes, 1 to 2 ms in duration, resembles that of spontaneously discharging fibrillation potentials over the loudspeaker. Reversible changes not seen in inherited disorders of muscle may characterize metabolic or toxic myopathies.34 Motor unit potentials usually remain normal in mild metabolic and endocrine myopathies, with little or no alteration in duration or amplitude. Despite ordinarily contrasting changes in the waveform of motor unit potentials between myopathies and lower motor neuron disorders,35,36 some cases with equivocal distinction challenge even experienced clinicians.95 Distal neuropathies have sick axon terminals with a random loss of muscle fibers within a motor unit. Similarly, recently reinnervated immature motor units with a few muscle fibers give rise to low-amplitude and short-duration motor unit potentials. Thus, in both instances of a neuropathic process, motor units show changes classically regarded as consistent with a myopathy.261 Conversely, motor unit potentials with a long duration may erroneously suggest neuropathic changes in myopathies with regenerating muscle fibers.75,222,302 Some of these potentials may be so delayed from the main unit as to cause the appearance of a satellite or late component. Complex potentials with normal

In a healthy subject, a mild voluntary contraction initially activates one or two motor units with small, type I muscle fibers at a rate of several impulses per second (see Fig. 35–36). Additional units are then recruited in a fixed order according to the size principle.97,150,326 During mild constant contraction, motor units typically discharge semirhythmically, slowly increasing, then decreasing the interspike intervals. Greater levels of muscle contraction change the firing rate to grade the muscle force. Thus increase in muscle force induces two separate but related changes in the pattern of motor unit discharge: (1) recruitment of previously inactive units and (2) higher firing rate of already active units. Which of the two plays a greater role is not known, but both mechanisms operate simultaneously. In a normal recruitment pattern, the number of discharging motor units increases appropriately for the muscle force generated by the effort. In neuropathy, a loss of functional motor units results in late or decreased recruitment associated with rapid firing of discharging units to compensate for the reduced number (see Figs. 35–37 and 35–38). In myopathy, a random loss of muscle fibers makes each motor unit less efficient, necessitating an early or increased recruitment to compensate for a weak force generated by each unit (see Fig. 35–39). Thus recruitment is said to be decreased or increased if fewer or a greater number of units discharge than the number expected for a given force being exerted. In parkinsonism and other disorders of basal ganglia, motor units often fire in groups at or near the rates of tremor.72 Upper motor neuron lesions such as spinal cord injury also alter motor unit forces and recruitment patterns.356 Unlike lower motor neuron dysfunction, however, the discharging units fail to fire rapidly to compensate for a decreased recruitment, because of the lack of central drive. In healthy subjects, the recruitment frequency, defined as the firing frequency of the initially activated unit at the time an additional unit is recruited, averages 5 to 10 Hz, depending on the types of motor units under study.141,294,295

Nerve Conduction and Needle Electromyography

953

FIGURE 35–38 A, Motor unit potentials from the extensor digitorum communis in a 20-year-old man with partial radial nerve palsy. Minimal (a and d), moderate (b and e) and maximal (c and f) voluntary contraction recruited only a single motor unit that discharged at progressively higher rates. B, Motor unit potentials from the extensor carpi ulnaris (a, b, and c) and extensor carpi radialis longus (d, e, and f) in the same subject. Maximal voluntary contraction recruited only a single motor unit firing at a high discharge rate. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

This measure often deviates from the expected range in patients with neuromuscular disorders,121 although normal and abnormal values overlap considerably. Another practical measure of recruitment is the ratio of the average firing rate to the number of active units.67 Healthy subjects have a ratio less than 5 with, for example, three units firing less than 15 Hz each. A ratio greater than 10, with two units firing over 20 Hz, indicates a loss of motor units. At greater effort (Fig. 35–40), simultaneous activation of many different units makes recognition of individual motor unit potentials difficult; hence the term interference pattern. Its analysis by the spike density and the average amplitude of the summated response provides a simple quantitative means of evaluating the number of firing units

activated with maximal contraction. This measure reflects the complex interaction among many different biologic factors. These include descending input from the cortex, number of discharging motor neurons, firing frequency of each motor unit, waveform of individual potentials, and probability of phase cancellation.

Measurements of Turns and Amplitude Weak voluntary effort mainly activates motor units composed of low-threshold type I muscle fibers. Wider ranges of motor units are represented in quantitative assessment of the interference pattern during strong muscle contraction.333 An automated technique to assess the number of

954

Nerve Conduction and Electromyography

duration, and number of turns with age. The ratio of turns to mean amplitude is increased in myopathy, especially at 10% to 20% of maximum force, whereas it is decreased in neurogenic disorders, mainly at a force of 20% to 30%.120 Similarly, the ratio of root mean square voltage to turns is increased in chronic neuropathies.109 Quantitative measurements of recruitment patterns in infants and young children also help differentiate primary muscle disease from neurogenic lesions.336 In contrast to evaluation of individual potentials, which allows precise description of motor units and their temporal stability, analysis of recruitment demonstrates the number and discharge pattern of all the motor units to characterize an overall muscle performance.

FIGURE 35–39 Low-amplitude, short-duration motor unit potentials from the biceps (a, b, and c) and tibialis anterior (d, e, and f) in a 7-year-old boy with Duchenne type muscular dystrophy. Minimal voluntary contraction recruited an excessive number of motor units in both muscles. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

“turns” counts directional changes of a waveform that do not necessarily cross the baseline during constant levels of muscle contraction.383 This method selects potential changes exceeding a minimum excursion usually of 100 V, measuring the amplitude from one point of change in direction to the next during a fixed time epoch.147 Automated methods usually report turns frequencies,171 the maximal ratio of turns to mean amplitude or peak ratio, and the number of time intervals between turns.237 Turns and spectral analyses of interference patterns, albeit indirectly, serve to characterize the physiologic properties of the motor units.81 With greater voluntary contraction, the number of turns increases faster than does the mean amplitude change between turns at low to moderate force levels, followed by a reversed pattern at higher force levels.265 Therefore, the relationship between the turns and amplitude critically depends on the level of effort during the recordings, as indicated by the shape of the “normal cloud.” For example, an increment of voluntary contraction from 10% to 30% of the maximal significantly increased the mean firing rate, number of turns, average amplitude, and rise rate.80 Clinical studies therefore require precise control of contractile force as a major determinant of waveform and firing properties. In one study,158 both low-threshold and high-threshold motor units showed a linear increase in mean amplitude,

FIGURE 35–40 Interference patterns seen in the triceps of a 44-year-old healthy man (a), tibialis anterior of a 52-year-old man with amyotrophic lateral sclerosis (b), and quadriceps of a 20-year-old man with limb-girdle dystrophy (c). Discrete single motor unit discharge in b stands in good contrast to abundant motor unit potentials with reduced amplitude in c. (From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.)

Nerve Conduction and Needle Electromyography

Lower and Upper Motor Neuron Disorders As mentioned earlier, the number and the average force contributed by each functional motor unit determine the recruitment pattern. A limited recruitment in disorders of the motor neuron, root, or peripheral nerve results from a reduced number of excitable motor units, which fire at an inappropriately rapid rate. In an extreme example, a single motor unit potential discharges at frequencies as high as 50 Hz, producing a discrete “picket fence” appearance at maximal effort (see Figs. 35–38 and 35–40). Patients with upper motor neuron lesions may also have characteristic firing behavior in the temporal discharge pattern of single motor units. In one study of 15 stroke patients with paretic tibialis anterior,118 low-threshold motor units fired within the lower end of the normal range, whereas high-threshold motor units, if recruited at all, discharged below their normal range. In another study with hemiparetic patients,130 two characteristic patterns emerged: compression in the range of motor neuron recruitment forces and failure to discharge motor units at a higher rate despite maximal effort to contract the paretic muscles. Thus an upper motor neuron or hysterical paralysis shows a slow rate of discharge despite a late recruitment and a reduced interference pattern. Additional characteristics of hysterical weakness or poor cooperation include irregular, grouped firing of motor units, not seen in a genuine paresis unless the patient also suffers from essential or other type of tremor.

Myopathy In myopathy, many units are recruited early to functionally compensate in quantity for a smaller force per unit. In typical myopathies (see Fig. 35–39), many axons discharge almost instantaneously with a slight voluntary effort, developing a full interference pattern at less than maximal contraction (see Fig. 35–40). In diseases of neuromuscular transmission, motor units also show an early recruitment, reaching a full interference prematurely for the same reason. Loss of whole motor units rather than individual muscle fibers in advanced myogenic disorders, however, may lead to a limited recruitment and incomplete interference pattern, mimicking a neuropathic change caused by a reduced number of motor units.

Involuntary Movement Electromyography helps characterize involuntary motor symptoms, which include neuromyotonia and related disorders associated with spontaneous discharges of motor units. In tremors, irregular bursts of motor unit potentials repeat at a fairly constant rate. During each burst, which varies in amplitude, duration, and waveform, many motor units are active simultaneously, showing no fixed temporal or spatial relationships. This pattern of grouping, when seen in association with a subclinical tremor burst, may mimic a polyphasic motor unit

955

potential. Electromyographic recordings can quantitatively delineate the rate, rhythm, and distribution of different types of tremor.328 Aberrant regeneration, in the setting of hemifacial spasm, for example, gives rise to clinical synkinesis or unintended coactivation of distant muscles not voluntarily contracted. In this condition, simultaneous recording shows the presence of time-locked discharge of motor units in multiple muscles involved in synkinetic movements. This should be distinguished from ordinary associated movement, often seen in the face, without a precise time relationship among co-contracting muscle groups.

OTHER ASPECTS OF ELECTROMYOGRAPHY Principles of Localization In addition to the studies of myopathies and neuropathies, electromyography plays a critical role in delineating the location and distribution of a lesion responsible for clinical abnormalities. In particular, a unique combination of clinical and electrodiagnostic findings helps elucidate the level of a radicular and plexus lesion.253 The clinical assessment depends on the distribution of sensory deficits and of motor involvement that consist of weakness, muscle atrophy, hyporeflexia, fatigue, cramps, and fasciculations. Electromyographic studies can confirm the clinical localization of a lesion if appropriate muscle groups are selected for needle examination based on neurologic findings225 (see Tables 35–24 and 35–25). In the upper limbs, motor deficits usually provide a more reliable localizing sign than sensory impairments (see Table 35–24). The reverse seems to prevail in the lower limbs, where anatomic peculiarity makes electrophysiologic localization of radicular lesions more difficult. Evidence of denervation in the deep cervical muscles innervated by the posterior, as opposed to anterior, rami indicates an intraforaminal lesion affecting the root or spinal nerve before the division into the two rami. Thus studies of paraspinal muscles, by documenting the involvement of the posterior rami, serve to distinguish between a radicular and a plexus lesion. Other muscles of interest in this regard include the rhomboids, supplied by the dorsal scapular nerve, and the serratus anterior, subserved by the long thoracic nerves. Because of their very proximal innervation, spontaneous activity recorded therein can also confirm a radicular as opposed to a plexus lesion. In the lumbosacral region, discs tend to compress the roots slightly above the level of their respective foramina before their lateral deviation toward the exit. Needle studies help identify the damaged root and confirm the diagnosis (see Table 35–25). As in the cervical region, the evidence of denervation in the paraspinal muscles locates the lesion to the root or spinal nerve, which is proximal to the origin of the posterior ramus. The reverse does not necessarily hold because

956

Nerve Conduction and Electromyography

Table 35–24. Innervation Patterns of the Cranial, Shoulder Girdle, and Upper Limb Muscles* Nerves Anterior Primary Rami Cervical plexus Spinal accessory nerve Phrenic nerve Brachial plexus Dorsal scapular nerve Suprascapular nerve Axillary nerve Subscapular nerve Musculocutaneous nerve

Long thoracic nerve Lateral pectoral nerve Medial pectoral nerve Radial nerve

Posterior interosseous nerve

Median nerve

Anterior interosseous nerve

Muscles

C2

Sternocleidomastoid Trapezius, upper, middle, lower Diaphragm

C3

C4

C5

C6

C7

C8

T1

d

Rhomboid Supraspinatus Infraspinatus Teres minor Deltoid, anterior, middle, posterior Teres major Brachialis Biceps brachi Coracobrachialis Serratus anterior Pectoralis major (clavicular part) Pectoralis minor Brachioradialis Extensor carpi radialis Triceps, long, lateral, middle heads Anconeus Supinator Extensor carpi ulnaris Extensor digitorum Extensor pollicis brevis Extensor indicis Pronator teres Flexor carpi radialis Abductor pollicis brevis Flexor digitorum pronfundus (I & II) Pronator quadratus Flexor pollicis longus

Ulnar nerve

Flexor digitorum profundus (III & IV) Flexor carpi ulnaris Adductor pollicis Abductor digiti minimi Interossei, volar (I–III), dorsal (I–IV)

Posterior Primary Rami

Cervical erector spinae

*Solid black indicates primary innervation; shading indicates secondary innervation. From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.

root compression may spare the paraspinal muscles. A cauda equina lesion, though asymmetrical in distribution, otherwise resembles a conus medullaris lesion, with bilateral involvement of the level ordinarily unaffected by a herniated lumbar disc. Electromyographic studies show fibrillation potentials

and large motor unit potentials in the myotomes of several lumbosacral roots, including paraspinal muscles17 and the urethral sphincter.114 These findings resemble those of an intrinsic conus involvement except for an asymmetrical distribution of the abnormalities with spread above the sacral region.

957

Nerve Conduction and Needle Electromyography

Table 35–25. Innervation Patterns of the Hip Girdle and Lower Limb Muscles* Nerves Anterior Primary Rami Lumbosacral plexus Femoral nerve

Obturator nerve Superior gluteal nerve

Inferior gluteal nerve Sciatic nerve Tibial division Peroneal division Common peroneal nerve Deep peroneal nerve

Superficial peroneal nerve

Tibial nerve

Medial plantar nerve Lateral plantar nerve

Posterior Primary Rami

Muscles

L2

L3

L4

L5

S1

S2

Iliopsoas Sartorius Rectus femoris Vastus lateralis, medialis Gracilis Adductor longus, brevis, magnus Gluteus medius Gluteus minimus Tensor fasciae latae Gluteus maximus

Semitendinosus, semimembranosus Biceps femoris, long head Biceps femoris, short head

Tibialis anterior Extensor digitorum longus Extensor digitorum brevis Extensor hallucis longus Peroneus longus Peroneus brevis Tibialis posterior Flexor digitorum longus Flexor hallucis longus Gastrocnemius, medial head Gastrocnemius, lateral head Soleus Abductor hallucis Abductor digiti minimi Interossei Lumbosacral erector spinae

*Solid black indicates primary innervation; shading indicates secondary innervation. From Kimura, J.: Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice, 3rd ed. New York, Oxford University Press, 2001, with permission.

The multifidus muscles receive innervation from a single root, in contrast to the rest of the paraspinal muscles, which are subserved polysegmentally.48 Nonetheless, paraspinal abnormalities usually do not reveal the exact level of the involved segment.144 Determination of the affected myotome, therefore, depends on careful exploration of the limb muscles to delineate the pattern of involvement. Anomalous innervation among different cord segments abounds, challenging the clinician in attributing any pattern of clinical or electromyographic findings to a specific spinal level.301 A computer-aided expert system may have some place in the evaluation of brachial plexus

injuries as advocated by some,111 although its clinical value remains unknown. Some studies report a high correlation among electromyographic evidence of denervation, myelographic abnormalities, and surgical findings.235 In one series,267 however, electromyography and magnetic resonance imaging (MRI) agreed in only 60% of patients. Because one or the other evaluation showed some abnormality in the remaining 40%, these two modes of study provided complementary diagnostic information. T2-weighted and short time to inversion recovery sequences of MRI can supplement the electrophysiologic evaluations of functional deficits by detecting denervated

958

Nerve Conduction and Electromyography

skeletal muscle through increased signal intensity. One study50 reported a close correlation between this abnormality and spontaneous activities on electromyographic examination. Electromyographic evaluation can uncover functional deficits that do not necessarily produce structural abnormalities.363 In addition, such studies may also elucidate the extent and chronicity of lesions,381 and can guide patient management by substantiating clinical progression or improvement.170 Electrodiagnosis plays a particularly important role in justifying surgical exploration if radiologic and clinical findings conflict.368 For example, extraforaminal compression of the L5 root by lumbosacral ligaments, though not detectable by myelography and other imaging studies, may cause denervation.281 Conversely, asymptomatic subjects with abnormal MRI scans of the lumbar spine must undergo further study to seek a physiologic and clinical correlation.21

volitional contraction, give rise to motor unit potentials. Structural or functional disturbances of the motor unit seen in diseases of the nerve or muscle lead to alterations of waveform and discharge patterns of their electrical signals. The study of motor unit potentials provides information useful in elucidating the nature of the disease. Certain characteristics of such abnormalities may suggest a particular pathologic process. As a clinical tool, electromyography serves best only if the examiner conducts the study and interprets the results in light of the patient’s history, physical examination, and other diagnostic findings. In fact, the study constitutes an extension of the physical examination, rather than an independent laboratory test. Therefore, it is most useful if performed by a physician thoroughly familiar with the patient’s clinical findings with the aim to prove or disprove the diagnostic impression.

Sequence of Abnormalities After nerve injury, needle examination of the affected muscle initially reveals poor recruitment of motor unit potentials, which may result from either structural or functional loss of axons. Fibrillation potentials and positive sharp waves appear 2 to 3 weeks after axonal degeneration. During active regeneration of motor axons, low-amplitude, polyphasic motor unit potentials have temporal instability. Subsequent development of high-amplitude, long-duration motor unit potentials with stable configuration signals completion of reinnervation. Axonal degeneration after a nerve injury is said to follow a length-dependent delay, with muscles subserved by a shorter nerve showing the signs of denervation earlier.256 If so, needle examination should detect fibrillation potentials and positive sharp waves first in the paraspinal muscles. In one study using multivariate estimates,299 however, paraspinal spontaneous activity showed no correlation with symptom duration. In clinical practice, therefore, this time relationship may not necessarily hold.78

SUMMARY Electromyographic studies analyze spontaneous and voluntarily activated muscle action potentials extracellularly. Following brief injury potentials coincident with the insertion of the needle, a relaxed normal muscle remains electrically silent except for the end-plate activities. Several types of spontaneous discharges all signal diseases of the nerve or muscle, although they carry different clinical implications. Both fibrillation potentials and positive sharp waves represent spontaneous excitation of individual muscle fibers. The complex repetitive discharges comprise high-frequency spikes derived from multiple muscle fibers, which discharge sequentially, maintaining a fixed order. In conventional electromyography, isolated discharges of a single motor unit, the smallest functional element of

REFERENCES 1. Ackil, A. A., Shahani, B. R., and Young, R. R.: Sural nerve conduction studies and late responses in children undergoing hemodialysis. Arch. Phys. Med. Rehabil. 62:487, 1981. 2. Afifi, A. K., Kimura, J., and Bell, W. E.: Hypothermiainduced reversible polyneuropathy: electrophysiologic evidence of axonopathy. Pediatr. Neurol. 4:49, 1988. 3. Aho, K., and Sainio, K.: Late irradiation-induced lesions of the lumbosacral plexus. Neurology 33:953, 1983. 4. Albers, J. W.: Clinical neurophysiology of generalized polyneuropathy. J. Clin. Neurophysiol. 10:149, 1993. 5. Albers, J. W.: Principles of sensory nerve conduction studies. Presented at the 49th Annual Meeting of the American Academy of Neurology, Boston, April 12–19, 1997. 6. Albers, J. W., Allen, A. A., Bastron, J. A., and Daube, J. R.: Limb myokymia. Muscle Nerve 4:494, 1981. 7. American Association of Electrodiagnostic Medicine: Consensus criteria for the diagnosis of partial conduction block. Muscle Nerve 22(Suppl 8):S225, 1999. 8. American Association of Electrodiagnostic Medicine: Practice parameter for electrodiagnostic studies in ulnar neuropathy at the elbow: summary statement. Muscle Nerve 22:408, 1999. 9. Amoiridis, G.: Median-ulnar nerve communications and anomalous innervation of the intrinsic hand muscles: an electrophysiological study. Muscle Nerve 15:576, 1992. 10. Argyropoulos, C. J., Panayiotopoulos, C. P., and Scarpalezos, S.: F- and M-wave conduction velocity in amyotrophic lateral sclerosis. Muscle Nerve 1:479, 1978. 11. Ashworth, N. L., Marshall, S. C., and Satkunam, L. E.: The effect of temperature on nerve conduction parameters in carpal tunnel syndrome. Muscle Nerve 21:1089, 1998. 12. Auger, R. G.: Role of the masseter reflex in the assessment of subacute sensory neuropathy. Muscle Nerve 21:800, 1998. 13. Baba, M., Gilliatt, W., and Jacobs, J. M.: Recovery of distal changes after nerve constriction by a ligature. J. Neurol. Sci. 60:235, 1983.

Nerve Conduction and Needle Electromyography 14. Bamford, C. R., Rothrock, R. J., and Swenson, M.: Average techniques to define the low-amplitude compound motor action potentials. Arch. Neurol. 41:1307, 1984. 15. Barkhaus, P. E., and Nandedkar, S. D.: Recording characteristics of the surface EMG electrodes. Muscle Nerve 17:1317, 1994. 16. Barkhaus, P. E., and Nandedkar, S. D.: On the selection of concentric needle electromyogram motor unit action potentials: is the rise time criterion too restrictive? Muscle Nerve 19:1554, 1996. 17. Bartleson, J. D., Cohen, M. D., Harrington, T. M., et al.: Cauda equina syndrome secondary to long-standing ankylosing spondylitis. Ann. Neurol. 14:662, 1983. 18. Bertram, M. F., Nishida, T., Minieka, M. M., et al.: Effects of temperature on motor unit action potentials during isometric contraction. Muscle Nerve 18:1443, 1995. 19. Bischoff, C., Stålberg, E., Falck, B., and Puksa, L.: Significance of A-waves recorded in routine motor nerve conduction studies. Electroencephalogr. Clin. Neurophysiol. 101:528, 1996. 20. Black, J. T., Bhatt, G. P., DeJesus, P. V., et al.: Diagnostic accuracy of clinical data, quantitative electromyography and histochemistry in neuromuscular disease: a study of 105 cases. J. Neurol. Sci. 21:59, 1974. 21. Boden, S. D., Davis, D. O., Dina, T. S., et al.: Abnormal magnetic-resonance scans of the lumbar spine in asymptomatic subjects. J. Bone Joint Surg. Am. 72:403, 1990. 22. Botte, M. J., Cohen, M. S., Lavernia, C. J., et al.: The dorsal branch of the ulnar nerve: an anatomic study. J. Hand Surg. Am. 15:603, 1990. 23. Bougle, D., Denise, P., Yaseen, H., et al.: Maturation of peripheral nerves in preterm infants: motor and proprioceptive nerve conduction. Electroencephalogr. Clin. Neurophysiol. 75:118, 1990. 24. Bouisset, S.: EMG and muscle force in normal motor activities. In Desmedt, J. E. (ed.): New Developments in Electromyography and Clinical Neurophysiology, Vol. 1. Basel, Karger, p. 547, 1973. 25. Braddom, R. I., and Johnson, E. W.: Standardization of H reflex and diagnostic use in S1 radiculopathy. Arch. Phys. Med. Rehabil. 55:161, 1974. 26. Brick, J. F., Gutmann, L., and McComas, C. F.: Calcium effect on generation and amplification of myokymic discharges. Neurology 32:618, 1982. 27. Bromberg, M. B., and Albers, J. W.: Patterns of sensory nerve conduction abnormalities in demyelinating and axonal peripheral nerve disorders. Muscle Nerve 16:262, 1993. 28. Bromberg, M. B., and Jaros, L.: Symmetry of normal motor and sensory nerve conduction measurements. Muscle Nerve 21:498, 1998. 29. Bromberg, M. B., and Spiegelberg, T.: The influence of active electrode placement on CMAP amplitude. Electroencephalogr. Clin. Neurophysiol. 105:385, 1997. 30. Brooke, M. H., Carroll, J. E., Davis, J. E., and Hagberg, J. M.: The prolonged exercise test. Neurology 29:636, 1979. 31. Bryant, P. R., and Eng, G. D.: Normal values for the soleus H-reflex in newborn infants 31–45 weeks post conceptional age. Arch. Phys. Med. Rehabil. 72:28, 1991. 32. Buchthal, F.: An Introduction to Electromyography. Copenhagen, Scandinavian University Books, 1957.

959

33. Buchthal, F.: The general concept of the motor unit. In Adams, R. D., Eaton, L. M., and Shy, G. M. (eds.): Neuromuscular Disorders. Baltimore, Williams & Wilkins, p. 3, 1960. 34. Buchthal, F.: Electrophysiological abnormalities in metabolic myopathies and neuropathies. Acta. Neurol. Scand. Suppl. 43:129, 1970. 35. Buchthal, F.: Diagnostic significance of the myopathic EMG. In Rowland, L. P. (ed.): Pathogenesis of Human Muscular Dystrophies: Proceedings of the Fifth International Scientific Conference of the Muscular Dystrophy Association, Durango, CO, June 1976. Amsterdam, Excerpta Medica, p. 205, 1977. 36. Buchthal, F.: Electrophysiological signs of myopathy as related with muscle biopsy. Acta Neurol. (Napoli) 32:1, 1977. 37. Buchthal, F.: Fibrillations: clinical electrophysiology. In Culp, W. J., and Ochoa, J. (eds.): Abnormal Nerves and Muscles as Impulse Generators. Oxford, UK, Oxford University Press, p. 632, 1982. 38. Buchthal, F., Engbaek, L., and Gamstorp, I.: Paresis and hyperexcitability in adynamia episodica hereditaria. Neurology 8:347, 1958. 39. Buchthal, F., and Rosenfalck, A.: Evoked action potentials and conduction velocity in human sensory nerves. Brain Res. 3:1, 1966. 40. Buchthal, F., Rosenfalck, A., and Behse, F.: Sensory potentials of normal and diseased nerves. In Dyck, P. J., Thomas, P. K., and Lambert, E. H. (eds.): Peripheral Neuropathy, Vol. 1. Philadelphia, W. B. Saunders, p. 442, 1975. 41. Buchthal, F., Rosenfalck, A., and Trojaborg, W.: Electrophysiological findings in entrapment of the median nerve at wrist and elbow. J. Neurol. Neurosurg. Psychiatry 37:340, 1974. 42. Buchthal, F., and Rosenfalck, P.: Electrophysiological aspects of myopathy with particular reference to progressive muscular dystrophy. In Bourne, G. H., and Golarz, M. N. (eds.): Muscular Dystrophy in Man and Animals. New York, Hafner Publishing Company, p. 194, 1963. 43. Buchthal, F., and Rosenfalck, P.: Spontaneous electrical activity of human muscle. Electroencephalogr. Clin. Neurophysiol. 20:321, 1966. 44. Buschbacher, R. M.: Body mass index effect on common nerve conduction study measurement. Muscle Nerve 21:1398, 1998. 45. Campbell, W. W.: The value of inching techniques in the diagnosis of focal nerve lesions. Muscle Nerve 21:1554, 1998. 46. Campbell, W. W.: Ulnar neuropathy at the elbow. Muscle Nerve 23:478, 2000. 47. Campbell, W. W., Pridgeon, R. M., and Sahni, K. S.: Short segment incremental studies in the evaluation of ulnar neuropathy at the elbow. Muscle Nerve 15:1050, 1992. 48. Campbell, W. W., Vasconcelos, O., and Laine, F. J.: Focal atrophy of the multifidus muscle in lumbosacral radiculopathy. Muscle Nerve 21:1350, 1998. 49. Carpendale, M. T. F.: Conduction time in the terminal portion of the motor fibers of the ulnar, median, and peroneal nerves in healthy subjects and in patients with neuropathy. Thesis, University of Minnesota, Minneapolis, 1956.

960

Nerve Conduction and Electromyography

50. Carter, G. T., and Fritz, R. C.: Electromyographic and lower extremity short time to inversion recovery magnetic resonance imaging findings in lumbar radiculopathy [short report]. Muscle Nerve 20:1191, 1997. 51. Cerra, D., and Johnson, E. W.: Motor nerve conduction velocity in premature infants. Arch. Phys. Med. Rehabil. 43:160, 1962. 52. Chang, C. W., Cho, H. K., and Oh, S. J.: Posterior antebrachial cutaneous neuropathy: case report. Electromyogr. Clin. Neurophysiol. 29:109, 1989. 53. Chaudhry, V., Corse, A. M., Freimer, M. L., et al.: Interand intraexaminer reliability of nerve conduction measurements in patients with diabetic neuropathy. Neurology 44:1459, 1994. 54. Cherington, M.: Accessory nerve: conduction studies. Arch. Neurol. 18:708, 1968. 55. Cholachis, S. C., Pease, W. S., and Johnson, E. W.: Polyphasic motor unit action potentials in early radiculopathy: their presence and ephaptic transmission as an hypothesis. Electromyogr. Clin. Neurophysiol. 32:27, 1992. 56. Chroni, E., and Panayiotopoulos, C. P.: F-wave quantitation in neuropathy. Muscle Nerve 18:786, 1995. 57. Claussen, G. C., Ahmad B. K., Sunwood, I. N., and Oh, S. J.: Combined motor and sensory median-ulnar anastomosis: report of an electrophysiologically proven case. Muscle Nerve 19:231, 1996. 58. Clay, S. A., and Ramseyer, J. C.: The orbicularis oculi reflex: pathologic studies in childhood. Neurology 27:892, 1977. 59. Colachis, S. C. III, Klejka, J. P., Shamir, D. Y., et al.: Amplitude of M responses: side to side comparability. Am. J. Phys. Med. Rehabil. 72:19, 1993. 60. Conrad, B., Aschoff, J. C., and Fischler, M.: Der Diagnostische Wert der F-Wellen-Latenz. J. Neurol. 210:151, 1975. 61. Conrad, B., Jacobi, H. M., and Prochazka, V. J.: Unusual properties of repetitive fasciculations. Electroencephalogr. Clin. Neurophysiol. 35:173, 1973. 62. Cornblath, D. R., Sumner, A. J., Daube, J., et al.: Conduction block in clinical practice. Muscle Nerve 14:869, 1991. 63. Crutchfield, C. A., and Gutmann, L.: Hereditary aspects of accessory deep peroneal nerve. J. Neurol. Neurosurg. Psychiatry 36:989, 1973. 64. Cruz Martinez, A., Barrio, M., Perez Conde, M. C., and Ferrer, M. T.: Electrophysiological aspects of sensory conduction velocity in healthy adults. 2. Ratio between the amplitude of sensory evoked potentials at the wrist on stimulating different fingers in both hands. J. Neurol. Neurosurg. Psychiatry 41:1097, 1978. 65. Cruz Martinez, A., Barrio, M., Perez Conde, M. C., and Gutierrez, A. M.: Electrophysiological aspects of sensory conduction velocity in healthy adults. 1. Conduction velocity from digit to palm, from palm to wrist, and across the elbow, as a function of age. J. Neurol. Neurosurg. Psychiatry 41:1092, 1978. 66. Cummins, K. L., Dorfman, L. J., and Perkel, D. H.: Nerve fiber conduction-velocity distributions. II. Estimation based on two compound action potentials. Electroencephalogr. Clin. Neurophysiol. 46:647, 1979.

67. Daube, J. R.: Needle Examination in Electromyography (Minimonograph no 11). Rochester, MN, American Association of Electromyography and Electrodiagnosis, 1979. 68. Daube, J. R.: AAEM Minimonograph #11: Needle examination in clinical electromyography. Muscle Nerve 14:685, 1991. 69. Daube, J. R.: Electrodiagnosis of muscle disorders. In Engel, A. G., and Franzini-Armstrong, C. (eds.): Myology. New York, McGraw-Hill, p. 764, 1994. 70. Daube, J. R., Kelly, J. J. Jr., and Martin, R. A.: Facial myokymia with polyradiculoneuropathy. Neurology 29:662, 1979. 71. de Carvalho, M., and Swash, M.: Fasciculation potentials: a study of amyotrophic lateral sclerosis and other neurogenic disorders. Muscle Nerve 21:336, 1998. 72. Dengler, R., Wolf, W., Schubert, M., and Struppler, A.: Discharge pattern of single motor units in basal ganglia disorders. Neurology 36:1061, 1986. 73. Denys, E. H.: AAEM Minimonograph #14: The influence of temperature in clinical neurophysiology. Muscle Nerve 14:795, 1991. 74. Desjacques, P., Egloff-Baer, S., and Roth, G.: Lumbrical muscles and the carpal tunnel syndrome. Electromyogr. Clin. Neurophysiol. 20:443, 1980. 75. Desmedt, J. E., and Borenstein, S.: Regeneration in Duchenne muscular dystrophy: electromyographic evidence. Arch. Neurol. 33:642, 1976. 76. Dessi, F., Durand, G., and Hoffmann, J. J.: The accessory deep peroneal nerve: a pitfall for the electromyographer. J. Neurol. Neurosurg. Psychiatry 55:214, 1992. 77. Devi, S., Lovelace, R. E., and Duarte, N.: Proximal peroneal nerve conduction velocity: recording from anterior tibial and peroneus brevis muscles. Ann. Neurol. 2:116, 1977. 78. Dillingham, T. R., Pezzin, L. E., and Lauder, T. D.: Cervical paraspinal muscle abnormalities and symptom duration: a multivariate analysis [short report]. Muscle Nerve 21:640, 1998. 79. Dioszeghy, P., and Stålberg, E.: Changes in motor and sensory nerve conduction parameters with temperature in normal and diseased nerve. Electroencephalogr. Clin. Neurophysiol. 85:229, 1992. 80. Dorfman, L. J., Howard, J. E., and McGill, K. C.: Influence of contractile force on properties of motor unit action potentials: ADEMG analysis. J. Neurol. Sci. 86:125, 1988. 81. Dorfman, L. J., and McGill, K. C.: AAEE Minimonograph #29: Automatic quantitative electromyography. Muscle Nerve 11:804, 1988. 82. Dumitru, D.: Single muscle fiber discharges (insertional activity, end-plate potentials, positive sharp waves, and fibrillation potentials): a unifying proposal. Muscle Nerve 19:221, 1996. 83. Dumitru, D., and King, J. C.: Motor unit action potential duration and muscle length. Muscle Nerve 22:1188, 1999. 84. Dumitru, D., King, J. C., and Nandedkar, S. D.: Comparison of single-fiber and macro electrode recordings: relationship to motor unit action potential duration. Muscle Nerve 20:1381, 1997.

Nerve Conduction and Needle Electromyography 85. Dumitru, D., King, J. C., and Nandedkar, S. D.: Motor unit action potentials recorded with concentric electrodes: physiologic implications. Electroencephalogr. Clin. Neurophysiol. 105:333, 1997. 86. Dumitru, D., King J. C., Rogers, W., and Stegeman, D. F.: Positive sharp wave and fibrillation potential modeling. Muscle Nerve 22:242, 1999. 87. Dyck, P. J., Kratz, K. M., Lehman, K. A., et al.: The Rochester Diabetic Neuropathy Study: design, criteria for types of neuropathy, selection bias, and reproducibility of neuropathic tests. Neurology 41:799, 1991. 88. Dyck, P. J., Lambert, E. H., and Nichols, P. C.: Quantitative measurement of sensation related to compound action potential and number and sizes of myelinated and unmyelinated fibers of sural nerve in health, Friedreich’s ataxia, hereditary sensory neuropathy, and tabes dorsalis. In Remond, A. (ed.): Handbook of Electroencephalography and Clinical Neurophysiology, Vol. 9. Amsterdam, Elsevier, p. 83, 1972. 89. Eisen, A.: The physiologic basis and clinical applications of needle EMG in neuromuscular abnormalities: principles and pitfalls in the practice of EMG and NCS. Presented at the 49th Annual Meeting of the American Academy of Neurology, Boston, April 12–19, 1997. 90. Eisen, A., and Danon, J.: The orbicularis oculi reflex in acoustic neuromas: a clinical and electrodiagnostic evaluation. Neurology 24:306, 1974. 91. Eisen, A., Schomer, D., and Melmed, C.: The application of F-wave measurements in the differentiation of proximal and distal upper limb entrapments. Neurology 27:662, 1977. 92. Ekstedt, J., and Stalberg, E.: How the size of the needle electrode leading-off surface influences the shape of the single muscle fibre action potential in electromyography. Comput. Prog. Biomed. 3:204, 1973. 93. Emeryk, B., Hausmanowa-Petrusewicz, I., and Nowak, T.: Spontaneous volleys of bizarre high frequency potentials (b.h.f.p.) in neuro-muscular diseases. Part I. Occurrence of spontaneous volleys of b.h.f.p. in neuro-muscular diseases. Part II. An analysis of the morphology of spontaneous volleys of b.h.f.p. in neuromuscular diseases. Electromyogr. Clin. Neurophysiol. 14:303 (Part I); 339 (Part II), 1974. 94. Emeryk-Szajewska, B., Kopec, J., and Karwanska, A.: The reorganization of motor units in motor neuron disease. Muscle Nerve 20:306, 1997. 95. Engel, W. K.: Brief, small, abundant motor-unit action potentials: a further critique of electromyographic interpretation. Neurology 25:173, 1975. 96. Engel, W. K., and Warmolts, J. R.: The motor unit: diseases affecting it in toto or in partia. In Desmedt, J. E. (ed.): New Developments in Electromyography and Clinical Neurophysiology, Vol. 1. Basel, Karger, p. 141, 1973. 97. Erim, Z., De Luca, C. J., and Mineo, K.: Rank-ordered regulation of motor units. Muscle Nerve 19:563, 1996. 98. Ertekin, C., Mungan, B., and Uludag, B.: Sacral cord conduction time of the soleus H-reflex. J. Clin. Neurophysiol. 13:77, 1996. 99. Ertekin, C., Nejat, R. S., Sirin, H., et al.: Comparison of magnetic coil and needle-electrical stimulation in diagnosis of lumbosacral radiculopathy. Muscle Nerve 17:685, 1994.

961

100. Ertekin, C., Nejat, R. S., Sirin, H., et al.: Comparison of magnetic coil stimulation and needle electrical stimulation in diagnosis of lumbosacral radiculopathy. Clin. Neurol. Neurosurg. 96:124, 1994. 101. Falco, F. J. E., Hennessey, W. J., Braddom, R. L., and Goldberg, G.: Standardized nerve conduction studies in the upper limb of the healthy elderly. Am. J. Phys. Med. Rehabil. 71:263, 1992. 102. Feasby, T. E., Brown, W. F., Gilbert, J. J., and Hahn, A. F. D.: The pathological basis of conduction block in human neuropathies. J. Neurol. Neurosurg. Psychiatry 48:239, 1985. 103. Feinberg, D. M., Preston, D. C., Shefner, J. M., and Logigian, E. L.: Amplitude-dependent slowing of conduction in amyotrophic lateral sclerosis and polyneuropathy. Muscle Nerve 22:937, 1999. 104. Ferrante, M. A., Olney, R., and Wilbourn, A. J.: Sensory nerve conduction study workshop. Presented at the 49th Annual Meeting of the American Academy of Neurology, Boston, April 12–19, 1997. 105. Ferrante, M. A., and Wilbourn, A. J.: The utility of various sensory nerve conduction responses in assessing brachial plexopathies. Muscle Nerve 18:879, 1995. 106. Finsterer, J., and Mamoli, B.: Satellite potentials as a measure of neuromuscular disorders. Muscle Nerve 20:585, 1997. 107. Fisher, M. A.: H reflexes and F-wave physiology and clinical indications. Muscle Nerve 15:1223, 1992. 108. Fisher, M. A.: Whither F waves. In Kimura, J., and Shibasaki, H. (eds.): Recent Advances in Clinical Neurophysiology. Amsterdam, Elsevier, p. 752, 1996. 109. Fisher, M. A.: Root mean square voltage/turns in chronic neuropathies is related to increase in fiber density. Muscle Nerve 20:241, 1997. 110. Fisher, M. A., Shivde, A. J., Teixera, C., and Grainer, L. S.: Clinical and electrophysiological appraisal of the significance of radicular injury in back pain. J. Neurol. Neurosurg. Psychiatry 41:303, 1978. 111. Fisher, W. S. III: Computer-aided intelligence: application of an expert system to brachial plexus injuries. Neurosurgery 27:837, 1990. 112. Fitz, W. R., Mysiw, W. J., and Johnson, E. W.: First lumbrical latency and amplitude: control values and findings in carpal tunnel syndrome. Am. J. Phys. Med. Rehabil. 69:198, 1990. 113. Fowler, C. J., Kirby, R. S., and Harrison, M. J. G: Decelerating burst and complex repetitive discharges in the striated muscle of the urethral sphincter, associated with urinary retention in women. J. Neurol. Neurosurg. Psychiatry 48:1004, 1985. 114. Fowler, C. J., Kirby, R. S., Harrison, M. J. G., et al.: Individual motor unit analysis in the diagnosis of disorders of urethral sphincter innervation. J. Neurol. Neurosurg. Psychiatry 47:637, 1984. 115. Franssen, H., and Wieneke, G. H.: Nerve conduction and temperature: necessary warming time. Muscle Nerve 17:336, 1994. 116. Franssen, H., Wieneke, G. H., and Wokke, J. H. J.: The influence of temperature on conduction block. Muscle Nerve 22:166, 1999.

962

Nerve Conduction and Electromyography

117. Fraser, J. L., and Olney, R. K.: The relative diagnostic sensitivity of different F-wave parameters in various polyneuropathies. Muscle Nerve 15:912, 1992. 118. Frontera, W. R., Grimby, L., and Larsson, L.: Firing rate of the lower motoneuron and contractile properties of its muscle fibers after upper motoneuron lesion in man. Muscle Nerve 20:938, 1997. 119. Fu, R., DeLisa, J. A., and Kraft, G. H.: Motor nerve latencies through the tarsal tunnel in normal adult subjects: standard determinations corrected for temperature and distance. Arch. Phys. Med. Rehabil. 61:243, 1980. 120. Fuglsang-Frederiksen, A., Monaco, M., and Dahl, K.: Turns analysis (peak ratio) in EMG using the mean amplitude as a substitute of force measurement. Electroencephalogr. Clin. Neurophysiol. 60:225, 1985. 121. Fuglsang-Frederiksen, A., Smith, T., and Hogenhaven, H.: Motor unit firing intervals and other parameters of electrical activity in normal and pathological muscle. J. Neurol. Sci. 78:51, 1987. 122. Fullerton, P. M., and Gilliatt, R. W.: Axon reflexes in human motor nerve fibres. J. Neurol. Neurosurg. Psychiatry 28:1, 1965. 123. Fusfeld, R. D.: Electromyographic abnormalities in a case of botulism. Bull. Los Angeles Neurol. Soc. 35:164, 1970. 124. Gadia, M. T., Natori, N., Ramos, L. B., et al.: Influence of height on quantitative sensory, nerve conduction, and clinical indices of diabetic peripheral neuropathy. Diabetes Care 10:613, 1987. 125. Gamstorp, I.: Normal conduction velocity of ulnar, median and peroneal nerves in infancy, childhood and adolescence. Acta Paediatr. Suppl. 146:68, 1963. 126. Garcia-Meriono, A., Cabello, A., Mora, J. S., and Liano, H.: Continuous muscle fiber activity, peripheral neuropathy and thymoma. Ann. Neurol. 29:215, 1991. 127. Gassel, M. M.: A study of femoral nerve conduction time. Arch. Neurol. 9:57, 1963. 128. Geiringer, S. R.: Anatomic Localization for Needle Electromyography. Philadelphia, Hanley & Belfus, 1994. 129. Geiringer, S. R.: Inching techniques are of limited use. Muscle Nerve 21:1557, 1998. 130. Gemperline, J. J., Allen, S., Walk, D., and Rymer, W. Z.: Characteristics of motor unit discharge in subjects with hemiparesis. Muscle Nerve 18:1101, 1995. 131. Gilliatt, R. W.: Sensory conduction studies in the early recognition of nerve disorders. Muscle Nerve 1:352, 1978. 132. Gilliatt, R. W.: Acute compression block. In Sumner, A. (ed.): The Physiology of Peripheral Nerve Disease. Philadelphia, W. B. Saunders, p. 287, 1980. 133. Gilliatt, R. W., Goodman, H. V., and Willison, R. G.: The recording of lateral popliteal nerve action potentials in man. J. Neurol. Neurosurg. Psychiatry 24:305, 1961. 134. Gilliatt, R. W., Le Quesne, P. M., Logue, V., and Sumner, A. J.: Wasting of the hand associated with a cervical rib or band. J. Neurol. Neurosurg. Psychiatry 33:615, 1970. 135. Gilliatt, R. W., and Thomas, P. K.: Changes in nerve conduction with ulnar lesions at the elbow. J. Neurol. Neurosurg. Psychiatry 23:312, 1960. 136. Godaux, E., and Desmedt, J. E.: Evidence for a monosynaptic mechanism in the tonic vibration reflex of the human

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

148.

149.

150. 151.

152.

153.

154.

masseter muscle. J. Neurol. Neurosurg. Psychiatry 38:161, 1975. Godaux, E., and Desmedt, J. E.: Exteroceptive suppression and motor control of the masseter and temporalis muscles in normal man. Brain Res. 85:447, 1975. Golovchinsky, V.: Ulnar-to-median anastomosis and its role in the diagnosis of lesions of the median nerve at the elbow and the wrist. Electromyogr. Clin. Neurophysiol. 30:31, 1990. Griep, P. A. M., Gielen, F. L. H., Boom, H. B. K., et al.: Calculation and registration of the same motor unit action potential. Electroencephalogr. Clin. Neurophysiol. 53:388, 1982. Gruber, W.: Ueber die Verbindung des Nervus medianus mit dem Nervus ulnaris am Unterarme des Menschen und der Saugethiere. Arch. Anat. Physiol. Med. Leipzig, p. 501, 1870. Gunreben, G., and Schulte-Mattler, W.: Evaluation of motor unit firing rates by standard concentric needle electromyography. Electromyogr. Clin. Neurophysiol. 32:103, 1992. Gutmann, L.: Atypical deep peroneal neuropathy in presence of accessory deep peroneal nerve. J. Neurol. Neurosurg. Psychiatry 33:453, 1970. Gutmann, L.: Median-ulnar nerve communications and carpal tunnel syndrome. J. Neurol. Neurosurg. Psychiatry 40:982, 1977. Haig, A. J., Talley, C., Grobler, L. J., and LeBreck, D. B.: Paraspinal mapping: quantified needle electromyography in lumbar radiculopathy. Muscle Nerve 16:477, 1993. Hakamada, S., Kumagai, T., Watanabe, K., et al.: The conduction velocity of slower and the fastest fibres in infancy and childhood. J. Neurol. Neurosurg. Psychiatry 45:851, 1982. Hausmanowa-Petrusewicz, I., and Jedrzejowska, H.: Correlation between electromyographic findings and muscle biopsy in cases of neuromuscular disease [abstract]. J. Neurol. Sci. 13:85, 1971. Hayward, M.: Automatic analysis of the electromyogram in healthy subjects of different ages. J. Neurol. Sci. 33:397, 1977. Heckmann, R., and Ludin, H. P.: Differentiation of spontaneous activity from normal and denervated skeletal muscle. J. Neurol. Neurosurg. Psychiatry 45:331, 1982. Heifernan, L., Rewcastle, N. B., and Humphrey, J. G.: The spectrum of rod myopathies. Arch. Neurol. 18:529, 1968. Henneman, E.: Relation between size of neurons and their susceptibility to discharge. Science 126:1245, 1957. Hess, K., Kern, S., and Schiller, H. H.: Blink reflex in trigeminal sensory neuropathy. Electromyogr. Clin. Neurophysiol. 24:185, 1984. Hodgkin, A. L., and Katz, B.: The effect of temperature on the electrical activity of the giant axon of the squid. J. Physiol. (Lond.) 109:240, 1949. Hömberg, V., Reiners, K., and Toyka, K. V.: Reversible conduction block in human ischemic neuropathy after ergotamine abuse. Muscle Nerve 15:467, 1992. Hopf, H. C.: Forearm ulnar-to-median nerve anastomosis of sensory axons. Muscle Nerve 13:654, 1990.

Nerve Conduction and Needle Electromyography 155. Hopf, H. C.: Clinical implications of testing brainstem reflexes and corticobulbar connections in man. In Kimura, J., and Shibasaki, H. (eds.): Recent Advances in Clinical Neurophysiology. Amsterdam, Elsevier, p. 39, 1996. 156. Horowitz, S. H.: Correlation of near-nerve sural conduction and quantified sensory testing in patients with diabetic neuropathy. Muscle Nerve 18:1202, 1995. 157. Horowitz, S. H., and Krarup, C.: Conduction studies of the normal sural nerve. Muscle Nerve 15:374, 1992. 158. Howard, J. E., McGill, K. C., and Dorfman, L. J.: Age effects on properties of motor unit action potentials: ADEMG analysis. Ann. Neurol. 24:207, 1988. 159. Hudgson, P., Gardner-Medwin, D., Worsfold, M., et al.: Adult myopathy from glycogen storage disease due to acid maltase deficiency. Brain 91:435, 1968. 160. Hudson, A. J., Brown, W. F., and Gilbert, J. J.: The muscular pain-fasci